Citation
Migration, family, gender, and reproductive health of Mexican-origin young women

Material Information

Title:
Migration, family, gender, and reproductive health of Mexican-origin young women
Creator:
Coleman-Minahan, Kate ( author )
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file (413 pages). : ;

Subjects

Subjects / Keywords:
Reproductive health -- United States ( lcsh )
Hispanic American women -- Health and hygiene ( lcsh )
Hispanic American women -- Family relationships ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Latina adolescents have higher adolescent birth rates than any race/ethnic group and Mexican-origin Latinas have higher birth rates than Puerto Rican and Cuban adolescent. The purpose of this dissertation is to examine how cultural constructions of gender and sexuality help explain the relationship between immigrant generation, family characteristics, and reproductive health behavior of Mexican-origin young women. I used the National Longitudinal Study of adolescent Health (Add Health) to examine how immigrant generation moderated the relationship between family structure and family socioeconomic status (SES) and the hazard of sexual initiation, the hazard of an adolescent birth, and the odds of contraceptive use at first sex among Mexican-origin adolescents. I then conducted lie history interviews with 21 first and second generation Mexican-origin young women in Denver, who are currently the same ages of the Add Health participants to understand the process of migration and how migrant families shape gender. I found that two biological parents, grandparents, and older siblings were protective in reducing risky sexual and reproductive health behaviors. Parental education increased the hazard of sexual initiation in the first and second generations, but reduced the hazard in the third generation. Having a mother not working as compared to having a professionally employed mother, =reduced the hazard of sexual initiation and adolescent birth in the first and second generations and increased the odds of contraceptive use in all three generations. Parent-adolescent communication about sex, particularly encouragement to delay intercourse, was associated with a delay in sexual initiation and a lower risk for adolescent childbearing. Extremely gendered households were associated with an increased risk for sexual initiation and an adolescent birth and these households were of lower SES and had more instability than moderately gendered households. Thus, culture cannot be separated from SES and structural influences. My results contribute to a more comprehensive cultural framework to understand health disparities by incorporating overlooked structural influences, including migration.
Thesis:
Thesis (Ph.D.)--University of Colorado Denver. Health and behavioral sciences
Bibliography:
Includes bibliographic references.
General Note:
Department of Health and Behavioral Sciences
Statement of Responsibility:
by Kate Coleman-Minahan.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
910882907 ( OCLC )
ocn910882907

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

MIGRATION, FAMILY, GENDER, AND REPRODUCTIVE HEALTH OF MEXICAN ORIGIN YOUNG WOMEN by KATE COLEMAN MINAHAN B.S.N, Fairfield University, 2002 M.S., University of Colorado Denver, 2008 A thesis submitted to the Faculty at the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Health and Behavioral Sciences 2015

PAGE 2

ii 2015 KATE COLEMAN MINAHAN ALL RIGHTS RESERVED

PAGE 3

iii This thesis for the Doctor o f Philosophy degree by Kate Coleman Minahan has been approved for the Health and Behavioral Sciences Program by Sheana Bull, Dissertation Advisor Patrick M. Krueger Exam Chair Fernando Riosmena Jean Scandlyn April 10, 2015

PAGE 4

iv Coleman Minahan, Kate (Ph.D. Health and Behavioral Sciences) Migration, F amily, Gender, and R eproductive H ealth of Mexican origin Y oung W omen Thesis directed by Professor Sheana Bull ABSTRACT Latina adolescents have higher adolescent birth rates than any race/ethnic group and Mexican origin Latinas have higher birth rates than Puerto Rican and Cuban adolescents. The purpose of the dissertation is to examine how cultural constructions of gender and sexuality help explain the relationship between immigr ant generation, family characteristics, and reproductive health behavior of Mexican origin young women. I used the National Longitudinal Study of Adolescent Health (Add Health) to examine how immigrant generation moderated the relationship between family s tructure and family socioeconomic status (SES) and the hazard of sexual initiation, the hazard of an adolescent birth, and the odds of contraceptive use at first sex among Mexican origin adolescents. I then conducted life history interviews with 21 first a nd second generation Mexican origin young women in Denver, who are currently the same ages of the Add Health participants to understand the process of migration and how migrant families shape gender. I found that two biological parents, grandparents, and older siblings we re protective in reducing risky sexual and reproductive health behaviors. Parental education increased the hazard of sexual initiation in the first and second generat ions but reduced the hazard in the third generation H aving a mother not working as compared to having a professionally employed mother, reduced the hazard of sexual initiation and adolescent birth in the first and second generations and increased the odds of contraceptive use in all

PAGE 5

v three generations Parent adolescent communi cation about sex, particularly encouragement to delay intercourse, was associated with a delay in sexual initiation and a lower risk for adolescent childbearing. Extremely gendered households were associated with an increased risk for sexual initiation and an adolescent birth and these households were of lower SES and had more instability than moderately gendered households. Thus, culture cannot be separated from SES and structural influences. My results contribute to a more comprehensive cultural framework to understand health disparities by incorporating overlooked structural influences, including migration. The form and content of this abstract are approved. I recommend its publication. Approved: Sheana Bull

PAGE 6

vi ACKNOWLEDGEMENTS I am grateful to t he Latino community in Denver for their support and to the women who shared their stories with me. I received funding from Sigma Theta Tau International, Alpha Kappa Chapter at Large and the National Institute of Nursing Research predoctoral fellowship: F3 1NR013821 01A1, and the Health and Behavioral Sciences Department. I am thankful to Sheana Bull, my dissertation advisor and my committee members Patrick Krueger, Fernando Riosmena and Jean Scandlyn, for their time, patience, guidance and support. Abby Fit ch helped me with logistical details and managing my fellowship. The HBS faculty, including Debbi Main and David Tracer provided feedback and support during my first few years in the program. Sarah Horton worked with me on an independent study and helped m e get started with my qualitativ e analysis. Charlene Barrientos Ortiz helped introduce me to the Latino community and facilitated participant recruitment. Servicios de La Raza and Westwood Unidos helped with recruitment. My family and friends listened to m y concerns and alwa ys boosted my confidence when needed. Finally, I am grateful to m y husband, Jason; w ithout his love and encouragement, I could not have completed this dissertation.

PAGE 7

vii TABLE OF CONTENTS CHAPTER I. CONTEXT, BACKGROUND A ND THEORY ................................ ............................. 1 Context ................................ ................................ ................................ ............................ 1 The Latino and Mexican origin populations in the U.S. ................................ ............ 1 Health disparities as a result of oppression ................................ ................................ 3 The inappr opriate framing of adolescent childbearing ................................ ............... 3 Specific Aims and Research Question ................................ ................................ ............ 9 Background ................................ ................................ ................................ ................... 10 Reproductive health in Latino adolescents ................................ ............................... 10 Theory ................................ ................................ ................................ ........................... 12 Segmented assimilation theory ................................ ................................ ................. 12 Theory of gender and power ................................ ................................ ..................... 19 Segmented assimilation theory, theory of gender and power, Mexican origin adolescents ................................ ................................ ................................ ................ 27 Gaps in the literature ................................ ................................ ................................ 28 II. METHODS ................................ ................................ ................................ ................... 29 Conceptual Model and Methodological Foundation ................................ .................... 29 Mixed Methods ................................ ................................ ................................ ............. 31 Phase I: Secondary Data Analysis ................................ ................................ ................ 34 Data ................................ ................................ ................................ ........................... 34 Analysis ................................ ................................ ................................ ..................... 41 Phase II: Qualitative Interviews ................................ ................................ .................... 44 Sample ................................ ................................ ................................ ....................... 44

PAGE 8

viii Data collection ................................ ................................ ................................ .......... 53 Analysis ................................ ................................ ................................ ..................... 55 Phase III: Analysis and Integration of Quantitative and Qualitative Data ................... 60 Analysis ................................ ................................ ................................ ..................... 60 III. REPRODUCTI VE HEALTH OUTCOMES, IMMIGRANT GENERATION, AND FAMILY IN MEXICAN ORIGIN ADOLESCENTS ................................ ..................... 62 Sexual Initiation ................................ ................................ ................................ ............ 63 Background ................................ ................................ ................................ ............... 63 Methods ................................ ................................ ................................ ..................... 68 Res ults ................................ ................................ ................................ ....................... 71 Discussion ................................ ................................ ................................ ................. 79 Conclusion ................................ ................................ ................................ ................ 85 Contraception at Sexual Initiation ................................ ................................ ................ 86 Background ................................ ................................ ................................ ............... 86 Methods ................................ ................................ ................................ ..................... 91 Results ................................ ................................ ................................ ....................... 96 Discussion ................................ ................................ ................................ ............... 104 Conclusion ................................ ................................ ................................ .............. 109 Age at First Birth ................................ ................................ ................................ ........ 110 Background ................................ ................................ ................................ ............. 110 Meth ods ................................ ................................ ................................ ................... 116 Results ................................ ................................ ................................ ..................... 119 Discussion ................................ ................................ ................................ ............... 130 Conclusion ................................ ................................ ................................ .............. 136

PAGE 9

ix Chapter 3 Conclusion ................................ ................................ ................................ .. 136 IV. MIGRATION PATTERNS, FAMILY STRUCTURE, AND GENDER ................. 143 Narratives ................................ ................................ ................................ .................... 143 Jocelyn: Independent migration. Growing up with a single mother who migrated alone at age 16. ................................ ................................ ................................ ....... 143 Isabel: Family stage. Mig rating at age three with her mother and siblings to meet her father in the U.S. ................................ ................................ ................................ ..... 148 Cristina and Arely: Family unit migration. Growing up with two biological parents who migrated to the U.S. together prior to the birth of the ir children. ................... 152 Narrative Summary ................................ ................................ ................................ ..... 157 Migration and family structure ................................ ................................ ............... 157 Family structure and ethnic identity ................................ ................................ ....... 159 Migration and gender ................................ ................................ .............................. 160 Migration and gendered discourses ................................ ................................ ........ 163 Migration, gendered discourses, and reproductive health outcomes ...................... 164 V. FAMILY STRUCTURE AND GENDER ................................ ................................ 166 ................................ ................. 167 Gendered division of labor ................................ ................................ ...................... 167 Gendered division of power ................................ ................................ .................... 168 Descriptions of households ................................ ................................ ..................... 168 ................................ ........................ 173 Teresa: Real versus ideal ................................ ................................ ......................... 173 ................................ .................. 175 Marriage, mot herhood, migration ................................ ................................ ........... 175 Moving away from marianismo and machismo ................................ ...................... 178

PAGE 10

x Ambiguity in discourses ................................ ................................ ......................... 182 Gender, immigrant generation, and reproductive health outcomes ........................ 184 Chapter 5 Conclusion ................................ ................................ ................................ .. 187 VI. THE ROLE OF PARENTS IN THE EDUCATION OF YOUNG WOMEN .......... 189 Sex Education ................................ ................................ ................................ ............. 189 Avoidance and discomfort ................................ ................................ ...................... 191 Virginity ................................ ................................ ................................ .................. 196 Cudate ................................ ................................ ................................ .................... 201 Vulnerability, responsibility, and first sexual experiences ................................ ..... 202 Contraception and first sex ................................ ................................ ..................... 206 Academic Education ................................ ................................ ................................ ... 207 Academic achievement of Mexican and Latino adolescents ................................ .. 209 Parent barriers to promoting academic success ................................ ...................... 211 Navigating two cultures ................................ ................................ .......................... 214 Messages of sacrifice ................................ ................................ .............................. 217 ................................ ............................... 220 Immigrant generation and education ................................ ................................ ...... 223 Gender and education ................................ ................................ .............................. 224 Education as a potential mediator ................................ ................................ ........... 225 Immigrant Generation, Adolescent Births, and Gendered Households ...................... 22 6 Differences by immigrant generation ................................ ................................ ..... 227 Differences by adolescent birth ................................ ................................ .............. 229 Differences by gendered households ................................ ................................ ...... 231

PAGE 11

xi Conclusion ................................ ................................ ................................ .............. 232 VI I MIXED METHODS ................................ ................................ ................................ 234 Integrat ing Quantitative and Qualitative Results ................................ ........................ 234 Similarity of the quantitative and qualitative samples ................................ ............ 236 Mixed methods limitations ................................ ................................ ..................... 240 tive Health Outcomes ................................ ...... 241 Background ................................ ................................ ................................ ............. 241 Methods ................................ ................................ ................................ ................... 243 Results ................................ ................................ ................................ ..................... 248 Discussion ................................ ................................ ................................ ............... 253 Conclusion ................................ ................................ ................................ .............. 263 Mediators of Parental Education ................................ ................................ ................. 264 Background ................................ ................................ ................................ ............. 264 Methods ................................ ................................ ................................ ................... 265 Results ................................ ................................ ................................ ..................... 268 Discussion ................................ ................................ ................................ ............... 273 Conclusion ................................ ................................ ................................ .............. 278 Grandparents, Siblings, and Reproductive Health Behaviors ................................ ..... 278 Grandparents ................................ ................................ ................................ ........... 279 Older siblings ................................ ................................ ................................ .......... 291 VIII. CONCLUSION ................................ ................................ ................................ ...... 304 Migration and Family ................................ ................................ ................................ 305 Migration ................................ ................................ ................................ ................. 305

PAGE 12

xii Family SES ................................ ................................ ................................ ............. 306 Family structure ................................ ................................ ................................ ...... 309 Family, Gender and Academics ................................ ................................ .................. 311 Gendered discourses ................................ ................................ ............................... 311 How parents educate their daughters ................................ ................................ ...... 314 Immigrant generation ................................ ................................ .............................. 319 Limitations ................................ ................................ ................................ .............. 328 Implications for future research ................................ ................................ .............. 329 Implications for practice and policy ................................ ................................ ........ 330 Measuring assimilation and acculturation ................................ ................................ .. 331 Strengths of SAT ................................ ................................ ................................ ..... 332 Weaknesses of SAT ................................ ................................ ................................ 333 Acculturation ................................ ................................ ................................ ........... 337 Reframing Latina Adolescent Childbearing ................................ ............................... 339 REFERENCES ................................ ................................ ................................ ............... 345 APPENDIX ................................ ................................ ................................ ..................... 367 A. APPENDIX FOR CHAPTER 2 ................................ ................................ ............. 367 B. APPENDIX FOR CHAPTER 3 ................................ ................................ ............. 383

PAGE 13

xiii LIST OF TABLES TABLE 1.1: Research Aims ................................ ................................ ................................ ............. 9 1.2: Adolescent R eproductive H ealth B ehaviors by N ativity, I mmigrant G eneration or A cculturation ................................ ................................ ................................ ..................... 11 2.1: Research Aims ................................ ................................ ................................ ........... 30 2.2: Mixing of Methods ................................ ................................ ................................ .... 33 2.3: Demographics of Qualitative Participants ................................ ................................ 47 2 .4: Interview Topics ................................ ................................ ................................ ........ 55 2.5: Qualitative Analysis Steps ................................ ................................ ......................... 57 2.6: Qualitative Analysis Chart ................................ ................................ ......................... 58 3.1: Independent Variables ................................ ................................ ............................... 69 3.2: Weighted M eans ( S tandard D eviations) and P roportions of S elect V ariables .......... 71 3.3: Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and F amily S tructure ................................ ................................ ................................ ................................ ........... 72 3.4: Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and F amily S tructur e ................................ ................................ ................................ ................................ ........... 74 3.5: Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and P arental E ducation ................................ ................................ ................................ .......................... 75 3.6: Hazard R atios (95% C onfi dence I ntervals) for S exual initiation and P arental employment ................................ ................................ ................................ ....................... 77 3.7: Summary of Family Structure and the Hazard of Sexual Initiation .......................... 81 3.8: Summary of Parental Education and the Hazard of Sexual Initiation ....................... 82 3.9: Summary of Parental Employment and the Hazard of Sexual Initiation ................... 83 3.10: Independent V ariables ................................ ................................ ............................. 92 3.11: Weighted Means (Standard Deviations) and Proportions of Select Variables ........ 97

PAGE 14

xiv 3.12: Odds Ratios (95% C onfidence I ntervals) for Contraceptive Use at First Sex and Family Structure ................................ ................................ ................................ ................ 98 3.13: Odds Ratios (95% C onfidence I ntervals) for Contraceptive Use at Fi rst Sex and Family Structure ................................ ................................ ................................ .............. 100 3.14: Odds Ratios (95% C onfidence I ntervals) for Contraceptive Use at First Sex and Parental Education ................................ ................................ ................................ .......... 101 3.15: Odds Ratios (95% C onfidence I ntervals) for Contraceptive Use at First Sex and Parental Employment ................................ ................................ ................................ ...... 103 3.16: Summary of Family Structure and the Odds of Contraceptive Use at F irst Sex ... 107 3.17: Summary of Parental Education and the Odds of Contraceptive Use at First Sex 107 3.18: Summary of Parental Em ployment and the Odds of Contraceptive Use at First Sex ................................ ................................ ................................ ................................ ......... 108 3.19: Independent V ariables ................................ ................................ ........................... 117 3.20: Weighted Means (Standard Devia tions) and Proportions of Select Variables ...... 120 3.21: Hazard R atios (95% C onfidence I ntervals) for Adolescent B irth and Family Structure ................................ ................................ ................................ .......................... 121 3.22: Hazard Ratios (95% C onfidence I ntervals) for Adolescent Birth and Family Structure ................................ ................................ ................................ .......................... 122 3.23: Hazard R atios (95% C onfidence I ntervals) for Adolescent Birth and Parental Educat ion ................................ ................................ ................................ ........................ 124 3.24: Hazard R atios (95% C onfidence I ntervals) for Adolescent Birth and Parental Employment ................................ ................................ ................................ .................... 126 3.25: Hazard of an Adole scent Birth Before and After Controlling for Age at First Sex ................................ ................................ ................................ ................................ ......... 129 3.26: Summary of Family Structure and the Hazard of an Adolescent Birth ................. 132 3.27: Summary of Parental Education and the Hazard of an Adolescent Birth .............. 133 3.28: Summary of Parental Employment and the Hazard of an Adolescent Birth ......... 134 3.29: Variable and Interaction Influence on Outcomes ................................ .................. 138

PAGE 15

xv 3.30: Variables, T heir I nteraction with I mmigrant G eneration, and S upport for H ypotheses ................................ ................................ ................................ ...................... 139 6.1: Migration, Family S tructure, Sex and A cademic E ducation, and Reproductive Outcomes by Immigrant Generation ................................ ................................ ............... 227 6.2: Migrati on, Family S tructure, Sex and A cademic E ducation, and Reproductive Outcomes by Adolescent Birth ................................ ................................ ....................... 230 6.3: Migration, Family S tructure, Sex and A cademic E ducation, and Reproductive Outcomes by G endered Household ................................ ................................ ................ 232 7.1: Descriptive Means and Proportions of Qualitative and Quantitative Participants .. 237 7.2 : Independent V ariables ................................ ................................ ............................. 244 7.3: Hazard Ratios for Sexual Initiation, Maternal Employment, and Maternal Characteristics ................................ ................................ ................................ ................. 249 7.4: Hazard Ratios for Adolescent Birth, Maternal Employment, and Maternal Characteristics ................................ ................................ ................................ ................. 252 7.5: Independent Variables ................................ ................................ ............................. 266 7.6: Hazard Ratios fo Characteristics ................................ ................................ ................................ ................. 269 7.7: ................................ ................................ ................................ ................................ ......... 272 7.8: The Presence of an Grandparent in the Home and Reproductive Health Outcomes 281 7.9: The Presence of an Older Sibling in the Home and Reproductive Health Outcomes ................................ ................................ ................................ ................................ ......... 293 8.1: Strengths and Weaknesses of SAT to Understand Reproductive Health Behaviors 335

PAGE 16

xvi LIST OF FIGURES FIGURE 1.1: Segmented Assimilation Theory ................................ ................................ ................ 15 1.2: The Theory of Gender and Power ................................ ................................ .............. 20 2.1: Concep tua l Model ................................ ................................ ................................ ...... 29 2.2: Mixed Methods Design Based on Teddlie and Tashakkori ................................ ....... 32 2.3: Conceptual Model for Aim 1 & Aim 2 ................................ ................................ ...... 34 2.4: Map of Mexico ................................ ................................ ................................ ........... 51 3.1: Conceptual Model for Aims 1 & 2 ................................ ................................ ............ 63 3.2: Hypotheses for Family Structure and Immigrant Generation Interaction ................. 65 3.3: Hypotheses for Family SES and Immigrant Generation Interaction ......................... 67 3.4 Hazard Ratio of Sexual Initiation, Immigration Generation by Mother's Employment ................................ ................................ ................................ ................................ ........... 79 3.5: Hazard Ratio of Sexual Initiation, Immigration Generation by Father's Employment ................................ ................................ ................................ ................................ ........... 79 3.6: Hypotheses for Family Structure Interactions ................................ ........................... 89 3.7: Hypotheses for Family SES Interactions ................................ ................................ ... 90 3.8: Hypotheses for Family Structure and Immigrant Generation Interaction ............... 113 3.9: Hypotheses for Family SES Interactions ................................ ................................ 115 3.10 : Immigrant Generation by Mother's Employment ................................ .................. 128 3.11 : Immigra nt Generation by Father's Employment ................................ .................... 128 ................................ ......................... 243 ployment ................................ ......................... 265 8.1: Conceptual Model ................................ ................................ ................................ .... 305

PAGE 17

1 C HAPTER I CONTEXT, BACKGROUND AND THEORY Nationally, Latino adolescents have the h ighest birth rates of any race/ethnic group and have higher HIV and STI rates than non Latino whites; 1 4 a reflection on structural constraints resulting in social disadvantage. 5 In just five years from now, Latinos will make up 25% of the adolescent population in the U.S. 6 These statistics underscore the importance of research with Latino, and specifically Mexican origin adolescents, in order to reduce reprod uctive health disparities and improve the quality of life for this vulnerable and growing population. This dissertation makes a significant contribution to the study of reproductive health behaviors of Mexican origin adolescents and young adults by providi ng a more comprehensive understanding of how cultural constructions of gender and sexuality can help explain the relationship between immigrant generation, family socioeconomic status (SES), family structure, and reproductive health behaviors. C ontext Th e Latino and Mexican origin populations in the U. S Between 2000 and 2010, the Latino population in the United States grew four times faster than the total population, totaling 16% of the U.S. population in 2010. 7 By 2050, Latinos will make up 29% of the U.S. population. 8 The Latino population is not homogenous. Latinos come from multiple countries, have migrated to the U.S. for different reasons, and have lived in the U.S. for varying lengths of time, including residing in the southwestern U.S. prior to the area becoming part of the U.S. Current public health experts encourage studying country of origin and differences between

PAGE 18

2 immigrant generation and time lived in the U.S. 9 11 The Mexican origin population is of specific interest because they are the largest Latino subgroup at 65.5% of the Latino population in the U.S. 12 Further, Mexican origin Latinos are more economically disadvantaged that many other Latino groups, ranking 11 out of 14 Latino groups on ranking fourth out of 14 on the share of the population without health insurance. 13 In this dissertation, when I refer to Mexican refer to previous research, I use the same term as specified in the particula r study. There are significant sexual and reproductive health disparities in the Latino and Mexican origin adolescent populations. Latino adolescents have higher HIV and STI rates than non Latino whites 2 4 A lthough there is no national epidemiologic data on differences by country of origin, one study found more Puerto Rican adolescents, but not Mexican or Cuban adolescents, reported a prior a sexually transmitted infection than white adolescents. 14 Between 1991 and 2005, non Latina black adolescents experienced the most rapid decline in birth rates dropping below the rate of Latina adolescents by 2005. T his decline has been attributed to improved contraceptive use by adolescents a nd to a lesser extent, delay s in sexual initiation. It is possible that Latinas did not experience this initi al decline because the Latina population, with higher adolescent fertility, w as increasing in size in the U.S. during that time. 15 Following the decline in non Latina black adolescent birth rates, Latina adolescents experienced the most rapid decline in adolescent birth rates since 2006, although they still had the highest adolescent birth rate of any race/ethnic group in 2012. 16,17

PAGE 19

3 R eproductive health outcomes among Latino subgroups and across immigrant generations remains understudied. Among Latinas, Mexican origin adolescents have a higher birth r ate than Puerto Rican and Cuban adolescents (55.5 births per 1,000 15 19 year olds, 43.7, 24.4, respectively) 18 but there is little data on sexual initiation or contracept ive use by Latino subgroup. Most studies have shown that as acculturation or immigrant generation increases, Latino and Mexican origin adolescents have earlier sexual initiation, are more likely to use contracepti on and have a higher risk for pregnancy an d childbirth 10,19 29 The reasons behind these differences, such as socioeconomic factors and cultural factors, have been hyp othesized though rarely studied directly. Health disparities as a result of oppression Patricia Hill Collins argues th at oppression is grounded in an intersection of race, conceptualizing oppres sion by uncovering the connections among race, class and 30(pp36, 29) I argue that sexual and reproductive health disparities of Mexican origin young women are a product of historical, political, economic, and so cial oppression and I use this framework to understand how the intersection of race, class, and gender are intertwined with the processes of migration and acculturation, and ultimately influence reproductive health behaviors. The inappropriate framing of a dolescent childbearing Reproductive health includes behaviors and outcomes such as sexual intercourse, pregnancy, childbirth, abortion, and sexually transmitted infections, as well as their

PAGE 20

4 meanings. 31 I focus this dissertation on adolescent childbearing because of chronic racial/ethnic disparities relat ed to these outcomes. I includ e sexual initiation and contraceptive use because they are two important behaviors directly linked to pregnancy and subsequent childbearing. Although abortion rates are one of the reasons Latina and black adolescents have similar pregnancy rates but different birth rates, 18 abortion is beyond the scope of this study. Adolescent childbearing has received attention in the literature and in public discourse as an explanation for povert y and social disadvantage. Researchers have demonstrated that infants born to adolescent mothers are unhealthier than infants of adult mothers. 32 35 Children born to adolescents mothers are at a higher risk for child abuse and neglect, 33,36 less likely to finish high school, 33,37 are more likely to become adolescent parents themselves, 37 39 and sons of adolescent mothers are more likely to be incarcerated. 40,41 Researchers also have shown that women were less likely to have graduated from high school if they experienced an adolescent birth as compared to women who miscarri ed during adolescence. 42 Totaling the cost of the outcome s above, including incarceration, lost wages due to low educational attainment and increased health care spending for infants and children, in 2010 adolesc ent childbearing cost U.S. taxpayers a n estimated 9.4 billion dollars 43 However, some economic experts suggest the consequences of adolescent childbearing has been overestimated. 5 Through prior research, my own research and five yea rs of clinical practice with young mothers and their children, I firmly believe that adolescent childbearing does not improve life chances for adolescents or their children and that we should support delays in childbearing. However, public discourse focusi ng on the social and economic

PAGE 21

5 consequences of adolescent childbearing and oversimplifying poverty or culture as causes of adolescent childbearing without addressing the structural factors that perpetuate poverty and influence culture, has framed adolescent childbearing as a social problem caused by poor young women of color and has created a perception that fertility control 44,45 For example, the preventing teen and unplanned pregnancy, we can significantly improve other serious social problems includ ing poverty (especially child poverty), child abuse and neglect, father suggests tha pregnancy. 46 Overestimation of the consequences of adolescent childbear ing Empirical evidence suggests there has been overestimation of medical and economic consequences of adolescent childbearing. For example, after controlling for SES the relationships between poor child outcomes, such as low birth weight and low child cog nitive functioning, and adolescent births weaken or disappear. 34 Children born to adolescent mothers were more likely to be abused, neglected, and plac ed in foster care that children born to adult mothers, however, the authors acknowledge that there is no way to know if delaying the birth would result in improved parenting. 36 Further, Kearney a nd Levine 5 assert that adolescent childbearing has margin al direct economic consequences. A dolescent childbearing matters because it is a symptom of socioeconomic inequality

PAGE 22

6 and this inequality is the main reason th e U.S. has a higher adolescent birth rate than other developed countries. Oversimplification of poverty and culture as causes of adolescent childbearing There is a body of literature that acknowledges the bidirectional causality between impoverishment an d adolescent reproductive health disparities and foc uses on social and economic disadvantage as a cause of adolescent births. 34,45,47 51 However, this research does not often address the structural factors related to poverty that perpetuate the public perception that poor women are responsible for their own poverty and the social problems of soc iety. Studies may acknowledge that Latinos are more likely to be poor than non Latino whites 1,52,53 and Mexican origin families are one of the most socially and economically disadvantaged groups 54,55 with Mexican origin adolescents experienci ng some of the lowest academic and vocational outcomes of other Latino subgroups and immigrant subgroups. 56 58 However, reasons for these economic disparities are often overlooked. For example, a thorough literature review of Latino adolescent reproductive health mentioned poverty and structural barriers that increase risk of adolescent births, such as documentation status, however the conclusion paragraph on policy implications stated, l and reproductive health and their access to culturally appropriate, youth friendly health care services, including reproductive health care. 31(p318) While this is a valid recommendation, there was no recommendation of policies to address structural factors, such as improving access to resources for undocumente d families. A conclusion highlighting the growing Latino population and high adolescent

PAGE 23

7 birth rates as a problem without mention of structural factors contributing to disadvantage may be misconstrued by the public. Much of the U.S. public is not aware tha t the U.S. has enacted p olicies that have promoted legal and temporary migration from Mexico, producing social networks that continue migratory flows. 59 61 Nor are they aware that federal and state legislation aimed to discourage Mexico U.S. migration by reducing access to resources, such as health care, food stamps, and funding for higher education through legislation such as 62,63 have only exacerbated health, educational, and socioeconomic disparities am ong the Latino population, specifically the Mexican origin population. 1 Similarly, border and deportation policies have had a number of unintended consequences, such as increased deaths from border crossing 65 67 and se p arating parents from their U.S. born children. 68 T hese barriers should be framed at the forefront of the problem of adolescent childbearing to avoid furthering stigma and blame of the Mexican origin population Additionally, some researchers have shown that after controlling for SES, Latina adolescents co ntinue to have higher birth rates than non Latino black and non Latino white adolescents and go on to suggest that cultural and traditional values are the reasons behind these differences. 31 This leads to cultural stereotypes marianismo and machismo familismo and respeto that suggest that Lat inos encourage adolescent childbearing that contribute to public hypervigilance of Latina fertility. 69,70 A qualitative study on the challenges pregnancy 1 In some states improvements have been made including allowing undocumented Colorado residents to pay in state college tuition. 64

PAGE 24

8 prevention health care providers face when following recommendations for culturally U.S. based values that guide teenage pregnancy preventio n programs and the values of 71 While there are cultural differences, this type of statement suggests that Hispanic values are the cause of adolescent pregnancy when in fact foreign b orn Mexican origin adoles cents and adolescents in Mexico have lower adolesc ent birth rates than U.S. born Mexican adolescents 29,69 Legislation media, and public discourse ha ve often led to a negative categorization and discrimination of Mexicans. Unfortunately, some public health and social science research has been used by a portion of the U.S. public to blam e the Mexican and Latino population for excessive populati on growth and the social and financial deterioration of t he U.S. 72 Farmer, Connors, and Simmon 73 have c ommented that social science, as well as medicine, has focused on individual and cultural factors and deemphasized structural 73(p189) As I have argued, when a social issue is taken takes the focus away from deeper issues. Thus, I assert that research should not be solely reproductive health disparities of Mexican origin young women into a historical, political, economic, cultural, and social context We should also be cautious of how we frame reproductive health disparities. While I did not measure these structural variables directly, I considered how racism, sexism and classism are produced and reproduced through the process of migration and acculturation which are then transmitt ed and acted

PAGE 25

9 out To avoid contributing further to stereotypes, I paid close attention to how I framed conclusions of this work. Further, I hypothesized that socioeconomics and cultural co nstructions of gender and sexuality can help explain differences in reproductive health behaviors by immigrant generation. Specific Aims and Research Question Overarching research question: How do cultural constructions of gender and sexuality help explai n the relationship between immigrant generation, family SES, family structure and reproductive health behavior of Mexican origin adolescents and young adults? Table 1 1 : Research Aims Phase I Ai m Research Question Aim 1: To examine whether the relationship between family structure and reproductive health behaviors in male and female Mexican origin adolescents and young adults varies by immigrant generation. (RQ1a) What is the relationship bet ween immigrant generation and reproductive health behaviors in Mexican origin adolescents and young adults? (RQ1b) Does immigrant generation moderate the relationship between living with two biological parents and reproductive health behaviors in Mexican origin adolescents? (RQ1c) Does immigrant generation moderate the relationship between the presence of a grandparent and reproductive health behaviors in Mexican origin adolescents? (RQ1d) Does immigrant generation moderate the relationship between the number of siblings in the home and reproductive health behaviors in Mexican origin adolescents? (RQ1e) Does immigrant generation moderate the relationship between having an older biological sibling and reproductive health behaviors in Mexican origi n adolescents? Aim 2: To examine whether the relationship between family SES and reproductive health behaviors (age at first sex, contraceptive use first and most recent sex, age at first birth) in male and female Mexican origin adolescents and young ad ults varies by immigrant generation. (RQ2a) Does immigrant generation moderate the relationship between origin adolescents and young adults? (RQ2b) Does immigrant generation moderate th e relationship between origin adolescents and young adults? (RQ2c) Does immigrant generation moderate the relationship between n origin adolescents and young adults? (RQ2d) Does immigrant generation moderate the relationship between origin adolescents and young adults?

PAGE 26

10 Table 1 1 : Research Aims (RQ2e) What family variables (supervision, co mmunication, relationship quality, gendered expectations); sex education variables (parental expectations of sex, parental communication about sex); religion mediate the relationships between family SES variables and reproductive health behaviors? Phase I I Aim 3: To understand how cultural constructions of gender and sexuality derived through family differ across immigrant generations of Mexican origin young women. (RQ3) How do cultural constructions of gender and sexuality derived through the family diff er across immigrant generations of Mexican origin adolescent females? (RQ4) How have the processes of migration and acculturation (consonant, selective, dissonant) impacted Mexican cultural constructions of gender and sexuality? Phase III Overarching research question How do cultural constructions of gender and sexuality help explain the relationship between immigrant generation, family SES, family structure and reproductive health behavior of Mexican origin adolescents/young adults? B ackground Reproductive health in Latino adolescents Prior research has focused on behavioral risk factors for adolescent childbearing including drug or alcohol abuse, 24,74 multiple sexual partners, 24,75 low power in a sexual relationship, 76 and a la ck of knowledge about birth control. 77 79 Individual protective factors include attending school 80,81 and after school or religious activities. 74,80,82 Often due to methodological issues, such as small sample sizes, reproductive health outcomes within Latino subgroups and across immigrant generations remains understudied. Although there is some conflicting evidence, in general studies have shown that as acculturation or immigrant generation increases, Latino and Mexic an origin adolescents have earlier ages at sexual initiation, are more likely to use contracepti on and have a higher risk for pregnancy and childbirth (Table 1.2).

PAGE 27

11 Table 1 2 : Adolescent R eproductive H eal th B ehaviors by Nativity, I mmigrant G eneration or A cculturation Study Survey Sample Results by nativity or generation Results by acculturation Brindis et al. 1995 Public high schools San Jose, CA Native born Latinos and immigrant Latinos Sexual initiatio n risk: Native born Latinos Foreign Ford & Norris 1993 Household sample Detroit Mexican and Puerto Rican Acculturation Scale (Language) Sexual initiation risk: Contraception: Gilliam et al 2007 Convenience s ample West Chicago Foreign born vs native Latinos, male and female Sexual initiation risk: Foreign Sexual initiation risk: Guarini et al. 2011 Add Health Wave I III First, second, third plus generation Latinos, ma le and female Overall sexual risk: Increased immigrant Guilamo Ramos et al. 2005 Add Health Wave I II Mexican, Cuban, PR, male and female, continuous length of time in US and native born Sexual initiation risk: Recent immigrants English at Native born Spanish No relationship between years in U.S. and sexual initiation in early immigrants. Jimenez et al. 2002 P regnancy prevention program Los Angeles Latinas (1) foreign born (2) native born English speakers (3) native b orn Spanish speakers Sexual initiation risk: Foreign Contraception: Foreign born Latinos = Kaplan et al. 2002 F amily planning clinics Los Angeles Latin o/as Acculturation scale : (1) language (2) traditional gender roles Sexual initiat ion risk: Pregnancy risk : Killoren & Deutch, 2013 NLSY97 Foreign born and U.S. born Latinos Overall sexual risk: Foreign Overall sexual risk: Increased acculturation = Landale 2009 Add He alth Wave I III First, second, third plus generations Mexican American First event is marriage: First event is birth: First & third generation =

PAGE 28

12 Table 1 2 : Adolescent R eproductive H eal th B ehaviors by Nativity, I mmigrant G eneration or A cculturation Lee & Hahm 2009 Add Health Wave I III Foreign born, native born Latinos More sexual par tners: STIs: Overall risk: Foreign born English at McDonald e t al. 2009 NLSY97 First, second, third plus generations Latinos (Mex, Cuban, Chicano, PR, other) male and female Sexual initiation ris k: Increased generation Contraception: Childbirth risk : Increased generation = Sexual initiation risk: Contraception: Minnis et al. 2001 F amily planning clinics San Francisco Foreign born, n ative born Latinos Sexual initiation risk: Foreign Contraception: Foreign Pregnancy risk : Native born 15 T heory I have argued that the study of repro ductive health behaviors in Mexican origin young women requires consideration of a historical, political, economic, social, cultural context which inherently creates a very large scope of inquiry. In order to focus my work and create a reasonably sized pro ject while incorporating a socio cultural context, I have integrated segmented assimilation theory and the theory of gender and power to form a theoretical framework. Furthermore, theory is often lacking in public health and medical research despite studie s that show improved outcomes using a theoretical framework. Segmented a ssimilation t heory Acculturation and assimilation have been used interchangeably although they are and th eir economic adaptation and upward mobility into a native U.S. middle class, while acculturation is the acceptance or rejection of particular cultural beliefs, values and

PAGE 29

13 practices of the host culture. 83 Researchers have conceptualized and operationalized both variables so inconsistently and inaccurately, often ignoring structural factors, that some public health experts have encouraged eliminating them as varia bles. 9,84,85 I agree that the definition and measurement of acculturation has been so misconstrued that its worth is in jeopardy. However, we cannot overlook the social, cultural, and economic processes that that influence migration, settlement, and ada ptation. In this dissertation, I conceptualized assimilation as economic incorporation to a particular native U.S. group, including population and location, and acculturation as rejection, acceptance, and negotiation of cultural beliefs and practices of bo th the native and host culture. While not mutually exclusive, acculturation may influence subsequent economic assimilation. I hoped this work would lead to a more comprehensive definition of acculturation. Classic assimilation theorists from the 1920s 30s, such as Parks and Gordon, believed that native language and culture of immigrants were barriers to successful economic assimilation, and only a complete loss of ethnic identity and adoption of an American cultural identify was the way to achieve upward mo bility and economic success. 86 These th eories were based on white European immigrant experiences that did not consider ethnic differences among immigrants, nor did they address structural factors such as political, economic, or social climates of the U.S. Hunt and colleagues specifically argue issues that impact both cultural change and health, such as the practical realities of immigrant life a 84(pp981 2)

PAGE 30

14 Sociologists Alejandro Portes and Min Zhou constructed segmen ted assimilation susceptible to downward mobility and what allows them to bypass or to get out of this 86(p975) Portes and Zhou 87 suggest that immigrants may not assimilate into a mainstream U.S. middle class but into the social class of the specific location in the U.S. where they migrated, often a poor urban area. However, not all immigrants living in impoverished neighborhoods become downwardly mobile and therefore, Portes and Zhou at the group level : Assimilation into white middle class America and subsequent upward mobility Maintenance of strong ethnic identities and biculturism increasing upward mobility Segmented assimilation theory has defin ed three background determinants: (1) human or parental capital/ family socioeconomic status; (2) family structure; and (3) modes of incorporation (government policy, societal reception, and the strength of the co ethnic community) (Figure 1.1). Drawing o n segmented assimilation theory, I focus on the influence of family structure and family SES on reproductive health outcomes in Mexican origin adolescents.

PAGE 31

15 Figure 1 1 : Segmented A ssimilation T heory 88 Family structure and reproductive health in Mexican origin adolescents. Segmented assimilation theory conceptualizes protective family structure as a stable family with two biological parents. Immigrant families have been documented as more likely to be extended or to include non kin than native U.S. families 89 91 and research has shown that immigrant children and children of immigrants are more likely to live in two biological parent families than their native born peers. 89,92 Growing up in a stable two parent family has been linked to upward mobility in second generation immigrants, even despite a low family SES. 57 In fact, immigrant and native children with two biological parents have fared better in life than children from any other family type because p arents tend to have a higher SES, provide more supervision, more encouragement, and hold higher academic expectations for their children 92 Even after controlling for race and family income, living with two biological parents, compared to single or mixed parent families has been associated with decreased risk of sexual intercourse, 80,93 96 increased contra ceptive use in females, 97 and decreased risk for adolescent birth. 82,98 100 D espite co residing families being poorer than non co residing families, 101,102 Monserud and Elder 103 found that in single parent families, co residing grand parents w ere

PAGE 32

16 beneficial to educational attainment. Segmented assimilation theory does not specifically address grandparents, but Perez 104 argues that the presence of grandparents in the household has been shown to buffer harsh effects of poverty and to promote upward mobility in both immigrant and African American U.S. families. Similarly, Bleakley and 105 found that adolescents who sought sexual health information from grandparents were more likely to hold ideals that would delay sexual initiation. Thus, the presence of a grandparent may provide added support and supervision and reinforce traditional cultural values. Although used in segmented assimilation theory and in the literature, traditional cultural values has not been well defined n or consistently measured. Similar to the concept of acculturation, traditional cultural values is a concept often removed from a historical, political, social and economic context and may rely upon and contribute to cultural stereotypes. In this disserta tion, I plan to better understand values and actions of Mexican families in order to examine the usefulne ss of traditional cultural values as previously measured in the literature Evidence has shown that Mexican origin adolescents in immigrant families ha ve more siblings than those from native born families 106 and Mexican origin siblings sp end more time together than non Latino white siblings. 107,108 Coupled with the cultural values of familismo and respeto 107,109 or the importance of family and respect, often encouraged by Latino families, siblings may have a stronger influen ce on reproductive health behavior of adolescents in immigrant families. Although supported in the literature, it is important to note that familismo and respeto are still stereotypes, albeit positive ones, and warrant careful consideration before using th em to describe Latino families. Sibling behavior has often been correlated, particularly if siblings have a close relationship 107 or

PAGE 33

17 when younger siblings model an older siblin 110 Prior research has found that adolescents with an older sibling were more likely to engage in risky behaviors, including sexual initiation, 111,112 more likely to have an adolescent birth if their older sister had an adolescent birth, 38 with younger siblings receiving less parental monitoring, 111,112 Conversely, siblings with close relationships with older siblings were found to have lower sex ual risk behavior, 113 and participants with older siblings may have a decreased risk f or early childbearing due to an older sibling instilling the same protective cultural values as their parents, receiving contraceptive advice from their siblings, or by learning Family SES and reproductive health. Fa mily SES is a key element in segmented assimilation theory. Immigrant families have a lower SES than native families and that predisposes adolescents to downward assimilation. 54,87 However, high parental SES is related to upward mobility in the second generation. 57 Higher family SES has been associated with better academic, vocational, and health outcomes in children and their transition to adulthood, including higher rates of high school and college graduation, professional employment, decreased risk for depression and risky behaviors (substance abuse, STIs, adolescent childbearing etc. ). 31,74,99,114 119 Parents with higher education tend to be employed at higher status jobs and have a higher family income than parents with lower education. 31 Higher SES provides access to resources for children; families can afford to live in safer neighborhoods with better schools, have better health care, and can afford out o f school activities. 116

PAGE 34

18 The relationship between specific components of family SES and adolescent parental SES index is used 22,26,120 Parental education has been shown to decrease risk for sexual intercourse, 93,118 increase use of bir th control, 97,121 and decrease risk for an adolescent birth. 115 However, Aneshensel 93 found no relationship between risk for sexual education was statistically significant. I address this gap by studying education and employment type of both mother and father. Patterns of assimilation and acculturation. According to segmented assimilati on theory, mobility of the second generation depends on background determinants (family structure, family SES, and modes of incorporation) as well as external barriers including racism (skin color), bifurcated labor markets (economic ladder), and an urban ). Portes, Fernandez Kelly, and Haller 57 believe that the background factors determine the type of acculturation of the adolescent and family (Figure 1.1). Consonant acculturation occurs when high er SES families acculturate to the host society together; selective acculturation occurs when families and their children choose what aspects of their ethnic culture they wish to retain while embracing other aspects of U.S. or local community culture such as being bilingual and bicultural; and dissonant acculturation occurs when children reject protective aspects of their ethnic culture that leads to conflict within the family. Selective acculturation has been associated with decreased risky behaviors and u pward socioeconomic mobility, 57,87,122 although this has not yet been tested with sexual or reproductive health outcomes.

PAGE 35

19 This theoretical framework is advantageous in the study of immigrants because it emphasizes the relationship between indiv idual (race, ethnicity), family (SES and structure), and structural factors (policy, segregated schools) and provides testable questions, many of which have been supported in the literature. 122 125 I addressed three limitations of segmented assimilation theory in this dissertation. First, the theory focuses on the second generation alone. I studied first, second, and third generations because first and third generations are also nec essary to und erstand the full process of acculturation and economic assimilation. Second, the theory does not address gender and how background factors, acculturation, or expected outcomes may differ by gender 9,126 As a consequence, I have integrated the theory of gender and power with segmented assimilation theory in order to understand the role of gender in the process of acculturation and subsequent reproductive behaviors in adolescents and yo ung adults. Finally, the theory does not address how there can be both consequences of acculturation, such as earlier age of first sex and a greater number of sexual partners of youth, and beneficial effects such as increased condom use. 19,24,31 Integrating the theory of gender and power, which seeks to understand how women have an increased sexual health risk du e to gendered inequalities, is helpful to examine benefits and consequences of acculturation Theory of gender and power Connell 127 developed the theory of gender and power in order to organize and explain the structures that produce gendered inequalities, usually benefiting men. This theory examines gender through three main structures: (1) the sexual division of labor, economic differences in the workplace and the home; (2) the sexual division of power,

PAGE 36

20 relationship and institutional differences in power; and (3) cathexis. Wingood and DiClemente 128 provide a helpful definition of cathexis: At the societal level, this struc ture dictates appropriate sexual behavior for women and is characterized by the emotional and sexual attachments that women have with men. This structure constrains the expectations that society has about women with regard to their sexuality and, as a cons equence, shapes our perceptions of ourselves and others and limits our experiences of reality (p. 544). This theory has been adapted to explain how these structures place women at risk for HIV 128 which is also useful for reproductive health risk in general (Figure 1.2). All three structu res are intertwined and necessary to understand gendered inequalities. Relationship power between the parents of the adolescent is pertinent to this study but power in the Figure 1 2 : The Theory of Gender and Power 128 Cultural constructions of gender and sexuality. It is important to understand cultural constructions of gender and sexuality when studying reproductive health behaviors in adolescents because identity shapes action and g ender identity will play a powerful role in reproductive health behavior. 129 a product of acting social beings trying to make sense of the world in which they find 130(p130) While culture can influence and even restrain human behavior, it is

PAGE 37

21 not detached from individuals and their practices. Culture is dynamic, fluid, shaped many factors, and is produced and repr oduced by individuals. 130 Gender is no longer believed to be purely biological, but historically, socially, and culturally constructed into male and female identities. 131,132 I acknowledg e that gender identity is not dichotomous and may include transgender male to female, female to male and non confirming gender, however this is beyond the scope of this work. Sexuality, defined as sexual thoughts, experiences, and acts both of individuals and a larger society, is also socially constructed and fluid. 133 Gender and sexuality are not mutu ally exclusive historical, and structural context. 133(p876) I conceptualized cultural constructions of gender and sexuality as shared meanings and values, such as gender roles and expectations of gender and sexuality, of family, pe ers, and society from which individuals create their own personal meanings and values. This process is also fluid and circular; while shared cultural meanings influence those of the individual, through behavior and self expression, the individual then infl uences shared cultural meanings. Studying cultural constructions of gender and sexuality provided a more comprehensive understanding of why women make particular reproductive health decisions. In this dissertation, I did not examine political economic for ces in themselves. However, gender and sexuality are produced and reproduced through families and political economic changes provide pressures on the family. Hondagneu Sotelo 134 states in her ethnography of gender in migration and settlement of Mexican origin adults in California, While larger forces provide pressures, resources, and opportunities, it is patriarchal gender relations in family and community, together with generation, class, and

PAGE 38

22 culture, that determ ine how those pressures and opportunities lead to particular patterns of migration and settlement. Patriarchal gender relations mediate between broad economic and political factors and actual patterns of migration. 134(pxxiv) This statement is also applicable to patterns of acculturation and subsequent economic assimilation. Children and adolescents learn and negotiate cultural constructions of gender and sexuality through their families. However, both fami lies and immigrant generation. Gendered division of labor and power: U.S. and Mexico. When studying the processes of migration and acculturation, we must first understand d ifferences between the sending and receiving countries. The U.S. may have more gender equity than Mexico, the U.S. 135 and women are still largely responsible for household activities and childcare. 136 In Mexico, rapid political and economic changes have begun to change gender roles that once include men as the f inancial providers and women the keeper of the household. 137 Hirsch 138 and Gutmann 137 found changing gender and identities and roles in Mexico were related to a demographic transition as well as cultural movements such as feminist and gay rights movements. Gendered division of labor and power: Migration and adaptation. It is often assumed that South North migration improves the lives of women because they become immersed in a more gender equal society. Hondagneuo Sotelo 134 found that when men migrated to the U.S. prior to 1965 versus after 1965, there was more gender equity in the household division of labor and decision making because men often migrated first and learned to domestically care for themselves without a woman present. Similarly,

PAGE 39

23 migration may weaken male dominance through greater opportunities for women to work outside of the home. 138 Yet migration may also result in men reinforcing masculinity and dominance in the home through their desire to compensate for low status employment in the U.S. 139,140 Similarly, women may be more dependent on men if they are undocumented and exploitation and unjust treatment of immigrant women in the U.S. labor market may incre 140 My research co ntributes to a better understanding of how the gendered division of labor and power change s with migration and acculturation. Gendered division of labor and power: Adolescents. the home has been associated with increased power and dec ision making in the household and can encourage education and employment as alternatives to early childbearing, particularly for girls. However, parental supervision has been associated with improved adolescent educational and health outcomes and mothers w orking outside the home may have less time to spend supervising children. 82,141 Qualitative and quantitative research has found that girls and boys received differential treatment in Latino and Mexican origin families, with sisters, especially older sisters, having more domestic responsibilities, fewer privileges, and more parental control than their brothers, particularly in families with a proclivity toward Mexican culture rather than Anglo culture. 108,142 145 Less is known about the impact of differential treatment on adolescent health outcomes, although McHale 108 had a more negative impact on youth well being when familism was low suggesting the meaning of differential treatment may be more influential on outcomes than the actual

PAGE 40

24 treatment itself. A strict division of labor in the home that burdens women and girls may reinforce gender roles and expectations of women caring for the home and family and may subsequently, even if not intended, devalue the importance of education and independence 142 perhaps resulting in earlier childbearing. The gendered division of labor power in the household, including how decisions are made and how family members are treated may influence how girls construct their worth and how they make decisions. For example, girls who witness intimate partner violence (IPV) between their paren ts are twice as likely to experience IPV as adults 146 I n homes where their mothers have less decision making power than fathers, girls may be less likely to advocate for power in their own relationships. Lower relationship power has been associated with increased risk of sexually transmitted infections and unintended pregnancy. 128 My research contributes to a better understanding of the relationship between the gendered division of labor and power in the home and adolescen t reproductive behaviors. Cathexis: U.S. and Mexico. Cathexis encompasses both the societal norms and affective relationships that exist at the societal and familial levels that influence and Latina reproductive macho found in the literature. 84,137 Gutmann 137 has argued that the continued use of machismo aggressive, controlling male dominating behaviors, and marianismo subservient and docile female behaviors akin to the virgin Mary, are not only highly nuanced, but can promote a negative racialization of Latinos and Mexicans an d reproduce gender inequities in the U.S. Frank and Heuveline 69 also argue that researchers should not

PAGE 41

25 assume that Mexican Mexican origin women have higher fertility rates than Mexican women in Mexico and foreign born Mexican origin women living in the U.S. for more than six years. the U.S. than in Mexico because is sex ubiquitous in U.S. media and society. However, U.S. parents are often uncomfortable talkin g about sex with their children and sex is still considered a potentially offensive word 147,148 Thus, cathexis is ambiguous in both cou ntries. Cathexis: Migration and adaptation. Hirsch 138 found in Mexico childbearing the construction of female gender ident ities and what it means to be a woman Delays in childbearing and declines in family size in first generation Mexican origin women in the U.S. revealed a change in the cultural construction of gender through the process of migration. Constructions of sexua lity also change with migration. A U.S. Mexico transnational survey by Parrado and Flippen 149 found that women who migrated from Mexico had more egalitarian beliefs about sexuality (i.e. equal responsibility for pregnancy and STI prevention, less acceptance of extramarital affairs, biological need for sex as a myth), while men who migrated have m ore traditional beliefs (i.e. women have responsibility for pregnancy and STI prevention, more acceptance of extramarital affairs for men only, and biological need for sex as a truth). These beliefs will influence how families construct gender and sexualit y These beliefs are particularly important in adolescents who are in the process of choosing what constructions to embrace, reject, or negotiate as they form their own identities. In the transition to adulthood, native born

PAGE 42

26 U.S. adolescents struggle to ch an Immigrant and second generation adolescents must make sense of an adult and adolescent world, but also between native U.S. world and a country of origin world 1 50 potentially resulting in a transition to adulthood fraught with even more conflict and turmoil than native born adolescents. Cathexis: Adolescents. Female virginity is an example of an emotional attachment to sex and a social norm. Many Latina adolesce nts are strongly encouraged by their family to remain a virgin until marriage 109,143,151 154 and many of them embrace this value. 109,153 Garcia 151 found that the majority of second generation Latina adolescents famili acculturation is not completely linear, the participants were quick to point out that while they constructed virginity differently than their mothers, they verbalized that they did not share constructions of virginity of white native U.S. adolescents. Similarly, motherhood is an important goal for many Latina adolescents 79,109,143 but contrary to cultural stereotypes, these adolescents do not want to become pregnant as adolescents; they desire to finish high school and attend college first. 109,143,151,153 In addition to encouraging virginity and motherhoo d, immigrant f amilies from all over the world often have stricter controls on adolescent daughters. There is more supervision and more rules and regulations for daughters, especially in the area of sexuality and dating, 142 144,150 likely influencing reproductive health behaviors. Despite gender roles and expectations in many Latino families, Latin a and Mexican origin adolescents do not view themselves as victims in the face of gendered expectations or

PAGE 43

27 inequalities. They are aware of the sexual division of labor, power, and cathexis, as well as ambiguities in both Mexican and U.S. culture, and they seek to find ways to negotiate gendered roles and expectations of multiple cultures. 109,143,151,153 My dissertation adds to current research on gender and sexuality in Mexican origin adolescents and links these constructions with reproductive health outcomes, which has not yet been done. Segmented assimilation theory, theo ry of gender and power, Mexican origin adolescents Integrating segmented assimilation theory and the theory of gender and power, the family plays a large role in the acceptance, rejection, or negotiation of constructions of gender and sexuality from the co untry of origin or U.S native community. While traditional cultural values have been measured qualitatively or less often, quantitatively through surveys, 20,109,143,145,153,154 they are usually examined in relationship to the health outcome, not in relationship to independent variables such as family structure and SES. Howev er, we can postulate from previous research that Mexican origin immigrant families with two married parents 57 and grandparents, 104,105 and with more siblings, 106 in less educated and underemployed migrant families, may ascribe to more traditional gender roles and expectations. 139,140 These traditional gender roles and expectations may then decrease risk for sexual initiation and unmarried adolescent childbearing, and increase risk for non contracept ive use in first and second generation families. However, as often seen in the immigrant paradox literature, as time in the U.S. increases, families may replace traditional cultural values and behaviors with the less protective values and behaviors of disa dvantaged U.S. communities. Thus, third generation adolescents may be vulnerable to risky sexual behaviors without protective cultural values to buffer the

PAGE 44

28 negative influence of poverty, increasing risk for downward economic assimilation and adverse health outcomes. 9,31,87 Gaps in the literature inequities in the U.S. setting will enhance our understanding of both the theoretical link between gender and migration and the role of family and gender dynamics i n migrant 140(p606) My dissertation el ucidates the link between gender, migration, and family dynamics, and reproductive health outcomes by addressing three main gaps in the literature. First, I addressed the heterogeneity of Latinos by studying adolescent reproductive health outcomes in multi ple immigrant generations of Mexican origin 84(p978) Second, I tested and extended theory by integrating the theory of gender of power with segmented assimilation theory to examine how family influences on reproductive health behavior by immigrant generation work through cultural constru ctions of gender and sexuality. Finally, instead of hypothesizing that cultural values are the reasons for differences in reproductive health behaviors by immigrant generation, I was able to link qualitative data on how the process of acculturation (conson ant, selective, dissonant) and constructions of gender and sexuality influence reproductive health behaviors.

PAGE 45

29 CH A P TER II METHODS C onceptual Model and Methodological Foundation inal Study (CLIS) 1992 2002 examined the social context, self identity, mental health, academic, and vocational outcomes of first 1.5 (immigration prior to age 12), and second generation adolescents. Participants and their families who were economically di sadvantaged at Wave I of the survey but were upwardly mobile by the final wave were interviewed in order to understand how the types of acculturation (dissonant, consonant, selective) influenced economic mobility. 57,155 This design served as the methodological foundation for my own work. This study tested my adaptation of the segmented assimilation conceptual model. 57 Figure 2 1 : Conceptual Model ( my additions in red ) 88

PAGE 46

30 Based on this conceptual model and methodological foundation, I used a cross sectional mixed quantitative and qualitative methods design. This design involved three phases. In p hase I I analyz ed data from the nationally representative National Longitudinal Study of Adolescent Health (Add Health) to understand the relationship between family structure, family SES and reproductive health behaviors across immigrant generations. In phase 2 I conduc ted qualitative life history interviews with 21 first and second generation Mexican origin women, 27 39 years old, recruited in the Denver metro area to understand how gender and sexuality are constructed in Mexican origin families. In phase 3 I integrated both methods to conduct a mixed methods analysis and explore how cultural constructions of gender and sexuality mediate the relationship between family structure and SES and reproductive health behaviors across immigrant generations of Mexican origin youn g women (Table 2.1). Although these are two different samples, as I explain in more detail in the mixed methods section, the qualitative participants are currently the same ages of the Add Health participants so they theoretically could have participated i n Add Health. This allows me to draw some inferences between the two samples. Table 2 1 : Research A ims Phase I Aim Research Question Aim 1: To examine whether the relationship between family structur e and reproductive health behaviors in male and female Mexican origin adolescents and young adults varies by immigrant generation. (RQ1a) What is the relationship between immigrant generation and reproductive health behaviors in Mexican origin adolescent s and young adults? (RQ1b) Does immigrant generation moderate the relationship between living with two biological parents and reproductive health behaviors in Mexican origin adolescents? (RQ1c) Does immigrant generation moderate the relationship betwee n the presence of a grandparent and reproductive health behaviors in Mexican origin adolescents? (RQ1d) Does immigrant generation moderate the relationship between the number of siblings in the home and reproductive health behaviors in Mexican origin a dolescents?

PAGE 47

31 Table 2 1 : Research A ims (RQ1e) Does immigrant generation moderate the relationship between having an older biological sibling and reproductive health behaviors in Mexican origin adolescents? Aim 2: To examine whether the relationship between family SES and repr oductive health behaviors (age at first sex, contraceptive use first and most recent sex, age at first birth) in male and female Mexican origin adolescents and young adults varies by immigrant generation. (RQ2a) Does immigrant generation moderate the r elationship between origin adolescents and young adults? (RQ2b) Does immigrant generation moderate the relationship between origin adolescents and young adults? (RQ2c) Does immigrant generation moderate the relationship between origin adolescents and young adults? (RQ2d) Does immigrant generation moderate the r elationship between origin adolescents and young adults? (RQ2e) What family variables (supervision, communication, relationship quality, gendered expectations); sex education variables (pare ntal expectations of sex, parental communication about sex); religion mediate the relationships between family SES variables and reproductive health behaviors? Phase II Aim 3: To understand how cultural constructions of gender and sexuality derived throu gh family differ across immigrant generations of Mexican origin young women. (RQ3) How do cultural constructions of gender and sexuality derived through the family differ across immigrant generations of Mexican origin adolescent females? (RQ4) How have the processes of migration and acculturation (consonant, selective, dissonant) impacted Mexican cultural constructions of gender and sexuality? Phase III Overarching research question How do cultural constructions of gender and sexuali ty help explain the relationship between immigrant generation, family SES, family structure and reproductive health behavior of Mexican origin adolescents/young adults? M ixed Methods According to the Journal of Mixed Methods Research mixed methods is d efined findings, and draws inferences using both qualitative and quantitative approaches or 156 Using two different typologies, 157,158 I created a typology for this dissertation: A partially mixed, partially sequential and partially concurrent, equal sta tus mixed design (Figure 2.2).

PAGE 48

32 Figure 2 2 : Mixed M ethods D esign B ased on Teddlie and Tashakkori 158 Teddlie and Tashakkori 158 believe that sequential designs can be ideal for dissertations because they allow the student to strengthen current knowledge in the literature in one phase and take an explorator y approach in the other. I built my quantitative analysis based on previous quantitative analyses in the literature while my qualitative piece was more exploratory. I began cleaning, coding and preliminarily analyzing Add Health data, based on a priori the mes from theory and literature during the summer of 2012 while awaiting COMIRB approval for qualitative interviews. Once interviews began in March of 2013, the iterative process shown for the data collection and analysis phases in Table 2.2 began. Intervie ws were completed April 2014 and analysis for both phases continued concurrently through December 2014.

PAGE 49

33 Table 2 2 : Mixing of M ethods QUANTITATIVE QUALITATIVE Conceptualization and design Add Health data s et chosen. Participants 24 32 years old in 2008, 29 37 in 2013. Mexican origin population 1800 with three distinguishable generations. Inclusion criteria based on Add Health. Participants 27 39 years old in 2013, within a few years in age of the Add Heal th participants. Latest age of migration was to U.S. was 18. Inclusion criteria included, first generation women who migrated prior to adolescence, or age 19. Second generation was defined as native born to Mexican born parents or Mexican born who mig rated prior to the age of six based on literature and language/education Second generation women were defined similarly. Immigrant generation models tested using different ages of migration. All foreign born participants, regardless of age of migratio n were considered first generation based on qualitative findings and supported in the literature. Interviews of Mexican born women who migrated prior to the age of six were not at all similar to native born women and could not be classified as such. P reliminary Analysis (Quant) and Data Collection (Qual) Living with two biological parents significantly associated with later age at first sex and later first birth Tried to get more in depth information on childhood, family transitions (divorce, death ) and parental relationship Parental education and employment associated with reproductive health outcomes differently by immigrant generation Asked more specifically about potential reasons education and employment work differently, supervision, etc. Parental supervision, communication, support, expectations for sex and school, variables coded and analyzed as mediators Relationship with parents, communication, parental expectations emerged in interviews Added in parental expectations and parental i nvolvement in school as mediators The value of education and role parents played in education emerging as influential factor on reproductive health Added in sex education variables, school, home, etc. as mediators Differences in how participant learn ed about sex and parental discussions about sex emerged in interviews Birth order and sibling variables added to models Birth order, siblings, and gender of siblings emerging as important factor in learning about sex, reproductive and educational outcom es Analysis Four qualitative patterns were used to select variables tested as mediators in family structure and family SES models Emergence of four patterns that influenced reproductive outcomes: educational expectations, parental supervision and supp ort, parental sexual expectations and education, gendered division of labor, power and cathexis Rich qualitative data was used to provide context and support hypotheses on how quantitative relationships work Rich description of family processes, specifi cally how gender and sexuality were produced and reproduced and negotiated in families

PAGE 50

34 Mixed methods are not without limitations and the biggest methodological risk in this design was the use of two different samples. Onwuegbuzie and Johnson 159(p560) argue that unless the qualitat ive and quantitative data is from the same sample, the validity of inferences between the two samples and methods is in jeopardy though they suggest that, sample, the pr sample, 27 41 year old Mexican origin females, who migrated prior to the age of 19, provided similarity in the quantitative and qualitative sample because the qualitative local partic ipants now could have theoretically been participants in Add Health. P hase I: Secondary D ata A nalysis Figure 2 3 : Conceptual Model for Aim 1 & Aim 2 Data I used data from all four waves (1994 2008) of The National Longitudinal Study of Adolescent Health (Add Health) from the University of North Carolina. This is the largest nationally representative longitudinal study of adolescents focusing on adolescent health and consists of both in home and in school s urvey data. The sampling frame included 26,666 high schools categorized by size, school type (public, private, Catholic), census region (Northeast, South, Midwest, West), percent white and urbanization (urban, rural, suburban), for a total of >100,000 adol escents and parents. Add Health contains an in home survey, in school survey, parental surveys, and school administrator surveys in Family SES Family Structure Immigrant generation First, se cond, third BEHAVIOR First sex sex, BC use, first birth IV DV Moderators

PAGE 51

35 Wave I. The in home survey for Wave I, 1994 1995 included 20,745 7 12th graders. Wave II in 1996, followed up with 14,738 o f the Wave I participants, now in grades 8 12 (those in grade 12 in Wave I were dropped from Wave II but re entered in Wave III). Wave III, 2001 2003, includes 13,690 participants from Wave I in addition to 1,507 opposite sex partners of the original part icipants, all 18 26 years old. Wave IV 2007 2008, includes 15,701 participants now 24 32 years old (52 respondents were 33 34 years old). All interview questionnaires were completed on laptops and the participants completed more sensitive questions using t he laptop and headphones. 160 Th ese data were appropriate for this project in three ways. First, Add Health provided detailed information on risk factors including sexual behavio r. Second, the sample was large and includes Mexican origin adolescents with information on place of examine differences between first, second, and third generation immigr ants, as well as age at immigration as a continuous variable. Third, the study was longitudinal allowing me to compare timing of reproductive health behaviors across groups with event history survival analysis. Study Sample I restricted the sample to Mex ican origin adolescents. I classified participants as Mexican origin if they considered themselves Mexican or Chicano. Although descendants of any Latin American country may classify themselves as Chicano, the definition of Chicano generally refers to pers ons of Mexican descent and who associate themselves with a political/social movement. 161 In addition to participants who self repor ted Mexican origin, I included participants who were born in Mexico or

PAGE 52

36 had a parent born in Mexico even if they did not self report Mexican origin because self identification is not the outcome of my analysis. Item 1: Are you of Latino or Hispanic origin? (yes or no) Item 2: What is your Hispanic or Latino background? May choose more than one Mexican, Mexican American, Chicano will all be coded as Mexican. If the participant choses more than one, they will still be included in the analysis because theore tically, they will have some socialization to Mexican culture if at least one parent is from Mexico. Item 3: Which category best describes your racial background? Power analysis. Because this was a secondary data analysis and the sample size predetermin ed, I calculated the minimum coefficient, or effect size, for the outcomes variables (sexual initiation and adolescent birth) to detect with a power of .80 and a two tailed alpha of .05. I used the proportion of Mexican origin adolescents in Add Health who had sex and/or a first birth by Wave III from Fomby, et al., 120 which is modest because a larger proportion will have experienced an event by Wave IV. There are 1,702 participants who self identified as Mexican origin and I illustrate the range in the minimum detectable coefficient size among Mexican origin adolescents by showing the effect sizes for the least common and most common outcomes. The rarest outcome in the ages examined is giving birth; I can detect a coefficient of 0.2567 and a hazard ratio of 0.7504. The most common outcome is having sexual intercourse; I can detect a coeffici ent of 0.1538 and a hazard ratio of 0.8419. A lthough I hope to find a larger effect size, a ccording to Agresti and Finlay 162 the ability to detect an effect size below 0.3 is Variables. I created a codebook of variables (See appendix). The process of cleaning and coding the data was quite extensive starting in July 2012 and ending in

PAGE 53

37 December 2013, because there are four waves of the survey and quest ions are asked differently depending on the wave. Dependent or outcome variables Item 1: Age at first sex: (continuous) 1. Have you ever had sexual intercourse? 2. In what month and year did you have sexual intercourse for the very first time? Month and year were only asked in Wave I & II. Age in single years was asked in Wave III & IV. Age at first sex was asked at each wave, regardless if a participant had their first sex in prior waves, creating a wealth of discrepancies. Almost half the sample who had se x, 48.8%, had some discrepancy in reporting. Some of this could be the difference in rounding with month and year asked in Wave I & II and whole ages in Waves III & IV. 163 analysis of discrepa ncies in sexual history between Wave I & II showed using the earliest wave, along with other methods of reconciliation, did not affect the association with independent variables. I ran logistic regression models to test for significant differences in respo ndents with: (1) any discrepancy; (2) >=1 year; (3) >=2 year; and (4) >=3 year discrepancies ran models using the latest age at sex recorded as well as the mode and mean of responses (see appendix for discrepancy models for age at first sex and age at firs t birth). Similar to Upchurch, 163 I found no difference in relationship to independent variable and no major factors, such as gender, age, SES factors that were related to consistent discrepancies in reporting. I re conciled discrepancies in age at first sex between waves by taking the earliest wave response in the assumption that recall bias increases with each wave. Participants who had their first sex prior to Wave I were not excluded because it removes those with the earliest ages at first sex thus skewing the mean age of first sex

PAGE 54

38 toward later ages. 164 Furthermore, I ran logistic regression descr iptive analyses and found that those who had sex greater than three yea rs prior to wave one were mostly male and mostly third generation which would then bias results if those participants were dropped. Although most variables are measured at wave I and those with earlier ages at first sex may have experienced differences in family structure or other variables at the time of their first sex if their sexual initiation was prior to Wave I, living with two biological parents is measured as those who have always lived in the home, so that relationship would not be affected. Six re spondents were dropped because they listed only one age at first sex and it was prior to age 11. Sex at this early of an age is not likely to be consensual. 165 Item 2: Contraceptive use at first sex: (dichotomous) Did you or your partner use any method of birth control the first time you had sexual intercourse? Co ntraception at first sex was asked in Waves I III. The data were taken from a single question asking about contraception at first sex as well calculating the first sex and use of contraception from the relationship history list for those missing on the sin gle direct question. Unfortunately, 17.9% of participants were missing on this outcome variable. I attempted to account for this missingness in multiple ways, including listwise deletion, multiple imputation by chained equations for missing predictor varia bles only, and multiple imputation by chained equations for missing predictor and outcome variables. The results of these analyses are discussed in detail in Chapter 3. Item 3: Age at first birth: (continuous) Age at first birth was a calculated variable because outcomes of pregnancies were asked during a chronological relationship history section. It was compared to both the age

PAGE 55

39 of children listed in the live birth section and the household roster for that wave. Schoen 166 developed a formula to calculate the age at participant s first birth using the age of the oldest biological child in the home, multiplying by 12 to get months, adding 6 and then subtracting from the interview date. If no birth was mentioned at that time or resulted in a 16% increase in first births that were not reported in the relationship history. Only 15% of respondents had any discrepancy in age at first birth and only 3.5% had more than a year discrepancy. Similar to age at first sex, I chose the earliest age at first birth to reconcile discrepancies. Independent variables. In addition to first, second, and third generation, I initially hop ed to include a generation 1.5, foreign born participants who migrated prior to the age of six, based on previous literature. 58,167,168 I quickly learned that I would not have enough power to include generation 1.5, because generation 1.5 only made up 5.9% of the sample. I decided to use three generations and classify first generatio n as foreign born participants who migrated after the age of six and second generation immigrants as native born, to foreign born parents or foreign born participants who migrated prior to the age of six, based on previous literature. Age six was initially chosen because immigration prior to age six means that the child will attend all of elementary school in the U.S. and not speak with an accent. However, the two qualitative participants who were born in Mexico and migrated prior to age six were so fundame ntally different in their life experiences and outcomes than native born young women (likely due to the difficulties of not being born a U.S. citizen), I could not continue to group any foreign born participants with native born participants. I ran models using various definitions of immigrant

PAGE 56

40 generations, including using the four categories 1.5, six category decimal generations, as recommended by Rumbaut 169 but I did not have enough power to run the models with any more than three categories. I then looked at differences in outcomes as well as demographic variables (i.e., family, language) and found that generation 1.5 were not any more similar to the native born than they were the foreign bo rn who migrated after age six (See appendix). Based on these preliminary analysis, and my qualitative interviews, I then decided to classify all foreign born immigrants as first generation, which is also supported in the literature. 21,26,28,122,170 Third generation immigrants are native born children to native born parents with foreign born grandparents. Add Health did not ask it is made up of any native born parti cipant with native born parents who classify themselves as Mexican or Chicano. They could potentially be fourth or fifth generation. used in the literature. Item 1: Generation (dummy variables as below) Item 2: First generation: age moved to U.S. (continuous) 1. Were you born in the Unites States? (yes or no) 2. In what country were you born? (Mexico or U.S.) 3. In what month and year did you did you first move to the U.S.? (continuous) 4. Where you born a U.S. citizen? (yes or no) (This may be used as a co variate) 5. Was your mother/father born in the U.S.? (yes or no) This question was asked for resident mother or father only. This may or may not be a biolo gical parent. For non resident biological parents, this question was not asked. 6. Where was he/she born? (Mexico or U.S.) First generation: Participant born in Mexico to Mexican born parents. Participants born in Mexico to parents born outside Mexico or the US were excluded. Second generation: Participant born in the U.S. with one or both resident parents born in Mexico

PAGE 57

41 Third and subsequent generation: Participant and resident parents born in the U.S. but self reports to be of Mexican or Chicano or igin Covariates. Variables are discussed in greater detail in Chapters 3 and 5. However, I controlled for gender, dichotomous variable, in each model. I stratified models by gender, but cutting the sample size in half, greatly reduced the power in each mo del. I interacted immigrant generation by gender for each reproductive health outcome and the interactions were not statistically significant which confirms there is no analytical weakness by controlling for gender and not stratifying. Moderators: I inter acted the following variables with immigrant generation. Aim 1: Family structure: Household Roster: to you? How old is {NAME}? Have you and {NAME} always lived in the same household? For how many years have you and {NAME} lived in the same household (1) Two biological parents in home always versus all other families; (2) a grandparent or great grandparent present versus no grandparent or great grandparent present; (3) the number of siblings in the home; and (4) the presence of an older sibling in the home versus no older sibling. Aim 2: Family SES Resident Mother/Father : How far in school did she go? What kind of work does she do? Does she work for pay? Has she worked for pay at any time in the l ast 12 months? Education was continuous. Employment was categorized as categorical (1) unemployed (2) unskilled worker (3) skilled worker (4) professional (See Appendix code book). or Analysis Stata 11 and Stata 13, 171 were used to conduct the quantitative analysis. All participants without a sample weight were dropped and models were adjusted for sample weights and survey design

PAGE 58

42 Descriptive analysis. A descriptive analysis with means, proportions and standard deviations of variables, was performed and included with each respective analysis. Logistic regression. I used logistic regression to analyze the relationship between independent variables (immigrant generation ), moderators, and control variables, and the dependent variable, contraceptive use at first sex. Survival analysis. I chose discrete time because Cox models are best used when time is measured truly continuously. Although age at first birth was measured in months, age at first sex was measured in years. Another assumption of the Cox model is that because time is con tinuous there will not be many events that occur at exactly the same time. Using age in years, there are many participants with events occur ring at the exact same age and thus discrete time was a better analysis for my data. 172 Missing data. With the exception of the large number of missing on contraceptive use at first sex, five participants were dropped due to missing on whether or had sex, and six participants were dropped who had an unknown pregnancy outcome. I was able to narrow down a probable age at first sex for 16 respondents who were missing on age at first sex using the most recent wave they had not yet had sex and the most recent wave they first had sex. For predictor variables, family income had the highest missing at 37%, parental education and employment less than 5% missing, and all other variables less than 1% missing. I used ten multiple imputation cycles by chained e quations 173 which assumes that missing can be missing at random based on variables observed in the data set (MAR) instead of missing completely at random (MCAR) as list wise deletion assumes. I built and rebuilt the imputation models

PAGE 59

43 to ensure the psuedo or R 2 for all family SES variables being predicted was at least 0.2 and that I maintained correlations between the variables. The quantitative analysis used secondary de identified data and was not considered hum an subjects research by the Colorado Institutional Review Board (COMIRB) Protocol 12 1100. Limitations. The Add Health dataset is the largest and most comprehensive data set that can answer the research questions, but it does have a few limitations. First, although longitudinal, the respondents were adolescents in 1994 and the data may be considered outdated. Second, I was limited to variables already in the data set and I was not able to measure gender or traditional values. Third, as with most surveys, re call bias and self report bias are inherent, although Add Health went to lengthy measures to reassure participants about privacy and a longitudinal design reduces recall bias Fourth, the Add Health sampling plan was based on high schools, thus the sample l eaves out adolescents most at risk for childbearing, those who dropped out of school. Fifth, the sample size was relatively small to split it into three immigrant generations which impacted model power. Strengths. This is the first study, to my knowledge, to interact family structure and family SES by immigrant generation to understand their influence on reproductive health outcomes. This is also the first study to use a longitudinal data set to examine differences in adolescent births in multiple immigran t generations. This analysis provided a better understanding of how reproductive health behaviors differ among varying immigrant generation and how family SES and family structure moderate these relationships.

PAGE 60

44 P hase II: Qualitative Interviews Sample Locat ion and recruitment. 138(p32) Therefore, the qualitative sample does no t involve randomization or power calculations, but adds to an understanding of how family characteristics and cultural constructions of gender and sexuality influence reproductive health behaviors in Mexican origin adolescents and young adults. Qualitative interviews were conducted with participants in the Denver metro area. In 2011 Colorado ranked 7 th in Latinos as a percent of population and 78% of Latinos are of Mexican origin. 174 The metro Denver area is an ideal location to interview Mexican origin participants because 32% of the populatio n is Latino 174 and there are many long s tanding Latino communities. COMIRB approved my initial protocol, 12 1432, on December 17, 2012. I planned to use stratified purposive sampling through social networks with assistance from key informants in Mexican origin communities in the Denver metro ar ea, 175 recruiting throu gh passive snowball recruitment If a participant referred a friend and family member to join the study, she would receive $5 as incentive and compensation for her time. I had worked in a community clinic serving the metro Denver community, participated in the CCTSI community training in 2012, and I maintained relationships with community agencies, attending and/or volunteering at events, becoming a member or donating to organizations. I was limited by my COMIRB approved passive recruiting because the poten tial participant had to contact me, someone they did not know. My first interview was in March, 2013 and I had no subsequent interviews until July, 2013 due to

PAGE 61

45 my inability to recruit more actively until I had COMIRB amendments approved. My final interview was conducted in April, 2014 with interview recruiting and data collection lasting approximately 13 months. Retrospectively, I would have chosen a more active strategy to begin with instead of waiting months at a time for COMIRB approval for each change t o my protocol. In addition to the 21 women who completed interviews, four women were screened and did not qualify and two women were screened, qualified, and then canceled or did not show up to the interview. Fifteen of my participants were recruited thro ugh three main venues. I was invited to speak at a community organization on the west side of Denver about my study, bring flyers, and leave a signup sheet which resulted in four interviews. Next, I was introduced to a research assistant who then recruited a coworker and the sample snowballed resulting in six interviews. Finally, a woman from a local community agency contacted potential participants for me and with permission, gave me maining six interviews were recruited by word of mouth from various key informants. Segmented assimilation theory posits that most immigrants live in impoverished urban areas and I planned to sample from urban impoverished areas in metro Denver. I intende d to recruit from a few similar comm unities to reduce confounding of community differences. Because recruiting turned out to be much more difficult and time consuming than I originally planned, I decided to recruit from the whole metro Denver area and s urr ounding suburbs on the assumption that there w ould not be significant sampling bias from one side of town to another. I excluded anyone living outside the metro area, specifically in rural areas which may have introduced more bias because of the

PAGE 62

46 difference s those who chose to live in a rural versus urban or suburban area. No participants were living south of Denver; the farthest west was Golden; the farthest north was Commerce City; and farthest east, Aurora, all considered urban/suburban areas. Particip ants. I interviewed a total of 21 Mexican origin young women, 27 39 years old, living in the metro Denver area between March 2013 and April 2014. Eleven participants were first generation immigrants, born in Mexico, and ten participants were second generat ion, born in the U.S. to Mexican born parents. In 2008 the Add Health participants were 24 32 years old and in 2013 they were 29 37 years old. The initial selection criteria for the adolescent in the U.S. included: age 28 36 (study design started in 2012), female, migration directly to Colorado/Denver from Mexico, and migrated prior to age 18, or born in the US with at least one parent born in Mexico. The ages of inclusion were subsequently widened to 27 41 to facilitate recruitment and the average age of t he 21 qualitative participants was 31.8 which is still similar to 32.3, the average age of Mexican origin Add Health participants in 2013. Demographics and outcomes of qualitative participants are presented in Table 2.3. Men and women are important in the study of gender and sexuality and the quantitative sample controlled for gender and allowed me to detect differences between males and females However, I interviewed females in the qualitative phase in order to ensure the dissertation was feasible. Women are a justifiable starting point for research on gender and sexuality because they are the gender often responsible for contraceptive decisions. 176 I had no inclusion or exclusion criteria based on sexual orientation and all 21 women had sexual relations hips with males, one is currently in a same sex relationship.

PAGE 63

47 Table 2 3 : Demographics of Qualitative Participants First generation Second generation Age at interview (mean) 32.9 30.7 Age at migration (N) <6 18.2% 6 13 27.4% 13 17 18.2% >17 36.4% Age at migration (mean) 12.2 Origin in Mexico* (N) Chihuahua 81.8% 60% Jalisco 18.2% 10% Other 40% Family structure Two biological or adoptive parents 36.4% 70.0% Presence o 0.0% 0.0% Number of siblings 3.6 2.2 Family SES~ Mother's years of education 6.4 7.6 Father's years of education 8.2 9.2 Mother not working 36.4% 60.0% Mother unskilled 54.5% 20.0% Mother skilled/professional 9.0% 20.0 % Father not working 0.0% 10.0% Father unskilled 63.6% 60.0% Father skilled/professional 18.2% 30.0% Reproductive health outcomes Age at first sex 17 17.6 Contraceptive use at first sex 45.5% 70.0% First birth <20 years old 45% 30% Highest level of education completed Less than high school 54.5% 0.0% High school degree 27.3% 60% College or higher 18.2% 40% Observations 11 10 Parents origin for second generation, parents could be from different places re measured at Wave I in Add Health and no qualitative participants lived with grandparents during adolescence, but 4 first gen and 3 second gen lived with grandparents at some point in their lives ~Two first generation participants did not know their fath ers' education or employment

PAGE 64

48 Age. I chose to interview young adults instead of adolescents for three reasons. First and most importantly, the quantitative phase of this dissertation is a randomly selected national sample of adolescents and the qualitati ve phase is local because nationally sampling for qualitative interviews is not feasible nor appropriate In order to 159 I chose participants who are roughly the same age as the participants in the Add Health sample would be in the present in order to have the ability to make meaningful inf erences between the two samples. This means that any of my qualitative participants theoretically could have participated in the Add Health study. Second, similar to Portes, et al., 57 I was able to study the reproductive health ou tcomes (first sex and first birth) in addition to the acculturation process that occurred during adolescence, because these events are likely to have already occurred in young adults. Third, young adults were able to articulate the most important factors t hat they felt influenced their decisions and behaviors (since they have occurred) and while the limitation recall bias and they may not remember what was actually occurring in their families at younger ages, I argue that the most important events in our li ves as adults stand out and are remembered because of the influence they had on our lives. First generation participants were slightly older than second generation participants (mean age 32.9 versus 30.7, respectively. Immigrant generation. As previously d iscussed in the Phase I above, I initially planned on defining second generation as native born to foreign born parents or foreign born who migrated prior to the age of six. Citizenship and documentation played a large role in the lives of the two women I interviewed who migrated before age six and I could

PAGE 65

49 not keep them with classified with their native born peers. All participants born in Mexico were classified as first generation. My quantitative and qualitative results, in addition to previous research 21,58,167,169 found that first generation immigrants are not homogeneous and outcomes are often related to age at migration. My sample of 11 first generation immigrants hav e a variety of ages of migration: two (18.2%) migrated at or prior to the age of six; three (27.4%) migrated between six and 13 years old; two (18.2%) migrated between 13 and 16 years old and four (36.4%) migrated between 17 and 19 years old. The one parti cipant who migrated at 19 years old is slightly outside the sampling frame A lthough she migrated during adolescence at 19, she was married and migrated with her husband. She was at a different life stage than the other participants. I decided to interview the 11 th participant, frame, and although I used her data in the analysis, I qualified any data when it came directly from her experience. While another participant migr ated at 18 and was also married and pregnant, she married and migrated almost immediately after she found out she was pregnant during high school. Although she too was in somewhat of a different life stage, she did not plan this stage and had a similar e xperience to the other women who migrated as adolescents. In the second generation, all of the participants except two had both parents born in Mexico, and one of the two born in the U.S. was actually raised in Mexico from infancy to ten years old despite being born in Texas. While having one parent born in fundamentally different from the second generation 169 quantitat ively the number of 1.75

PAGE 66

50 generation participants in Add Health was too small to analyze separately. Qualitatively I did not believe the experience of the one participant with a parent born and raised in the U.S. differed dramatically enough from the other second generation women to need to be replaced. Although I studied differences between first, second, third generation in the quantitative sample, this is too many categories of qualitative participants to keep the study feasible. I aimed for 8 10 partici pants in each generation in order to reach saturation within groups and three generations would result in 24 30 participants instead of 16 20. An increase in 8 10 participants would require a significant increase in time and resources in addition to the on going quantitative analysis. Additionally, the third may be fifth or sixth generation, resulting in a more heterogeneous population which would have been difficult to r ecruit for. Therefore, I only interviewed first and second generation adolescents. Childhood place of residence. I acknowledge that immigrants from Mexico are a heterogeneous group and that area of origin within Mexico matters. I would have liked to recru it women who were from one or two areas from Mexico, similar to Hirsch, 138 to ensure that the differences I found based on family characteristics were not in fact differences based on the area in Mexico they came from However, due to recruitment difficulties, there was no inclusion criteria based on location in Mexico. Denver tends to have a few prominent sending communities, Chihuahua and Jalisco, and nine (81.8%) of the 11 first generat ion women were from Chihuahua, and six of them grew up in or near Ciudad Juarez (Figure 2.4 177 ). Two (18.2%) were from Jalisco. Seven of the 11 women

PAGE 67

51 migrated directly to the Denver met ro area, while four lived in either Texas or New Mexico for a few years prior to moving to Denver. All the women spent the majority of their time in the US in the Denver area. Figure 2 4 : Map of Mexico 177 Area of origin for second generation women was slightly more diverse including Jalisco, Michoacn, Aguascalientes, Coahuila, and Durango, but six of 10 has at least one parent from Ch ihuahua. Most participants described small to mid size towns. Seven of the 10 women were born and raised in the Denver metro area. Of the three who were born elsewhere in the US, one moved to Denver from Texas at seven years old, one grew up primarily in G reely, Colorado but moved around the country while her migrant worker parents looked for work, and one participant was born and raised in New Mexico and moved to Denver two years ago. While the participant from New Mexico is the only participant who did no t spend her adolescence in urban/suburban Colorado, I do not Chihuahu a Jalisco

PAGE 68

52 believe her experience was substantially different than the rest of the sample. When I referred to her in the results section, I qualified that she was not raised in Colorado. Family structure an d family SES. I had no inclusion criteria related to family structure or SES. Table 2.4 shows there was a greater number of second generation participants living with two biological parents than first generation participants and no participants in either g eneration lived with a grandparent during their adolescence, however, four participants had lived with a grandparent at some point in their lives. As expected, first generation participants had a greater nu mber of siblings than the first There is not a lo t of difference between parental education and parental employment although second generation parents had slightly more years of educati on and were more likely to have be en employed in skilled or professional jobs probably because second generation parent s had been residing in the U.S. longer than first generation parents. discussed in more detail in Chapter 4. Reproductive health and educational outcomes. Although prior research found that first generation adolescents tend to have a later age at first sex than second generation adolescents, 21,23,29 first and second generation participants in my qualitative sample had very similar ages at first sex (17 and 17.6, respectively). This difference, although very unlikely statistical significance, could be related to a second generation outlier who had her first sex at age 23, and it exemplifies one of the limitations to purposive nonrandom sampling. Age at first sex is lower in two biological parent families and there were more first generation participants who experienced family transi tions and

PAGE 69

53 fewer grew up with biological parents than the second generation participants (36.4% versus 70%). As expected a greater number of second generation participants used contraception at their first sex. Slightly more first generation participants ex perienced an adolescent birth. I compared the differences between first and second generation women in the sample to first and second generation women in the Denver metro area to assess representativeness. In my sample, 100% of second generation women had a high school diploma or higher when about 45.5% of first generation women had a high school Boulder metro area, 81.4% of native born Latinos (most similar to my second generation women) had a high school diploma or higher and 45.8% of foreign born Latinos had a high school diploma or higher. 174 A 2011 study at the University of Colorado Hospital enrolled 320 foreign born and native born Latinas in the Denver Aurora metro area and 93.6% of native born women ha d a high school degree or higher and 51.4% of foreign born women had a high school degree or higher. 178 The first generation women in my sample were similar to the expected level of education i n Castillo 178 study and 174 and my second generation women were slightly more educated than average. This could be because six of the second generation women were snowball recruited from one women were more likely to be similar to her in some way. Data collection Interviews. I conducted semi structured life history interviews with young women in order to understand how their fami ly, migration experience, and life experience

PAGE 70

54 help shape their constructions of gender and sexuality. Life history interviews have been used as an interview method in social science research 138,179 182 and they use open ended questions to direct the participant to recall life experiences generally in chronological order, underscoring the importance of location and timing of events. 182 This method was appropriate for my research question in three ways. First, it allowed me to explore how young women make sense of their lives, the context behind t heir behaviors, and the sequence of life events 183 which was particu lar helpful to study processes that are not largely conscious, such as the process of rejection or embracing cultural constructions of family. Second, this method provided a more informal and indirect way to acquire information on more sensitive topics tha t are difficult to get at with other research methods, such as home life and sexual practices. 180 Finally, Abu Lughod 179 argues that behaviors allows for a better understanding of how he or she is embedded in the larger local social lif e, thus life history interviews are ideal to study how cultural constructions of gender and sexuality help explain the relationship between immigrant generation, family SES and family structure, and reproductive health behaviors. I designed the life histor y interviews similarly to Hirsch, 138 chronologically and from less sensitive to more sensitive topics in order to build rapport (Table 2.4). Each woman was given time to read the informe d consent and I answered any questions they had. Many women received a copy of the consent in advance and had ample time to look over it. I had a COMIRB approved waiver of consent for foreign born women so they did not have to sign their name due to fears related to documentation status. I did my best to reassure participants of privacy and build some rapport prior to asking sensitive

PAGE 71

55 questions. I presented myself openly, described why I was doing the study, and stated I am a nurse working with young mother s and previously worked at Clinica Family Health. My experience as a nurse, particularly interviewing patients all day long and my comfort discussing sensitive and sexual topics helped the participants feel more comfortable. (Please see appendix for interview guide). th eir home or workplace. A community organization graciously allowed me to conduct interviews in their building. All interviews were conducted in a quiet and private location although there were young children present during a few of the interviews. Each int erview lasted one to two hours and participants were compensated $20 for their time. Three women completed shorter follow up interviews, receiving $10 to provide clarification on topics covered in the primary interview. Seven of the women received $5 for e ach referral who joined the study. The women were given a choice to conduct the interview in English or Spanish. All 21 women spoke Spanish, five women spoke only Spanish, and 16 women were bilingual. I conducted the five Spanish interviews by myself beca use I am proficient in Spanish. I digitally recorded the interviews and took notes. Analysis De identified interviews were transcribed word for word by three research assistants. Spanish interviews were transcribed in Spanish and then translated in Eng lish Table 2 4 : Interview T opics Childhood and family life (parents, siblings), holiday celebrations, family relationships, parental expectations, gendered division of labor in household Migration experiences, border crossing, initial reactio ns to U.S., memories from home country Education, employment, future plans Dating, boyfriends/husbands, sex, family planning

PAGE 72

56 so there is an English and Spanish transcript for all interviews conducted in Spanish. My primary research assistant wa s bilingual and bicultural. Her parents were born in Guatemala and she has a BA in Public Health with experience transcribing interv iews for prior research studies. She t ranscribed 11 of the interviews and translated four of the Spanish interviews. A second research assistant, a MPH student, transcribed eight English interviews. A third bilingual and bicultural research assistant was a lso an MPH student was second generation Mexican and bilingual, and she transcribed the final Spanish interview because my primary research assistant was not available. I transcribed the first interview myself. I conducted a content analysis on the intervi ews using a five step process developed from Ulin 184 and Strauss and Corbin; 185 reading, coding, displaying, reducing, and interpreting (Table 2.5). Beginning with step 1, reading, as immediately as possible after the interview, I transcribed my notes and placed demographic and outcome da ta into a chart. This approach made sure I did not forget to ask any important information and if I did, I contacted the participant to answer a few missed questions or I set up a follow up interview. After the research assistant transcribed an interview, I listened to the entire interview while reviewing the transcript and made any necessary corrections to the transcript and added notes about tone or background activity/conversations present in the interview. For the Spanish interviews, I listened to the i nterview in Spanish, checked the accuracy of the Spanish transcription and made sure I agreed with the English translation. This ensured accuracy of transcriptions. Although I conducted the Spanish interviews myself, because I am not a native Spanish speak er, I met with my research assistants on three occasions to discuss any discrepancies I found in translation and clarify anything I

PAGE 73

57 did not understand. This approach ensured I did not miss any cultural nuances due to cultural expressions I was not aware of I also kept a qualitative analysis journal, writing down any questions or thoughts I had while listening to interviews. Table 2 5 : Qualitative A nalysis S teps Step Process Data type Reading 1. Familiarity 2. Identify missing areas 3. Identify broad patterns Qualitative Coding 1. Develop code book 2. Open coding 3. Axial/coding sort Qualitative Displaying 1. Examine coding sorts for nuance 2. Develop hypotheses 3. Examine credibility Qualitative and quantitative Reduction Using coding sorts to make matrices and models to make data visible Qualitative and quantitative Interp retation 1. Drawing inferences from data displaying and reduction 2. Analyzing discordance between qualitative and quantitative data Qualitative and quantitative Second, I entered the transcripts in Atlas ti and perf ormed open domain coding based on a priori domains (Table 2.6). During this coding process I also reread the transcripts began to analyze broad themes that I used to assist in the quantitative analysis (see mixed methods chart). I continued writing comment s and questions in my qualitative analysis journal and kept detailed memos in Atlas ti which assisted with organization and provides transparency during data collection and analysis. After I domain coded the interviews, two experienced qualitative research ers each domain coded and subcoded an interview for inter rater reliability. I then went through each interview to recheck the domain codes to ensure I did not miss anything and that all the domains were the same for each interview.

PAGE 74

58 Next, I subcoded the text and disp layed data with matrices to facilitate subcoding (see Appendix for examples of displays). Once I was satisfied with subcodes I proceeded to step four, data reduction. I displayed data by immigrant generation and reproductive health outcomes. Through these displays I identified patterns and developed hypotheses. I frequently returned to the transcripts and to current literature in order to understand, derive meaning, and verify my hypotheses. Once I developed simple arguments to answer my research questions and explained patterns that emerged in vivo, I then returned to the literature to further develop, challenge, or support my arguments. Finally, I integrated my arguments with current literature and theory. The results are discussed in Chapters 4, 5, and 6. Throughout the dissertation, I used pseudonyms for all the participants. If the interview was conducted in Spanish, I provided the Spanish quote first followed by the English translation because translation and interpretation are important and I wanted to ensure I was true to what the participant said. I also edited the quotes for grammar while understanding of the quotes and avoided contributing to stereotypes of improper g rammar Table 2 6 : Qualitative A nalysis C hart A prior themes Conceptual factor Literature and theory Migration history of self or family Macro level factors, U.S. Mexico migration history and patterns, family SES/structure Political econ omic theory. Bohem, 2008; Mahler and Pessar 2001 & 2006 Family structure, parental employment, parent support, relationship with parents Structural Factors, family SES/structure Gender/Sexuality Constructions SAT TGP Documentation, and deportation expe riences Macro level factors SAT Childhood Family SES/structure, Gender and Sexuality Constructions Hirsch 2003; TGP Transnational contacts and experiences Gender and Sexuality Constructions Levitt 1998&2003, Mahler and Pessar 2001 & 2006, Boehm 2008, Hi rsch 2003 Friends, school, future aspirations Gender/Sexuality Constructions SAT Dating, sex, contraception Gender/Sexuality Constructions Hirsch 2003 TGP=theory of gender and power SAT=segmented assimilation theory

PAGE 75

59 in minority and poor persons. I saved all the word for word transcripts in the event a reader would like to see the original transcript. Limitations. Interviewing young adults was an important step to ensure I could make inferences between the quant itative and qualitative sample. However, young adults are removed from their adolescence and recall bias is more of a limitation, though the big events, such as first sex and first birth, were generally recalled with fervor. Self report bias due to embarra ssment or wishing to present oneself a certain way, either what she thought I wanted to hear or know or what she wanted her life to be like, is a limitation although it actually provided interesting data in itself. Similarly, I am aware of the fact that be ing a non Latina white may have hindered women from discussing race/ethnicity related topics Lorena Garcia discussed pros and cons of self disclosure of her Mexican origin background when interviewing Puerto Ricans and Mexican Americans 145 and Jennifer Hirsch, 138 a non Latina white American interviewing Mexican origin women in Atlanta and in Mexico, tried to deemphasize social class differences by not wearing her diamond engagement ring in the field and learning to knit. Most of my participants were given my contact information through people who they trusted or they met me in person at events which increased my credibility. The ability to speak Spanish and just being genuine, greatly imp roved my credibility and comfort of the participants. Most life histories are conducted using multiple interviews with each participant which was not feasible for my mixed methods study. I gained knowledge on a wide range of topics through the life course but there were times I wished I had more detail on particular topics. Despite this limitation, I collected sufficient data to answer my research questions. Snowball recruitment often results in a more homogenous sample, rather than

PAGE 76

60 a diverse sample. Final ly, the small sample size and single location limits generalizability of findings. Strengths. worldviews that underlie racial/ethnic and generational differences in the pathways to early family formation. Research on generational differences among Mexican American 26(p13) Qualitative interviews provide an structure and family SES influence reproductive health behaviors which is not found in quantitative studies. The se results are described in detail in Chapter 4. P hase III: Analysis and Integration of Quantitative and Qualitative Data Analysis In addition to the iterative process between the qualitative and quantitative phases as discussed above, I combined results from both phases to answer my research questions. In Chapter 7 I discuss results of Phase III. I used qualitative data to help explain relationships I found in the quantitative analysis. First, I used qualitative data to inform selection of mediators that age at sexual initiation and the hazard of an adolescent birth differently by immigrant differently by immigrant generat ion. Second, because grandparents and siblings have often been left out in studies examining reproductive health behavior and there are not quantitative variables on relationships with grandparents and siblings in Add Health, I used qualitative data to exp lain how grandparents and older siblings may protect young women from early sexual initiation and non contraceptive use.

PAGE 77

61 Limitations. In addition to the most obvious limitation of mixing a national quantitative data set with a local qualitative sample, a fter quantitative analyses using family structure and SES of moderators, I found that the largest difference in reproductive outcomes was between the first and third generation. This is not surprising because first and second generation immigrants have for eign born parents. Unfortunately, the qualitative analysis is in first and second generation immigrants and may not elucidate the largest quantitative difference between the first and third generation. Strengths. This is the first study, to my knowledge, to combine a quantitative national sample and a local qualitative sample. Social science and public health scholars all recommend the use of a mixed methods approach to understand reproductive health disparities. 31,83,126,186 189

PAGE 78

62 CHAPTER III REPRODUCTIVE HEALTH OUTCOMES, IMMIGRANT GENERATION, AND FAMILY IN MEXICAN ORIGIN ADOLESCENTS Prior r esearch has shown that first generation or foreign born Latino adolescents are less likely to engage in sex, less likely to give birth and have fewer sexual partners than their native born peers. However, foreign born Latinos are also less likely to use co ntraception or condoms than their native born peers. 19,21 29,190,152 Public health and social science researchers have yet to unravel the reasons why adolescents of differing immigrant generations make different reproductive health decisions. Drawing o n segmented assimilation theory that asserts family structure and SES can predict assimilation pathways, I hypothesized that family structure and family SES will infl uence reproductive health behaviors differently by immigrant generation. Although I am examining reproductive health outcomes and not economic assimilation, earlier sexual initiation, non contraceptive use, and adolescent childbearing may increase risk for downward economic assimilation. Figure 3.1 shows the conceptual model for Aims 1 and 2. In this chapter, I examined how immigrant generation moderated the relationship between family structure and family SES and reproductive health behaviors. The chapter is organized into three sections: sexual initiation, contraceptive use at first sex, and the hazard of an adolescent birth. I close with tying the three sections together and discussing how the analyses support segmented assimilation theory.

PAGE 79

63 Figure 3 1 : Conceptual Model for Aims 1 & 2 S exual Initiation Background Early sexual initiation is associated with an increased risk of sexually transmitted infections (STIs) and adolescent births. 191,192 Despite the rapid growth of the Latino population and disparities in STI and birth rates among Latino adolescents, 7,17,193 we have only begun to disentangle sexual health disparities by country of origin and immigrant generation. The substantial size and economic disadvantage of the Mexican origin population, along with poor socioeconomic outcomes in Mexican origin adolescents, 54,57,58 make it imperative to better understand sexual health disparities in this population. Reproductive health outcomes vary by immigrant generation and length of time in the U.S. 69,118 Although previous resea rch shows that first generation Latino adolescents have a later age at sexual initiation than subsequent generations, reasons behind this finding remain poorly understood. 21,28,23,29 Segmented assimilation theory suggests that tra ditional cultural values protect immigrant children from adverse outcomes, 57 including early onset of sexual activity. Although researchers seldom measure values directly, Mexican origin immigrant families with two married parents 57 and grandparen ts, 104,105 Family Structure (P arents, grandparents, siblings) Family SES ( P arental education, employment) Immigrant generation First, second, third Behavior Sexual i nitiation contraception first birth IV DV Moderator

PAGE 80

64 and in families where migrant parents have less education and work in unskilled jobs more often hold traditional values such as the importance of family or conservative sexual values, such as the importance of virginity 139,140,194,195 As time in the U.S. increases, families may replace traditional cultural values with the less protective values of disadvantaged native born communities thereby increasing the risk of downward assimilation and r isky health behaviors. 87 Drawing on segmented assimilation theory, my first hypothesis was that more traditi onal family structures that include two biological parents, grandparents, or more siblings, will provide greater protection against early sexual initiation for first and second generation adolescents than for third generation immigrants. Foreign born famil ies tend to be of lower SES than third generation families and yet they have later ages at sexual initiation, 53 perhaps because adults in the home are more likely to reinforce protective traditional cultural values such as the importance of family ( familismo ) and respect for others ( respeto ), especially elders. 105,109 The presence of two biological parents and grandparents provides supervision, guidance, and stability that increase the likelihood that children of immigrants will assimilate into the middle class. 57 Mexican origin adolescents in first and second generation families have more siblings than third generation adolescents 106 which may reflect stronger traditional values in the family and decrease the risk for early sexual initiation. While the presence of an older sibling in the home has been shown to increase risky sexu al behavior for younger siblings 111,112 adolescents in Mexican immigrant families spend more time with their siblings than in non immigrant Mexican families 196 and older siblings tend to have responsibility for younger siblings. 144,145 Thus, first and second generation adolescents may spend more

PAGE 81

65 time with a nd listen to the traditional advice and guidance of their siblings resulting in delayed sexual initiation as compared to third generation adolescents. Figure 3.2 shows my hypotheses for the interactions between family structure variables and immigrant gene ration us ing estimated hazard ratios. Figure 3 2 : Hypotheses for F amily S tructure and I mmigrant G eneration Interaction My second hypothesis was that families with less educated parents or parents who work in unskilled occupations will protect against the early onset of sex in first and second generation immigrant families than in third generation families. Parents with more education and who work in more skilled occupations may improve the future economic w ellbeing of their children. 87 However, low socioeconomic status (SES) families may be more likely to maintain traditional gender norms; 139,140 parents with less education and Two bio parents Hazard of sexual initiation First Second Third Immigrant generation HR 0.5 HR 0.5 HR 0.8 Grandparent Hazard of sexual initiation HR 0.5 HR 0.5 HR 0.8 First Second Thi rd Number of siblings Hazard of sexual initiation Hazard of sexual initiation First First Second Second Third Third HR 0.5 HR 0.8 HR 0.8 HR 0.5 HR 0.8 HR 1.2 Older sibling

PAGE 82

66 unskilled employment may hold traditional gender values that emphasize the importance of virginity until marriage and may have stricter dating rules based on those values. In fact, Romo and colleagues found that maternal education was negatively associated with the time mothe rs spent talking to their daughters about delaying intercourse. 194 Additionally, mothers who do not work outside the home may provide supervision which protects adolescents from early sexual initiation. 197 Figure 3.3 shows my hypotheses for the interactions between family SES variables and immigrant generation using estimated hazard ratios.

PAGE 83

67 Figure 3 3 : Hypotheses for Family SES and Immigrant Generation Interaction

PAGE 84

68 M ethods I used data from all four waves (1994 2008) of the National Longitudinal Study of Adolescent Health (Add Health). The survey began in 1994 with 20,475 7 12 graders and participants were interviewed again in 1996 (Wave II), 2001 2002 (Wave III), and 2007 2008 (Wave IV). 160 I restricted the sample to Mexican origin r espondents. I excluded five participants missing on ever having sexual intercourse, six participants missing on the presence of a parent in the home and seven participants with an age at first sex prior to 11 due to the unlikelihood that sex was consensual at that age. 165 After excluding respondents who were missing without a sample weight, the final sample size was 1,637 Mexican origin participants. Variables. The primary outcome of interest was age at first sex. The respondent ti the age at first sex recorded in the earliest wave under the assumption that recall bias increases over time. 163 My analysis included those who had their first sex prior to Wave I. If I excluded those participants, I would remove those with the earliest ages at first sex, specifically unequal numbers of males and third generation adolescents, which would bias the sample. 164 Table 3. 1 presents the independent variables and how they were measured. All independent variables were measured at Wave I with the exc eption of age and age squared which are time varying.

PAGE 85

69 Table 3 1 : Independent Variables Independent variables Measured Type of variable Immigrant generation First=Born in Mexico to Mexican born parents Second=Born in the U.S. to at least one Mexican born parent Third=Born in the U.S. to two U.S. born parents and classifies ethnicity as Mexican 167,198 Categorical Family structure Two biological parents Participant lived with two biological or t wo adoptive parents for his/her entire life Dichotomous Presence of a grandparent in the home Participant lived with a grandparent or great grandparent in the home at Wave I Dichotomous Number of siblings in the home T he number of biological, half, or step siblings in the home at Wave I Continuous Has an older sibling Participant has an older full biological sibling who may or may not be in the home at Wave I. Dichotomous Family SES Number of years of education: 0= never attended school 8= 8 th grade or less 10= no high school degree but more than 8 th grade 12=GED or diploma 16=college graduate Continuous ion As above Continuous 1=not working outside of the home 2=unskilled employed 199 3=skilled/professionally employed Categorical As above Categorical Control variables Age Age in years Time va rying continuous Age squared Age in years squared Time varying continuous Gender ~ 1=male 0=female Dichotomous Family income Logged annual household income, adjusted for inflation and family size at Wave I Continuous Resident father Presence of a father figure (biological, step) in the home Dichotomous *Because of the small sample size in first generation mothers for skilled (n=50) and professional employment (n=25), they were combined for all generations. ~I controlled for gender instead of estimating stratified models because there were no significant interactions between immigrant generation and gender. Analysis. I modelled the age at sexual initiation using complementary log log discrete time hazard models. I used a person year dataset and particip ants were censored

PAGE 86

70 at the year they experienced the event or the age of their last interview. 172 I right censored all participants at age 20 because I am concerned with the consequences of early sexual initiation and because participants with a first sex after age 20 are no longer at risk for an adolescent pregnancy. A sensitivity analysis showed no substantial differences in results when right censoring at age 17 and right censoring at the age of the last interview. I included age and age squared in the model beca use there is a concave relationship between age and first sex. Please see appendix for a summary of robustness testing. I first examined main effects of each family structure and family SES variable and then I interacted immigration generation with each p redictor to understand whether the relationship between the hazard of sexual initiation and the family structure or family SES variable differed across immigrant generations. I controlled for two biological parents in models that included a grandparent var iable to account for confounding. Roughly 23% of participants had no resident father and 5% no resident mother. For particular parent in the home w ere included. Because more educated parents tend to have more skilled or professional employment, I added an education variable in models that included an employment variable and added an employment variable in models that included an education variable to reduce confounding. I us ed ten multiple imputed data sets created by chained equations to address missing data. 173 Family income had the highest missing at 37%, parental education and employment were missing less than 5%, and all other v ariables were missing less than 1%. I also adjusted for the complex sampling frame of Add Health, including sampling weights to the U.S. population.

PAGE 87

71 Results Table 3. 2 presents weighted means and percentages of the study variables by sexual initiation (pri or to age 20) and immigrant generation. Almost 70% of participants had sexual intercourse prior to age 20. Adolescents who had their first sexual intercourse prior to age 20 are less likely than those who had intercourse after age 20 to have lived with two biological parents and grandparents and are more likely to have skilled or professional parents at Wave I. Across immigrant generations, SES improves with each immigrant generation as parents become more educated and are more likely to be skilled or profe ssional workers. The percentage of participants who had sex prior to age 20 also increases with each generation and the mean age at first sex decreases with each generation. Table 3 2 : Weighted means ( S tandard D eviations) and P roportions of S elect V ariables Ever had sex (<20 years old) Never had sex First generation Second generation Third generation N= 1126 (68.8%) 511 (31.2%) 357 (21.8%) 593 (36.2%) 687 (42.0%) Person years 7,464 3,098 1,983 4,673 3, 906 Variable Age at first sex 17.2 (2.6) 17.2 (3.3) 16.3 (3.8) Family structure Two biological or adoptive Parents 52.5% 68.2% 49.8% 72.0% 48.7% Presence of a grandparent 3.8% 8.8% 1.5% 7.1% 5.9% Older sibling 60.7% 67.1% 72.1% 61.8% 58.5% Number of siblings 2 (1.6) 2.1 (1.2) 2.4 (1.6) 2.4 (1.4) 1.6 (1.4) Family SES Mother's years of education 11 (3) 10.4 (3.3) 8.7 (3.1) 10 (2.7) 12.6 (2.2) Father's years of education 10.9 (3.2) 10.7 (3.1) 8.7 (3) 10.1 (2.5) 13 (2. 4) Mother unemployed 41.1% 45.2% 52.8% 47.1% 33.3% Mother unskilled 27.3% 29.8% 39.6% 34.7% 16.7% Mother skilled/professional 31.5% 24.9% 7.6% 18.3% 50.0% Father unemployed 12.5% 11.9% 14.0% 13.7% 9.7% Father unskilled 46.9% 46.7% 60.7% 51.3 % 33.5% Father skilled/professional 40.6% 41.4% 25.3% 35.0% 56.8% Control variables

PAGE 88

72 Table 3 2 : Weighted means ( S tandard D eviations) and P roportions of S elect V ariables Family income $29,200 (3,080) $28,600 (2,050) $17,000 (1,470) $26,900 (3,230) $37,100 (2,860) Resident father in home 71.9% 83.2% 77.0% 85.8% 65.8% Male 52.2% 42.7% 45.1% 48.4% 52.1% Wave I Age 16.1 (2.1) 16.1 (1.4) 16.7 (2.1) 16.1 (2.2) 15.8 (1.4) Family structure: Parents and grandparents Tables 3. 3 and 3. 4 show the hazard ratios for initiation of first sex and family structure. Table 3. 3 Model 1 sh ows that compared to the hazard for sexual initiation in the third generation, the hazard is lower in the first and second generation. Post estimation t test found that the hazard of sexual initiation between the first and second generation is also statist ically significant (p<0.001). Model 2 shows that having two biological or adoptive parents in the home reduces the hazard of sexual initiation by 38% (1 0.62). The hazard ratio for the second generation became closer to 1 and the significance dropped from <0.001 to <0.05 suggesting two biological parents is a mediator and one of the reasons the second generation has a lower hazard for sexual initiation than the third generation. Model 3 shows that the interaction between immigrant generation and two biologi cal parents is not significant. Model 4 shows that having a grandparent in the home reduced the hazard of sexual initiation by 31% (1 0.69), (p<0.10). Model 5 found no significant interaction between immigrant generation and having a grandparent in the hom e. Table 3 3 : Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and F amily S tructure Model 1 Model 2 Model 3 Model 4 Model 5 First generation 0.68** 0.68** 0.66* 0.67** 0.67** (0.53 0.88) (0.52 0.89) (0.47 0.92) (0.51 0.88) (0.51 0.88) Second generation 0.71*** 0.79* 0.80* 0.81* (0.60 0.85) (0.66 0.95) (0.53 1.03) (0.66 0.96) (0.66 0.98) Third generation Ref. Ref. Ref. Ref. Ref. Two bio parents 0.62*** 0.59** 0.62*** 0.62***

PAGE 89

73 Table 3 3 : Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and F amily S tructure (0.54 0.73) (0.43 0.81) (0.53 0.72) (0.53 0.72) First gen X two bio parents 1.08 (0.75 1.54) Second gen X two bio parents 1.14 (0.65 1.99) Third gen X two bio parents Ref. Grandpare nts 0.73 (0.45 1.04) (0.36 1.48) First gen X grandparents 1.38 (0.22 8.44) Second gen X grandparents 0.83 (0.23 2.99) Third gen X grandparents Ref. Age 16.06*** 15.91*** 15.92*** 15.67*** 15.66*** (9.04 28.51) (8.96 28.26) (8.97 28.26) (8.82 27.85) (8.79 27.89) Age*Age 0.92*** 0.92*** 0.92*** 0.92*** 0.92*** (0.91 0.94) (0.91 0.94) (0.91 0.94) (0.91 0.94) (0.91 0.94) Male 1.30** 1.30** 1.29** 1.28** 1.28* (1.09 1.54) (1.0 8 1.55) (1.08 1.55) (1.07 1.54) (1.06 1.55) Constant 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) Observations 10,562 10,562 10,562 10,562 10,562 *** p<0.001, ** p<0.01, p<0.05 p<0.10 Family structure: Siblings Table 3. 4 Model 1 and Model 2 show that the number of siblings is not associated with age sexual initiation in any immigrant generation. Model 3 shows that having an older sibling in the home reduced the hazard o f sexual initiation of 15% (1 0.85) (p<0.10). Model 4 shows there is no significant interaction between having an older sibling and immigrant generation.

PAGE 90

74 Table 3 4 : Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and F amily S tructure Model 1 Model 2 Model 3 Model 4 First generation 0.69** 0.75 0.69** 0.78 (0.54 0.88) (0.46 1.22) (0.54 0.88) (0.55 1.13) Second generation 0.72*** 0.59** 0.71*** 0.62*** (0.60 0.87) (0.41 0.85) (0.60 0.85) (0.48 0.79) Third generation Ref. Ref. Ref. Ref. Number of siblings 0.98 0.96 (0.92 1.05) (0.84 1.09) First gen X siblings 0.97 (0.82 1.16) Second gen X siblings 1.10 (0.93 1.30) Third gen X siblin gs Ref. Older biological sibling 0.82 (0.73 1.02) (0.61 1.09) First gen X older sibling 0.85 (0.54 1.33) Second gen X older sibling 1.28 (0.88 1.86) Third gen X older sibling Ref. Age 16.11*** 15. 92*** 16.02*** 15.98*** (9.04 28.71) (8.91 28.43) (9.05 28.36) (8.97 28.44) Age*Age 0.92*** 0.92*** 0.92*** 0.92*** (0.91 0.94) (0.91 0.94) (0.91 0.94) (0.91 0.94) Male 1.29** 1.29** 1.29** 1.30** (1.09 1.54) (1.08 1.54) (1.09 1.54) (1.09 1.54) Constant 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) Observations 10,562 10,562 10,562 10,562 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Family SES: Parental education Table 3. 5 shows the hazard ratios for initiation of first sex and parental education. initiation in the first generation (1.07*1.01=1.08) (p<0.08), but is not significantly different from one in the second (1*1.01=1.01) and third generation. Model 3 shows that

PAGE 91

75 Mode sexual initiation increases by 1% (1.10*0.92) in the first generation, and by 7% (1.15*0.92) in the second generation, but decreases by 7% (1.15*0.92) in third generation. Table 3 5 : Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and P arental E ducation Model 1 Model 2 Model 3 Model 4 First generation 0.37* 0.60** 0.18** (0.55 1.03) (0.15 0.90) (0.4 4 0.81) (0.07 0.50) Second generation 0.84 0.75 0.65** 0.12*** (0.66 1.06) (0.29 1.94) (0.50 0.85) (0.05 0.30) Third generation Ref. Ref. Ref. Ref. Mother's education 1.02 1.00 (0.99 1.06) (0.94 1.06) First gen X mother's education (0.99 1.17) Second gen X mother's education 1.01 (0.93 1.09) Third gen X mother's education Ref. Father's education 0.99 0.92** (0.95 1.03) (0.86 0.97) First gen X father's education 1.10* (1.01 1.20) Second gen X father's education 1.15*** (1.07 1.25) Third gen X father's education Ref. Not working mother 0.99 0.96 (0.77 1.27) (0.75 1.23) Unskilled mother 0.95 0.92 (0.69 1.31) (0.66 1.28) Skille d/Professional mother Ref. Ref. Not working father 1.09 1.00 (0.75 1.58) (0.68 1.48) Unskilled father 1.14 1.09 (0.90 1.44) (0.86 1.38) Skilled/Professional father Ref. Ref. Age 16.00*** 15.96*** 14.18*** 14.17*** (9.52 26.89) (9.49 26.85) (6.97 28.84) (6.95 28.88) Age*Age 0.92*** 0.92*** 0.93*** 0.93*** (0.91 0.94) (0.91 0.94) (0.91 0.95) (0.91 0.95)

PAGE 92

76 Table 3 5 : Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and P arental E ducation Male 1.28** 1.28** 1.44*** 1.45*** (1.08 1.53) (1.08 1.52) (1.23 1.70) (1.23 1.72) Fa mily Income 0.97 0.97 0.97 0.98 (0.85 1.10) (0.85 1.10) (0.84 1.13) (0.84 1.14) Resident Father 0.72* 0.72* (0.56 0.93) (0.56 0.93) Constant 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) Ob servations 10,007 10,007 8,039 8,039 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Family SES: Parental employment Table 3. 6 shows the hazard of sexual initiation and parental employment. Model 1 found no significant association between sexual initiation a Figure 1 graphs the results from Table 3. 6 Model 2, and shows that the hazard for first generation adolescents with mothers who are not working was reduced by 0.45 (0.38*1.19) as compared to first generation adolescents with skilled or professional mothers. Post estimation tests found that the hazard ratio between first generation adolescents with mothers who are not working and third generation moth ers who are not working was significantly different (p=0.045) but the hazard ratio between first and second generation mothers who are not working is not significantly different. Table 3. 6 signif icantly different than 1. Model 4 shows a significant interaction between generation 3.4 graphs the results from Table 3. 6 Model 4, and shows that the hazard of sexual initiation for second generation adolescents with unski lled fathers was reduced by 19% (0.50*1.61) as compared to second generation adolescents with a skilled/professional father. Post estimation tests found the hazard ratio between second generation adolescents with unskilled fathers and third generation

PAGE 93

77 adol escents with unskilled fathers was significantly different at p=0.001. The hazard for sexual initiation for third generation adolescents with unskilled father increased by 1.61 (1*1.61) as compared to third generation adolescents with skilled or profession al fathers. Post estimation tests found the hazard of sexual initiation for first generation adolescents with unskilled fathers was moderately significantly different than the hazard of sexual initiation for third generation adolescents with unskilled fath ers (p=0.039). Figure 3.5 employment. Table 3 6 : Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and P arental E mploym ent Model 1 Model 2 Model 3 Model 4 First generation 1.60 0.60** 0.77 (0.55 1.03) (0.70 3.66) (0.44 0.81) (0.43 1.40) Second generation 0.84 1.04 0.65** 0.98 (0.66 1.06) (0.73 1.48) (0.50 0.85) (0.66 1.46) Third generation Re f. Ref. Ref. Ref. Unemployed Mother 0.99 1.19 (0.77 1.27) (0.80 1.78) Unskilled Mother 0.95 1.05 (0.69 1.31) (0.65 1.71) Professional Mother Ref. Ref. First gen X not working mother 0.38* (0.16 0.91) First g eneration X unskilled Mother 0.45 (0.17 1.18) First generation X skilled/profess mother 1.00 (1.00 1.00) Second generation X unemployed mother 0.68 (0.39 1.18) Second gen X unskilled mother 0.78 (0.43 1.43) Secon d gen X skilled/profess mother Ref. Not working Father 1.09 1.43 (0.75 1.58) (0.74 2.77)

PAGE 94

78 Table 3 6 : Hazard R atios (95% C onfidence I ntervals) for S exual I nitiation and P arental E mploym ent Unskilled Father 1.14 1.61** (0.90 1.44) (1.15 2.26) Professional Father Ref. Ref. First gen X not working father 0.69 (0.26 1.81) First gen X unskilled father 0.66 (0.35 1.27) First gen X skilled/profess father Ref. Second gen X not working father 0.58 (0.30 1.13) Second gen X unskilled father 0.50** (0.32 0.76) Second gen X skilled/profess father Ref. Mother's education 1.02 1.02 (0.99 1.06) (0.98 1.06) Father's education 0.99 1.00 (0.95 1.03) (0.96 1.04) Age 16.00*** 16.10*** 14.18*** 14.36*** (9.52 26.89) (9.63 26.93) (6.97 28.84) (7. 09 29.10) Age*Age 0.92*** 0.92*** 0.93*** 0.93*** (0.91 0.94) (0.91 0.94) (0.91 0.95) (0.91 0.95) Male 1.28** 1.28** 1.44*** 1.40*** (1.08 1.53) (1.07 1.54) (1.23 1.70) (1.18 1.67) Family Income 0.97 0.97 0.97 1.00 (0.85 1.10) (0.85 1.10) (0.84 1.13) (0.85 1.16) Resident Father 0.72* 0.72* (0.56 0.93) (0.55 0.92) Constant 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) Observations 10,007 10,007 8,039 8,039 *** p<0.00 1, ** p<0.01, p<0.05, p<0.10

PAGE 95

79 Figure 3 4 : Hazard R atio of S exual I nitiation, I mmigration G eneration by Mo ther's E mployment Figure 3 5 : Hazard R atio of S exual I nitiation, I mmigration G eneration by F ather's E mployment Discussion Table 3. 7 summarizes the results of family structure models and support for segmented assimilation theory. Extending segmented assimilation theory and prior research, I found that although living with two biological parents and a grandparent protected against sexual initiation equally in each immigrant generation, I found that 0.60 0.80 1.00 1.20 1.40 1.60 1st gen 2nd gen 3rd gen Not working Unskilled Skilled/Professional 0.60 0.80 1.00 1.20 1.40 1.60 1st gen 2nd gen 3rd gen Not working Unskilled Skilled/Professional

PAGE 96

80 parental education and employment influenced the hazard of sexual initiation quite differently across immi grant generations. My results challenge the assumption that improved parental education and employment always protect adolescents from risky sexual behaviors. This new information provides a crucial first step in unraveling the complex relationship between family, immigrant generation and sexual initiation in the disadvantaged and growing Mexican origin adolescent population. I advanced current research by examining whether the relationship between family structure and the hazard of sexual initiation differ ed across immigrant generations. Supporting previous research and segmented assimilation theory, 95,96 I found that living with two biological parents decreased the hazard of sexual initiation. The influence of grandparents has been overlooked in adolescent reproductive health literature and extending previous research and segmented assimilation theory I found that the presence of a grandparent in the home also decreased the hazard of sexual initiation. In addition to the benefits of two biological parents in the home, grandparents may also aid in supervision and encourage protective traditional values, such as the importance of virginity. 105,109 My findings challenge previous research that found an increased risk of sexual initiation in adolescen ts with an older sibling in the home. 111,112 I found that having an older sibling in the home decreased the hazard of sexual initiation in all immigrant generations. Previous studies did no t stratify by race/ethnic origin and siblings in Mexican origin families spend more time together than non Latino white siblings 107,200 which could symbolize closer relationships between siblings which have been found to have lower sexual risk behavior. 113 Further, older siblings tend to have more domestic

PAGE 97

81 responsibility in the house than younger siblings, includin g responsibility to care for younger siblings. 108,144,145 Older siblings may spend more time with, be closer to, and feel more responsible for their younger siblings, resulting in more guidance to delay sexual initiation. Further research is needed to understand how older siblings may protect younger siblings from sexual initiation in Mexican origin families. Table 3 7 : Summary of Family Structure and the Hazard of Sexual I nitiation Variable HR Supported hypothesis? Supported SAT? Two bio parents 0.62* ** (0.54 0.73) Partial: No significant interaction by immigrant generation but two biological parents reduced the hazard of sexual initiation in all generations. Two parents partially mediated the hazard between the second and third generations. Supports SAT Grandparents 0.69 (0.45 1.04) Partial: No significant interaction by immigrant generation but a grandparent in the home reduced the hazard of sexual initiation in all generations Extends SAT, no prior mention of grandparents although extending th e theory they should be protective Number of siblings 0.98 (0.92 1.05) No: No significant main effect or interaction Neither supports nor rejects. Siblings not mentioned in SAT. Older sibling 0.86 (0.73 1.02) Partial: No significant interaction but an older sibling decreased the hazard of sexual initiation in all generations. Extends SAT. No prior mention on siblings although migrant families would have closer relationships. *** p<0.001, ** p<0.01, p<0.05, p<0.10 Table 3. 8 summarizes parental education models and support for segmented assimilation theory. Table 3. 9 summarizes parental employment models and support for segmented assimilation theory. The positive relationship between parental education and employment and the hazard of sexual ini tiation in the first and second generation supports my hypotheses. These findings are particularly noteworthy given that previous research has found that higher parental SES is protective for many health and educational outcomes and reduces the risk for se xual initiation in Latinos. 31,88,93,115,118 I n addition to

PAGE 98

82 the assumption that improved parental socioeconomic status is always protects adolescents, it is often assumed that traditional cultural values are eroded with Mexico to U.S. migration and immersion in a more gender equal society. Because I f ound similar influences of parental SES in first and second generation adolescents, I agree with recent studies who found that this is not always the case and traditional cultural values such as conservative or patriarchal sexual values, may be reinforced in immigrant families due to structural constraints such as workplace exploitation and lack of documentation status. 139,140 Further, Gonzalez Lopez 19 5 found that migrants from rural areas in Mexico held more conservative sexual views as compared to migrants from urban areas, largely because of increased access to higher education and professional employment for men and women in urban areas. Perhaps le ss educated and underemployed Mexican born parents encourage more traditional cultural values, have stricter rules for dating, and provide more supervision, particularly when mothers do not work outside the home while more educated and professionally emplo yed Mexican born parents have more liberal sexual views and may be more permissive. In third generation adolescents, I found a negative relationship between parental education and employment of parents and the hazard of sexual initiation, suggesting that t raditional cultural values eventually erode over time because parental SES becomes protective by the third generation Table 3 8 : Summary of P arental Education and the H azard of S exual I nitiation Variable HR Supported hypothesis? Supported SAT? Mother's education 1.00 (0.94 1.06) Partial: In the first education increased, the hazard of sexual initiation decreased. The interaction was not significantly different from 1 in the second or third generation. Rejects and extends SAT which asserts that higher parental education improves adolescent outcomes. Higher parental education works differently in the first generation. First gen X mother's education (0.99 1.17) Second g en X mother's education 1.01 (0.93 1.09) Third gen X mother's education Ref.

PAGE 99

83 Table 3 8 : Summary of P arental Education and the H azard of S exual I nitiation Father's education 0.92** (0.86 0.97) Yes. In the first and second reduced the hazard of sexual initiation as compared to the third genera tion. Rejects and extends SAT which asserts that higher parental education improves adolescent outcomes. Higher parental education works differently in the first generation. First gen X father's education 1.10* (1.01 1.20) Second gen X father's educ ation 1.15*** (1.07 1.25) Third gen X father's education Ref. *** p<0.001, ** p<0.01, p<0.05, p<0.10 Table 3 9 : Summary of P arental Employment and the H azard of S exual I nitiation Variable HR Sup ported hypothesis? Supported SAT? Not working Mother 1.19 (0.80 1.78) Partial: In the first generation, a mother not working reduced the hazard of sexual initiation compared to a professional mother. This interaction was not significant in the second ge neration. Rejects SAT that asserts skilled and professional parents improve second generation outcomes. Extends SAT by showing skilled and professional parents are not always protective for all outcomes. Unskilled Mother 1.05 (0.65 1.71) Professi onal Mother Ref. First gen X not working mother 0.38* (0.16 0.91) First generation X unskilled Mother 0.45 (0.17 1.18) First generation X skilled/profess mother 1.00 (1.00 1.00) Second generation X unemployed mother 0.68 (0.39 1.18) Second gen X unskilled mother 0.78 (0.43 1.43) Second gen X skilled/profess mother Ref. Not working father 1.43 (0.74 2.77) Partial: In the second generation, an unskilled father reduced the hazard of sexual initiation compared to a professional father. In the third generation, an unskilled father increased the hazard of sexual initiation compared to professional fathers. This interaction was not significant in the first generation. Rejects SAT that asserts skilled and professional parents imp rove second generation outcomes. Extends SAT by showing skilled and professional parents are not always protective for all outcomes. Unskilled father 1.61** (1.15 2.26) Professional father Ref. First gen X not working father 0.69 (0.26 1.81) First generation X unskilled father 0.66 (0.35 1.27) First generation X skilled/profess father Ref. Second generation X unemployed father 0.58 (0.30 1.13) Second gen X unskilled father 0.50** (0.32 0.76) Second gen X skilled/profess fath er Ref. *** p<0.001, ** p<0.01, p<0.05, p<0.10

PAGE 100

84 Segmented assimilation theory predicts upward economic assimilation in the second generation in families with two biological parents, higher parental SES, with reinforcement of traditional protective values. While I provide direct support for the importance of two biological parents and our findings suggest that erosion of traditional protective values is one of the reasons the third generation has a higher risk for sexual initiation despite higher par ental SES, the positive relationship between higher parental SES and risk for sexual initiation in the first and second generation elucidates a relationship between parental SES and traditional values or conservative sexual values that is not fully explain ed with segmented assimilation theory. I advocate for future research to improve understanding and measurement of Mexican origin values and their influence on health behaviors, specifically to test if traditional values, stricter dating rules, and supervis ion are some of the reasons lower education and underemployment in the first and second generation protect against early sexual initiation. I acknowledge the structural reasons behind health disparities. However, simil ar to researchers who found an increa sed risk of HIV with higher SES in East Africa, 201 I assert there are consequences to defining adolescent sexual risk behaviors as a proble m of low income groups alone. When health care providers and public health programs target low SES families, we may miss opportunities for health education or even fail to offer sexually transmitted testing to adolescents of higher SES families. Supporting segmented assimilation theory, my study also suggests that Mexican origin immigrant families protect their adolescents from early sexual initiation through protective traditional values that outweigh the consequences of lower parental SES. It is important to note that I do not assume traditional cultural values that may protect against early sexual initiation are

PAGE 101

85 protective for all reproductive health behaviors such as contraceptive use. In addition to exploring how traditional values protect adolescents f rom risky sexual behaviors, we should find ways of helping higher SES immigrant families maintain protective traditional cultural values while continuing higher educated trajectories. My study is the first to use a discrete time survival model from a lon gitudinal nationally representative data set using interactions to examine how family structure and family SES influence the hazard of sexual initiation differently across immigrant generations. However, there are several limitations to my study. First, I did not have the ability to measure traditional values directly. However, the use of segmented assimilation theory and previous research support the conclusions I have drawn between family structure, family SES and traditional values. Due to sample size I was unable to test for differences in age at migration within the first generation, although previous literature found that age of immigration influences outcomes. 21,167,169 My sensitivity analysis using a six category age at migration variable, 169 found the hazard for sexual initiation increased as age of migration increased (see Appendix). Recall bias can affect the internal validity of the study, especially if adolescents did not accurately report their age at first sex, although the longitudinal design of the study decreases recall bias. Conclusion In this analysis, I have extended prior research and segmented assimilation theory by challenging the assumption that improved parental socioeco nomic status always protects adolescents from risky sexual behavior. Public health experts and clinicians should not overlook sexual and reproductive risk in higher SES immigrant Mexican origin families. Furthermore, scholars should consider the difference s in both adolescent

PAGE 102

86 reproduce health outcomes across immigrant generations and how factors that predict outcomes work differently across immigrant generations. Interventions that can prevent erosion of protective cultural values in Mexican origin families while improving academic and vocational trajectories as time and generations in the U.S. increase, may dramatically reduce sexual and reproductive health disparities in Mexican origin adolescents. C ontraception at Sexual Initiation Background Risk for adolescent pregnancy and birth is due to sexual intercourse, particularly earlier sexual initiation, and non contraceptive use. 31 Prior research has found that Latino adolescents are less likely to use contraception than non Latino white or non Latino black adolescents, 19,20,152,186,191,202 although less is known about contraceptive use within the diverse Latino population. Mexican origin adolescents are important to study due to the size of the population 6,55 and adverse outcomes, including increased risk for adolescent births, 18 compared t o other Latino groups, such as Puerto Ricans and Cubans. According to the literature, first generation Latino adolescents have a lower risk for early sexual initiation, 21,28,23,29 but they generally have a higher risk for non con traceptive use than second and third generation adolescents. 19,28,203 Segmented assimilation theory suggests that traditional cultural values such as the importance of family or conservative sexual values, such as the importance of virginity, can buffer second generation adolescents from the negative influence of poverty and deviant peer behavior often found in immigrant communites. 57 This theory has been supported for outcomes such as dropping out of high sch ool, drug use, and adolescent birth. 88,122,124,204

PAGE 103

87 However, the same traditional cultural values that may protect against adverse outcomes such as sexual initiation, as discussed in the previous section, may decrease the odds of contraceptive use. The importance of fam ily ( familismo ) 109 and the importance of virginity, 20,109,143,145,153,154 are strongly linked with the Catholic Church which also discourages contraception and promotes motherhood and childbearing 153 While traditional cultural values have been measured qualitatively or less often quantitatively through surveys, 20,109,143,145,153,154 they are usually examined in relationship to the health outcome, not in relationship to independent variables such as family structure and SES. However, I can postulate from previous research that Mexican origin immigrant families with two married parents 57 and grandparents, 104,105 and in less educated and underemployed migrant families, may ascribe to more traditional, conservative, or patriarc hal values. 139,140 As often seen in the immigrant paradox literature, as time in the U .S. increases, families may replace traditional cultural values and behaviors with the less protective values and behaviors of disadvantaged U.S. communities increasing risk for downward economic assimilation and adverse health outcomes. 9,31,87 Although families with two biologi cal parents may be more likely to promote traditional values discouraging contraception, previous research shows that having two biological parents in the home increases the likelihood of using contraception in all race ethnic groups. 97,191,202 I hypothesize t hat intact families including two biological or adoptive parents, will increase the odds of contraceptive use in all immigrant generations for similar reasons having two biological parents reduced the hazard of sexual initiation; two biological parents pro vide stability, supervision and emotional support. Although the

PAGE 104

88 presence of a grandparent in the home also provides more supervision and emotional support, and having a grandparent in the home also reduce s the hazard of sexual initiation, I hypothesized th at in first and second generation families a grandparent in the home would decrease the odds of contraceptive use because reinforcement of traditional and conservative cultural values that discourage the use of contraception 105 may outweigh the protection of additional support and supervision. Immigrant adolesc ents have more siblings than native born adolescents 106 which may reflect traditional va lues related to motherhood and result in less parental supervision, 111 both of which may decrease odds of contraceptive use in first generation adolescents. The presence of an older sibling in the home has been shown to increase risky sexual behavior for younger siblings 111,112 and because siblings in Mexican origin families spend more time together than non Hispanic white siblings, they may be even more influential younger siblings may be less likely to use contraception. 107,200 Alternatively, older siblin gs may educate their younger siblings about the importance of contraception. Prior research found that having an older sibling reduced the hazard of sexual initiation in Mexican origin adolescents, 205 thus older siblings may be protective in Mexican origin families. Figure 3.6 shows my hypotheses for the interactio ns between family structure variables and immigrant generation using estimated hazard ratios.

PAGE 105

89 Figure 3 6 : Hypotheses for F amily S tructure I nteractions Lower parental education and lower SES decreases odds o f contraceptive use. 31,118,186,206 Migran t parents with less education and unskilled employment may encourage traditional cultural values, 139,140 emphasizing the importance of virginity and discouraging the use of contraception. The traditional values reinforced by undereducated and underemployed migrant parents may compound the lower SES of migrant families as compared to native born families resulting in lower odds for contraceptive use in first and second generation families, particularly with low parental SES. Figure 3.7 shows my hypotheses for the interactions between family SES variables and immigrant generation using estimat ed hazard ratios. Two bio parents First Second Third Immigrant generation OR 1.5 OR 1.5 OR 1.2 Grandparent OR 0.8 OR 0.8 OR 1.2 First Second Third Number of siblings First First Second Second Third Third OR 0.8 OR 1.2 OR 1.2 OR 0.6 OR 1 OR 0.8 Older sibling Odds of contraceptive use Odds of contraceptive use Odds of contraceptive use Odds of contraceptive use

PAGE 106

90 education Odds of contraceptive use First Second Third Immigrant generation OR 1.5 OR 1.5 OR 1.4 education OR 1.5 OR 1.5 OR 1.4 First Second Third Mother not working F irst First Second Second Third Third OR 0.6 OR 0.8 OR 0.8 OR 0.6 OR 0.8 OR 0.9 Unskilled mother Professional mother First Second Third OR 1.4 OR 1.4 OR 1.2 Not working father First Second Third OR 0.6 OR 0.6 OR 0.6 OR 0.6 OR 0.8 OR 0.8 Unskilled father Professional father First Second Third OR 1.4 OR 1.4 OR 1.2 Odds of contraceptive use Odds of contraceptive use Odds of contraceptive use Odds of contraceptive use Odds of contraceptive use Odds of contraceptive use Odds of contraceptive use First Second Third Figure 3 7 : Hypotheses for Family SES Interactions

PAGE 107

91 Methods I used data from all four waves (1994 2008) of The National Longitudinal Study of Adolescent Health (Add Health). The survey began in 1994 with 20,475 7 12 graders and participants were interviewed again in 1996 (Wave II), 2001 2002 (Wave III), and 2007 2008 (Wave IV). 160 I restricted the sample to Mexican origin respondents. I excluded 212 participants who had not had sexual intercourse by Wave III and thus were not asked whether they used contraception at their first sex. I excluded five participants who were missing on ever having sex, six participants missing on the presence of a parent in the home, and seven participants with an age at first sex pri or to 11 due to the unlikelihood that sex was consensual at that age. 165 After excluding respondents who did not have sample weight the final sample size was 1,425 Mexican origin participants. Those who had their first sex prior to Wave I were not excluded because it removes participants with the earlie st ages at firs t sex and removes unequal numbers of males and third generation adolescents. 164 Both adolescent childbearing and unmarried childbearing are of concern, however, I did not exclude any participants based on age at first sex. Seventy percent of participants had their first sex prior to age 20 and because contraceptive use at first sex was not asked in Wave IV, the average age of participants in Wave III was 23. While the small number of participants who had their first sex between 20 and 23 were no longer at risk for an adolescent birth, they are still at risk for relatively early childbearing and it is worth examining their contraceptive use patterns. Variables. Contraceptive use at first sex was asked in Waves I III as a single

PAGE 108

92 relationship in provided in the relationship history section of each wave and some participants had up to nine different relationships. As Schoen found that pregnancies and births were more likely to be missed when recalling them within a chronological relationship history 166 than just bei ng asked to list the dates of pregnancies and births, I measured contraceptive use at first sex from the single direct question asked once in Waves I III and I only used data asked from relationship history if the adolescent was missing on the single direc t question. Although the question did not specify the meaning of birth control and some adolescents may assume hormonal birth control, the very next question asked what type of birth control he/she used and included non hormonal forms such as condoms so so me may have been able to change their answer if they saw that birth control included hormonal. It is possible that birth control use was underestimated if adolescents only thought about hormonal forms as birth control. Table 3. 10 presents the independent variables and how they were measured. All independent variables were measured at Wave I. Table 3 10 : Independent V ariables Independent variables Measured Type of variable Immigrant generation First=Born i n Mexico to Mexican born parents Second=Born in the U.S. to at least one Mexican born parent Third=Born in the U.S. to two U.S. born parents and classifies ethnicity as Mexican 167,198 Categorical Family structure Two biological parents Partici pant lived with two biological or two adoptive parents for his/her entire life Dichotomous Presence of a grandparent in the home Participant lived with a grandparent or great grandparent in the home at Wave I Dichotomous Number of siblings in the home The number of biological, half, or step siblings in the home at Wave I Continuous Has an older sibling Participant has an older full biological sibling who may or may not be in the home at Wave I. Dichotomous Family SES

PAGE 109

93 Table 3 10 : Independent V ariables Number of years of education: 0= never attended school 8= 8 th grade or less 10= no high school degree but more than 8 th grade 12=GED or diploma 16=college graduate her Continuous As above Continuous 1=not working outside of the home 2=unskilled employed 199 3=skilled/professionally employed* Categorical As above Categorical Control variable s Age Age in years at Wave I Continuous Age at first sex Age at first sexual intercourse Continuous Gender ~ 1=male 0=female Dichotomous Family income Logged annual household income, adjusted for inflation and family size at Wave I Continuous Residen t father Presence of a father figure (biological, step) in the home Dichotomous *Because of the small sample size in first generation mothers for skilled (n=50) and professional employment (n=25), they were combined for all generations. ~I controlled for gender instead of estimating stratified models because there were no significant interactions between immigrant generation and gender. I controlled for age at Wave I interview because older participants may be more subject to recall bias and younger part icipants may be more likely to select socially expected responses. I also controlled for age at first sex because age is associated with contraceptive use; while older women are more likely to use contraception than younger women, 207 younger adolescents may be more consistent users than older adolescents, possibly due to longer relationships. 202,208 I controlled for gender instead of stratifying because there were no significant interactions between immigrant generation and gender and thus stratifying only reduced power in the mo dels. I also controlled for the presence families included mothers. I did not control for marital status at first sex due to the rarity of marriage prior to first sex; bet ween seven and 13% of Mexican origin participants

PAGE 110

94 were married prior to the age of 20 and many were married after their first sex but prior to their first births. Although children of a married adolescents birth may have slightly improved outcomes than unm arried adolescent births, due to financial resources and support, they are still of concern. 209 Analysis. I used logistic regression with the dichotomous variable contraceptive use at first sex as the outcome. Residual diagnost ics were performed and there were no substantial outliers and deviance residuals were within +/ 1.5 and the sum of squares deviance residual were within +/ 2.5 which are acceptable. 172 I first examined main effects of each family structure and family SES indepen dent variables and then I interacted immigration generation with each predictor in order to understand how the relationship between the odds of contraceptive use and the independent family structure or family SES variable differs by immigrant generation. B ecause grandparents are more often present in single parent families, 101 I controlled for two biological par ents in models with the grandparent variable to reduce confounding. About 26% of participants had no employment models, only participants with that particular parent in the home was included. Because more educated parents tend to have more skilled or professional employment, I added an education variable in models that included an employment variable and I added an employment variable in models that included an education vari able to reduce confounding. Missing data. For missing independent variables, family income had the highest missing at 37%, parental education and employment were missing less than 5%, and all other variables were missing less than 1%. Almost 18% of the out come variable,

PAGE 111

95 contraceptive use at first sex, is missing. Theoretically, missing data on the outcome variable is more concerning. However, methodologically it does not matter if the independent or dependent variables are missing and the dependent variable should be in the imputation model. 210 I used logistic regression to predict those who are missing on cont raceptive use at first sex and found they were more likely to be younger at Wave I (p<0.001), have an older age at first sex (p<0.001), be male (p=0.046), third generation (p<0.02), and have fathers with less years of education (p=0.011). I addressed the missingness of the outcome variable in three ways. First, I ran the models using list answered the question, limiting generalizability. List wise deletion also assumes that data is miss ing completely at random (MCAR) which was disproved with the models predicting those who are missing. Second, I built and rebuilt prediction models to impute the outcome variable in each generation to ensure the pseduoR 2 was at least 0.15 0.2 and the corre lation to the other variables in the models were maintained. I then used the same ten cycles of the multiply imputed data sets with chained equations to model contraception use at first sex and only used the imputed data for the predictor variables droppin g those missing on the outcome variable. Finally, I used the same ten cycles of the multiple imputed data sets to model contraceptive use at first sex imputing all variables, including the outcome variable. These three sets of models allowed me to find out if the reason relationships differed between the list wise deleted outcome model and the fully imputed model were related to poor prediction of the predictor or outcome variables. I presented the fully imputed models, however, all three sets of models are in the appendix.

PAGE 112

96 Additionally, I conducted the following sensitivity analyses. When stratifying by those who had sex prior to age 20 and those who had sex after age 20, I found no substantial differences, although as number of siblings increased in the t hird generation, the odds for contraceptive use increased. I did not find this effect in the first and or second generation though there is less stability in the estimates due to small sample sizes. After stratifying by immigrant generation, there were no substantial differences although the sample size in each generation was so small, there were few significant relationships. In models stratified by gender, having a father not working versus a professionally employed father increased the odds for contracep tive use the first and second generation although stability is poor due to small sample sizes. Having a father not working versus a professional father decreased the odds of contraceptive use in first generation females although having father not working i n the first gener ation is a very rare occurrence, reducing stability of the models. Results I presented the results of the fully imputed models because the results were generally consistent across all three imputation strategies and imputation is a more v alid way to handle missing data that is not MCAR. The relationships that I was not able to recover using the fully imputed models were unstable or significant at the p<0.10 level in the list wise deletion model (See appendix for all three groups of models) Table 3. 11 presents descriptive data. Over 60% of the sample used contraception at their first sex. Those who used contraception at their first sex were more likely than those who did not use contraception to live with two biological parents, have a gra ndparent in the home, have higher educated parents, a higher family income, and a

PAGE 113

97 later age at sexual initiation. Although second generation adolescents are more likely to live with two biological parents, we see that parental education, employment and inc ome improve across immigrant generations. Third generation adolescents have the earliest age at first sex and are more likely to have used contraception at their first sex than first or second generation adolescents. Table 3 11 : Weighted Means (Standard Deviations) and P roportions of Select V ariables Contraception at first sex No contraception at first sex First generation Second generation Third generation N= 879 546 241 611 573 Variable Contracep tion at first sex 47.5% 61.9% 67.1% Male 52.4% 54.1% 51.4% 52.9% 53.9% Two biological or adoptive parents 57.9% 43.0% 46.9% 69.1% 41.9% Presence of a grandparent 5.5% 2.8% 1.3% 5.2% 5.3% Old sibling 61.0% 62.0% 67.5% 64.0% 56.8% Number of siblings 2 (1.6) 2 (1.4) 2.4 (1.5) 2.4 (1.5) 1.6 (1.4) Resident father in home 73.3% 70.6% 74.1% 83.0% 63.5% Mother's years of education 11.3 (2.9) 10.5 (2.8) 9 (2.9) 10 (2.6) 12.5 (2.5) Father's years of education 11.2 (3.2) 10.5 (3.1) 8.6 (3.3) 10.3 (2.7) 12.7 (2.6) Mother unemployed 42.8% 40.2% 47.9% 46.5% 35.7% Mother unskilled 24.7% 31.1% 40.7% 33.0% 16.7% Mother skilled/professional 32.6% 28.8% 11.4% 20.1% 47.6% Father unemployed 12.3% 13.0% 13.4% 14.3% 10.4% Father unskilled 44.9% 48.6% 58.7% 47.1% 39. 3% Father skilled/professional 42.9% 38.4% 27.9% 38.6% 50.3% Family income $32,000 (3,590) $25,300 (2,290) $17,500 (1,540) $26,900 (3,700) $36,400 (3,070) Age at first sex 15.9 (1.9) 16.2 (2) 16.6 (2.2) 16.2 (1.7) 15.6 (2.1) Wave I Age 16.2 (2.5) 15.9 (2.5) 16.6 (2.8) 16.6 (2.4) 15.5 (2.1)

PAGE 114

98 Family structure: Parents and grandparents Table 3. 12 Model 1 shows that the odds of contraceptive use at first sex were 56% lower (1 0.43) in first generation adolescents as compared to third generation adolesc ents. The odds of contraceptives use at first sex in the second generation were not significantly different than the third generation. A post estimation t test showed that difference between the odds of contraceptive use in the first generation as compared to the second generation was significant at p=0.052. Model 2 shows that the odds for contraceptive use for adolescents with two biological parents were roughly 1.8 times the odds for adolescents who have not lived with two biological parents for their who le lives. The odds ratio for the second generation became further from one and lost significance with the addition of two biological parents, suggesting two biological parents suppresses the relationship between the odds of contraceptive use in the second and third generation. Model 3 shows that there was not a significant interaction between two biological parents and immigrant generation. Model 4 shows that adolescents with a grandparent in the home were twice as likely to use contraception at their first sex as adolescents without a grandparent in the home (p=0.055). Model 5 shows there was no significant interaction for having a grandparent in the home and immigrant generation. Table 3 12 : Odds R atios (95% Confidence I ntervals) for Contraceptive Use at First Sex and Family S tructure Model 1 Model 2 Model 3 Model 4 Model 5 First gen 0.43* 0.41* 0.45 0.42* 0.42* (0.20 0.92) (0.19 0.90) (0.20 1.01) (0.19 0.92) (0.19 0.93) Second gen 0.75 0.64 0.80 0.64 0.62* (0.49 1.15) (0.41 1.02) (0.46 1.40) (0.40 1.01) (0.38 0.99) Third gen Ref. Ref. Ref. Ref. Ref. Two biological parents 1.84*** 2.24** 1.88*** 1.88*** (1.38 2.46) (1.37 3.66) (1.40 2.51) (1.40 2.52)

PAGE 115

99 Table 3 12 : Odds R atios (95% Confidence I ntervals) for Contraceptive Use at First Sex and Family S tructure First ge n X two bio par 0.79 (0.37 1.70) Second gen X two bio par 0.66 (0.33 1.32) Third gen X two bio par Ref. Grandparents 2.03 1.55 (0.99 4.17) (0.56 4.30) First gen X grandparents 0.82 (0.05 14.01 ) Second gen X grandparents 2.11 (0.45 9.80) Third gen X grandparents Ref. Age at Wave I 0.93 0.94 0.95 0.95 0.95 (0.82 1.04) (0.84 1.07) (0.84 1.07) (0.84 1.07) (0.84 1.07) Age at first sex 1.09* 1.06 1.06 1.06 1.06 (1.01 1.18) (0.98 1.15) (0.98 1.15) (0.98 1.15) (0.98 1.15) Male 1.00 0.96 0.97 0.97 0.97 (0.75 1.33) (0.71 1.30) (0.72 1.31) (0.71 1.32) (0.71 1.32) Constant 1.79 1.58 1.45 1.40 1.44 (0.20 16.37) (0.16 15.38) (0.15 14.34) (0.15 13.12) (0.15 13.33) Observations 1,425 1,425 1,425 1,425 1,425 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Family structure: Siblings Table 3. 13 Models 1 and 2 show that the relationship between the number of siblings in the home and the odd s of contraceptive use at first sex were not significantly different from 1 and there was no significant interaction between the number of siblings in the home and immigrant generation. Model 3 shows the relationship between having an older sibling in the home and the odds of contraceptive use was not significantly different than one. However, Model 4 shows the odds of contraceptive use in second

PAGE 116

100 generation adolescents with an older sibling in the home were 1.6 times higher than third generation adolescents with an older sibling in the home (2.27*0.71) (p=0.068). Table 3 13 : Odds R atios (95% C onfidence I ntervals) for Contraceptive Use at First Sex and Family Structure Model 1 Model 2 Model 3 Model 4 First gen eration 0.41* 0.43* (0.18 0.89) (0.16 1.16) (0.20 0.93) (0.16 1.02) Second generation 0.71 0.58 0.75 0.46* (0.47 1.08) (0.26 1.29) (0.49 1.16) (0.21 1.00) Third generation Ref. Ref. Ref. Ref. Number of siblings 1.0 7 1.04 (0.96 1.20) (0.84 1.29) First gen X siblings 0.99 (0.75 1.31) Second gen X siblings 1.10 (0.79 1.54) Third gen X siblings Ref. Older biological sibling 0.98 0.71 (0.68 1.39) (0.44 1.15) First ge n X older sibling 1.16 (0.49 2.78) Second gen X older sibling (0.94 5.46) Third gen X older sibling Ref. Age at Wave I 0.93 0.93 0.93 0.93 (0.83 1.05) (0.83 1.04) (0.82 1.04) (0.82 1.04) Age at first sex 1.0 9* 1.09* 1.09* 1.09* (1.00 1.17) (1.00 1.17) (1.01 1.18) (1.01 1.18) Male 1.00 1.00 1.00 1.02 (0.75 1.33) (0.75 1.34) (0.75 1.33) (0.76 1.36) Constant 1.63 1.72 1.81 2.03 (0.18 15.07) (0.20 14.94) (0.19 17.31) (0.21 20.16) Observations 1,425 1,425 1,425 1,425 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Family SES: Parental education Table 3. 14 presents odds ratios for contraceptive use at sexual initiation and parental education. Model 1 shows that a one year increase in

PAGE 117

101 increased the odds of contraceptive use at first sex by 6% ([1 1.06]*100) though just beyond significance (p=0.11). Model 2 shows there was no significant interaction ow that the education and immigrant generation. However, the odds ratio for fathe first generation immigrants was 1.11 and just beyond significance (p=0.195), suggesting first generation adolescents as compared to third generation adolescents. Table 3 14 : Odds R atios (95% C onfidence I ntervals) for Contraceptive Use at First Sex and Parental Education Model 1 Model 2 Model 3 Model 4 First generation 0.55 0.54 0.61 0.22 (0.23 1.2 9) (0.08 3.53) (0.24 1.51) (0.04 1.36) Second generation 0.84 0.98 0.92 0.76 (0.52 1.38) (0.15 6.54) (0.51 1.66) (0.10 5.60) Third generation Ref. Ref. Ref. Ref. Mother's education 1.06 1.06 (0.99 1.13) (0.96 1.17) First gen X mother's education 1.00 (0.84 1.20) Second gen X mother's education 0.99 (0.83 1.16) Third gen X mother's education Ref. Father's education 1.03 0.99 (0.96 1.11) (0.89 1.11) First gen X father's educa tion 1.11 (0.95 1.29) Second gen X father's education 1.01 (0.87 1.17) Third gen X father's education Ref. Not working mother (0.93 2.55) (0.94 2.54) Unskilled mother 1.10 1.10 (0.64 1.91) (0.65 1.87)

PAGE 118

102 Table 3 14 : Odds R atios (95% C onfidence I ntervals) for Contraceptive Use at First Sex and Parental Education Skilled/Professional mother Ref. Ref. Not working father 1.12 1.08 (0.52 2.41) (0.51 2.29) Unskilled father 1.01 0.99 (0.57 1.78) (0.57 1.71) Skilled/Professional father Ref. Ref. Age at Wave I 0.92 0.92 0 .91 0.91 (0.81 1.05) (0.81 1.05) (0.80 1.03) (0.80 1.03) Age at first sex 1.08 1.08 (0.99 1.18) (0.99 1.18) (0.97 1.19) (0.98 1.19) Male 1.07 1.07 1.21 1.21 (0.77 1.47) (0.77 1.48) (0.85 1.72) (0.85 1.73) Family Income 1.27 1.27 1.26 1.28 (0.94 1.72) (0.94 1.72) (0.92 1.71) (0.93 1.77) Resident Father 1.12 1.12 (0.77 1.63) (0.77 1.63) Constant 0.41 0.39 1.02 1.48 (0.02 7.01) (0.02 7.30) (0.06 18.06) (0.10 22.96) Observations 1,333 1,333 1,047 1,047 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Family SES: Parental employment Table 3.1 5 presents odds ratios for contraceptive use at first sex and parental employment. Model 1 shows that the odds of contraceptive use were 1.54 times h igher (1.54*1) for adolescents with a mother not working as compared to a skilled or professionally employed mother (p=0.093). Model 2 shows there was no significant Models contraceptive use at first sex was not significantly different than 1 and there was no

PAGE 119

103 Table 3 15 : O dds R atios (95% C onfidence I ntervals) for Contraceptive Use at First Sex and Parental Employment Model 1 Model 2 Model 3 Model 4 First generation 0.55 0.68 0.61 0.39 (0.23 1.29) (0.18 2.61) (0.24 1.51) (0.10 1.44) Second generation 0.84 1.05 0.92 0.99 (0.52 1.38) (0.44 2.49) (0.51 1.66) (0.42 2.31) Third generation Ref. Ref. Ref. Ref. Not working Mother 1.54~ 1.71 (0.93 2.55) (0.86 3.40) Unskilled Mother 1.10 1.29 (0.64 1.91) (0.59 2.85) Skilled/professional Mother Ref. Ref. First gen X not working mother 0.61 (0.12 3.16) First gen X unskilled Mother 0.86 (0.17 4.28) First gen X skilled/profess mother Ref. Second gen X unemployed mother 0.82 (0.27 2.54) Second gen X unskilled Mother 0.59 (0.18 1.97) Second gen X skilled/profess mother Ref. Not working Father 1.12 0.68 (0.52 2.41) (0.20 2.31) Unskilled Father 1.01 1.08 (0.57 1.78) (0.45 2.60) Professional Father Ref. Ref. First gen X not working father 2.41 (0.48 12.11) First gen X unskilled father 1.70 (0.48 5.96) First gen X skilled/profess father 1.00 (1.00 1.00) Secon d gen X not working father 2.15 (0.58 7.98) Second gen X unskilled father 0.68 (0.23 2.01) Second gen X skilled/profess father Ref.

PAGE 120

104 Table 3 15 : O dds R atios (95% C onfidence I ntervals) for Contraceptive Use at First Sex and Parental Employment Mother's education 1.06 1.05 (0.99 1.13) (0.98 1.12) Father's education 1.03 1 .03 (0.96 1.11) (0.96 1.11) Age at wave I 0.92 0.92 0.91 0.91 (0.81 1.05) (0.81 1.05) (0.80 1.03) (0.80 1.03) Age at first sex 1.08 1.08 (0.99 1.18) (0.99 1.18) (0.97 1.19) (0.97 1.20) Male 1.07 1.07 1.21 1.19 ( 0.77 1.47) (0.77 1.48) (0.85 1.72) (0.82 1.72) Family Income 1.27 1.28 1.18 1.18 (0.94 1.72) (0.94 1.73) (0.89 1.58) (0.88 1.59) Resident Father 1.12 1.11 (0.77 1.63) (0.76 1.61) Constant 0.41 0.39 1.02 1.04 (0.02 7.01) ( 0.02 6.40) (0.06 18.06) (0.05 19.85) Observations 1,333 1,333 1,047 1,047 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Discussion This study is the first to examine how family structure and family SES influences the odds of contraceptive use at fi rst sex differently across immigrant generations of Mexican origin adolescents using a nationally representative longitudinal study. The results contribute to the literature in three main ways. First, advancing previous literature, 19,28,203 I found that second generation adolescents had similar odds of contraceptive use at first sex as third generation adolescents. This means that after controlling for age at first sex and gender, despite having foreign born pare nts as first generation adolescents do, second generation adolescents are more similar to third generation adolescents when it comes to contraceptive use. Second generation adolescents were more likely to have two biological parents in the home which may b e one of the reasons the odds of contraceptive use

PAGE 121

105 between the second and third generation became significant with the addition of two biological parents. Supporting previous literature I found that first generation adolescents were significantly less like ly to have used contraception at first sex than second and third generation adolescents. These findings extend segmented assimilation theory because traditional cultural values may not always be protective; the odds of contraceptive use increased with gene rations in the U.S. possibly eroding traditional cultural values that discourage contraceptive use. The socioeconomic advancement across generations could be another reason contraceptive use increased between the first and third generation, in addition to increased access to contraception and availability of comprehensive sex education in some U.S. schools. Table 3. 16 summarizes the family structure models and support for segmented assimilation theory. Supporting my hypothesis and segmented assimilation the ory, having two biological parents in the home increased the odds of contraceptive use across all three generations. I extended previous research and segmented assimilation theory by examining the influence of grandparents and siblings on contraceptive use at first sex in Mexican origin adolescents. Contrary to my hypothesis, my results suggest that having a grandparent increased the odds of contraceptive use equally in all immigrant generations. Only about 4% of adolescents in the sample lived with a grand parent at Wave I which could account for marginal level of significance. Perhaps the addition of an adult for supervision, emotional support, and economi c resources, outweigh conservative sexual values discouraging contraception use. Or perhaps grandparen ts care more about protecting their grandchildren from early childbearing, even if it conflicts with their own religious or traditional values. Future research, particularly with a larger sample size of

PAGE 122

106 adolescents with grandparents, could provide a cleare r understanding of how My results challenge previous research in a multiracial sample that found older siblings increased risky sexual behavior in younger siblings. 111,112 My results show the presence of an older sibling in the home increased the odds of contraceptive use in the second generation compared to the third generation. Similarly, the sexual initiation analysis found t hat the presence of an older sibling decreased the hazard of sexual initiation in all immigrant generations equally. Perhaps the reason my results are different than those in multiracial studies are because the relationship between siblings works different ly in Mexican origin families, particularly in immigrant families. Siblings in Mexican origin families spend more time together than non Latino white siblings 107,200 which could symbolize closer relationships between siblings which have been found to have lower sexual risk behavior. 113 Further, older siblings tend to have more domestic responsibility in the house tha n younger siblings, including responsibility to care for younger siblings. 108,144,145 Older siblings may spend more time with, be closer to, and feel more responsible for their younger siblings, resulting in more guidance and sexual education for younger siblings. The influence of siblings on reproduc tive health outcomes specifically in Mexican origin and immigrant families provides a new area of future research.

PAGE 123

107 Table 3 16 : Summary of Family S tructure a nd the Odds of Contraceptive U se at First Sex Va riable OR Supported hypothesis? Supported SAT? Two bio parents 1.84*** (1.38 2.46) Yes: No significant interaction by immigrant generation but two biological parents increased the odds of contraceptive use across all generations. Two parents suppressed the relationship between the second and third generation. Supports SAT Grandparents 2.03 (0.99 4.17) No: Grandparents increased the odds of contraceptive use across immigrant generations Extends SAT, no prior mention of grandparents although extending the theory they would decrease odds of contraceptive use due to conservative sexual values Number of siblings 1.07 (0.96 1.20) No: No significant main effect or interaction Neither supports nor rejects. Siblings not mentioned in SAT. Older sibling 0.71 (0.44 1.15) Partial: Significant interaction. Older siblings protective in the second generation as compared to the third generation. Extends SAT. No prior mention on siblings although migrant families would have closer relationships and older siblings m ay encourage contraceptive use. First gen x older sibling 1.16 (0.49 2.78) Second gen x older sibling (0.94 5.46) Third gen x older sibling Ref. *** p<0.001, ** p<0.01, p<0.05, p<0.10 Table 3. 17 summarizes the parental education models and support for segmented contraceptive use equally in all generations, although the odds ratio approached significance (p=0.11). If significant, this would have supported my hypothesis and segmented assimilation theory that asserts parental education protects against risky adolescent behaviors. Table 3 17 : Summary of Parental Education a nd the Odds of Contraceptive U se at Firs t Sex Variable OR Supported hypothesis? Supported SAT? Mother's education 1.06 (0.96 1.17) No. No significant main effect nor interaction for Neither supports nor rejects but model approaching significance approaches First gen X mother's education 1.00 (0.84 1.20)

PAGE 124

108 Table 3 17 : Summary of Parental Education a nd the Odds of Contraceptive U se at Firs t Sex Second gen X mother's education 0.99 (0.8 3 1.16) the odds of contraceptive use approached significanc e at (p=0.11). education decreasing risky adolescent behavior. Third gen X mother's education Ref. Father's education 0.99 (0.89 1.11) No. No significant main effect nor interaction for Neither supports nor rejects. First gen X father's education 1.11 (0.95 1.29) Second gen X father's education 1.01 (0.87 1.17) Third gen X father's education Ref. *** p<0.001, ** p<0.01, p<0.05, p<0.10 Table 3. 18 summarizes parental employment models and support for segmented assimilation theory. Extending previous research and s egmented assimilation theory, I found that adolescents with a mother not working were more likely to use contraception at their first sex as compared to adolescents with skilled or professionally employed mothers. This is somewhat counterintuitive since hi gher educated mothers tend to be more professionally employed. Similar to my sexual initiation analyses, a mother not working is likely providing more supervision and support than mothers who work outside the home. Table 3 18 : Summary of Parental Employment a nd the Odds of Contraceptive U se at First Sex Variable OR Supported hypothesis? Supported SAT? Not working Mother 1.54 (0.93 2.55) Partial: Having a mother not working increased the odds of contracep tive use across all three generations. No significant interaction. Rejects SAT that asserts skilled and professional parents improve second generation outcomes. Extends SAT by showing skilled and professional parents are not always protective for all out comes. Unskilled Mother 1.10 (0.64 1.91) Professional Mother Ref. Not working father 1.12 (0.52 2.41) No. There was no main effect nor significant employment. Neither rejects nor supports. Unskilled father 1.01 (0. 57 1.78) Professional father Ref. *** p<0.001, ** p<0.01, p<0.05, p<0.10

PAGE 125

109 My study is the first to use longitudinal nationally represented data and examine interactions to examine how family structure and family SES influence the odds of contr aceptive use differently across immigrant generations. However, there are several limitations to my study. First, the large amount of missing data on the outcome variable wh ich approached significance in the imputed predictor models, the findings I reported were consistently found in the list wise deletion models, imputed predictor models and the fully imputed imputation models increasing validity of findings. Further, I cond ucted sensitivity analyses stratifying by gender, immigrant generation, and first sex prior to age 20 and found no substantial differences. Second, I did not have the ability to measure traditional values directly or sibling or grandparent support. However the use of segmented assimilatio n theory and previous research support the conclusions I have drawn between family structure, f amily SES and traditional or conservative sexual values Third, due to sample size I was unable to test for differences in age at migration within the first generation, although previous literature found that age of immigration influences outcomes. 21,167,169 My sensitivity analysis using a six category age at migration variable, 169 found the odds of contraceptive use increased as age of migration decreased (see Appendix Table A.1), supporting previous literature. Finally, recall bias can affect the internal validity of the study, although the longitudinal design of the stu dy decreases recall bias. Conclusion I have advanced current literature by finding that family factors influenced reproductive health of Mexican origin adolescents differently across immigrant

PAGE 126

110 generations. This study challenges current literature on the in fluence of older siblings on risky sexual behaviors of younger siblings and advances research by showing that having a mother not working increases the odds of contraceptive use. Future research could provide a better understanding of how grandparents and older siblings may increase the odds of contraceptive use at first sex in Mexican origin adolescents and how parental education and employment influence first, second, and third generation adolescents differently, providing an important first step to ident ifying ways to reduce reproductive health disparities in Mexican origin adolescents. A ge at First Birth Background By 2020, Latino adolescents will comprise 25% of the adolescent population in the U.S. 6 Although they have experienced one of the more rapid declines the last two decades, Latinos still have the highest adolescent birth rate of any race/ethnic group. 17 Mexican origin Latinos make up 64% of the Latino population 55 and they have a higher birth rate than Puerto Rican and Cuban adolescents, 18 making it imperative to understand reproductive health disparities among Mexican or igin adolescents. Prior research has shown that age at sexual initiation is later in first and second generation immigrants than the third generation resulting in a decreased risk for an adolescent birth. 21,28,23,29,205 However, contraceptive use is often lower among first and second gen eration adolescents increasing risk for an adolescent birth. 19,28,203 Despite these opposing risk factors, some studies found that native born Latinas have an increased risk for adolescent birth than foreign born Latinas and that in the foreign born, risk for adolescent birth increases as time in the U.S. and acculturation increases. 24,29,115

PAGE 127

111 However, two studies found speaking Spanish at home, generally a proxy for less acculturation, was associated with an increased risk in pregnancy in Latinas, 98,115 demonstrating the lack of consensus in the literature on immigrant generation and risk for adolescent birth. Segmented assimilation theory suggests that traditional family values such as the importance of family or conservative sexual values, such as the importance of virginity, promote economic assimilati on in second generation adolescents by protecting them from risky behaviors, including adolescent childbearing. 57 Although traditional values found in qualitative studies, such as the importance of virginity, 20,109,143,145,153 are often used as a driver of reproductive health outcomes in Latinos, they are often speculative, not measured directly. 21,22,31 However, we can deduce from previous research that Mexican origin immigrant families with two married parents 57 and grandparents, 104 and in families where migrant parents have less education and work in unskilled jobs may be more likely to adhere to traditional conservative, or patriarchal cultural values. 139,140 As exposure to U.S. communities increases, families may replace traditional cultural values and behaviors with less protective values and behaviors of impoverished communities where migrants tend to re side, increasing risk for downward assimilation and risky health behaviors. 9,31,87 Living with two biological parents decreases the risk for adolescent birth in Latinas, through improved economic resources in the home and providing emotional support and supervision. 31,82 D rawing on segmented assimilation theory, I hypothesize that traditional and intact families that include two biological parents, grandparents, and more siblings, may be more protective against an adolescent birth for first and second

PAGE 128

112 generation adolescents than for third generation immigrants. Foreign born families tend to be of lower SES than third generation families and yet may have a lower risk for an adolescent birth, perhaps because in addition to parental monitoring and support, migrant parents and g randparents may be more likely to reinforce traditional cultural values such as the importance of virginity until marriage and discourage childbearing out of wedlock. 105,109,145,153 Mexican origin migrant families generally have more children than native born families which may reflect stronger traditional values related to child bearing and motherhood, increasing risk for married adolescent childbearing but decreasing risk for unmarried adolescent childbearing 91,106 Adolescents with an older si bling are more likely to engage in risky behaviors, including sexual initiation, 111,112 more likely to have an adolescent birth if their older sister had an adolescent birth, 38 and younger siblings receive less parental monitoring, 111,112 an adolescent birth. Converse ly, siblings with close relationships with older siblings were found to have lower sexual risk behavior. 113 Furthermore, prior research has shown that having an older sibling in the home protected adolescents from early sexual initiation 205 and an older sibling was associated with an increased odds of contraceptive use for the second generation, 211 suggesting that older siblings in Mexican origin families ma y protect their younger siblings from risky sexual behavior, including adolescent childbearing. Figure 3.8 shows my hypotheses for the interactions between family structure variables and immigrant generation using estimated hazard ratios.

PAGE 129

113 Figure 3 8 : Hypotheses for Family Structure and Immigrant Generation Interaction Higher SES is associated with a lower risk of an adolescent birth 31,115,118 and parents with more education and who work in more skilled occupations may improve the future economic wellbeing of their children through modeling and academic expectations which may delay childbearing. 87 However, migrant families with less educated parents or parents who work in unskilled occupations may be more protective against an adolescent birth in first and second generation families than in third generation families because low SES migrant families may be more likely to maintain traditional cultural values, 139,140 such as virginity, and have stricter dating rules and more parental supervision, especially in families with a mother not working outside the home. Figure 3.9 shows my hypothese s Two bio parents First Second Third Immigrant generation HR 0.5 HR 0.5 HR 0.8 Grandparent HR 0.5 HR 0.5 HR 0.8 First Second Third Number of siblings First First Second Second Third Third HR 0.5 HR 0.8 HR 0.8 HR 0.5 HR 0.8 HR 1.2 Older sibling Hazard of an adolescent birth Hazard of an adolescent birth Hazard of an adolescent bir th Hazard of an adolescent birth

PAGE 130

114 for the interactions between family SES variables and immigrant generation using estimated hazard ratios.

PAGE 131

115 Figure 3 9 : Hypotheses for F amily SES I nteractions

PAGE 132

116 Methods I used data from all fou r waves (1994 2008) of The National Longitudinal Study of Adolescent Health (Add Health). The survey began in 1994 with 20,475 7 12 graders and participants were interviewed again in 1996 (Wave II), 2001 2002 (Wave III), and 2007 2008 (Wave IV). 160 I restricted the sample to Mexican origin respondents. I excluded six participants missing on the presence of a parent in the home, seven participants with an age at first sex prior to 11 due to the unlikelihood that sex was consensual at that age, and six participants who were missing on the outcome of their pregnancy. 165 No first births occurred prior to Wave I among Mexican origin participants. After excluding respondents who were missing without a sample weight, the final sample size was 1,636 M exican origin participants. Variables. The outcome is age at first birth. The respondent was asked if she/he had ever been pregnant or gotten someone pregnant, when the pregnancy ended, and about the outcome of the pregnancy. Age at first birth was calcul ated because pregnancies and outcomes of pregnancies were asked in the chronological relationship history in Waves III and IV. For women who may have left out a pregnancy in the ing the age of the oldest biological child in the home. 166 I captured 16% increase in births using this correction. I reconciled the few discrepancies in age at first birth by taking the earliest wave response in the assu mption that recall bias increases with each wave. 163 Table 3. 19 lists the independent variables and how they were measured. All independent variables were measured at Wave I except marriage, which was a time dependent variable.

PAGE 133

117 Table 3 19 : Independent V ariables Independent variables Measured Type of variable Immigrant generation First=Born in Mexico to Mexican born parents Second=Born in the U.S. to at least one Mexican born parent Third=Born in the U.S. to two U.S. born parents and classifies ethnicity as Mexican 167,198 Categorical Family structure Two biological parents Participant lived with two biological or two adoptive parents for his/her entire life Dichotomous Presence of a grandparent in the home Participant lived with a grandparent or great grandparent in the home at Wave I Dichotomous Number of siblings in the home The number of biological, half, or step sibli ngs in the home at Wave I Continuous Has an older sibling Participant has an older full biological sibling who may or may not be in the home at Wave I. Dichotomous Family SES Number of years of education: 0= never attended sc hool 8= 8 th grade or less 10= no high school degree but more than 8 th grade 12=GED or diploma 16=college graduate Continuous As above Continuous ent 1=not working outside of the home 2=unskilled employed 199 3=skilled/professionally employed* Categorical As above Categorical Control variables Age Age in months Time varying continuous Age squared Age in months s quared Time varying continuous Gender ~ 1=male 0=female Dichotomous Family income Logged annual household income, adjusted for inflation and family size at Wave I Continuous Resident father Presence of a father figure (biological, step) in the home Dicho tomous Married Age at marriage, marriage end, second marriage, second marriage end Time varying dichotomous Age at first sex Age at first sexual intercourse Continuous *Because of the small sample size in first generation mothers for skilled (n=50) and professional employment (n=25), they were combined for all generations. ~I controlled for gender instead of estimating stratified models because there were no significant interactions between immigrant generation and gender.

PAGE 134

118 Analysis. I used complementa ry log log discrete time hazard models to examine age at first birth. I used a person month dataset and participants were censored at the month and year they experienced the event or the age of their last interview. 172 I right censored all participants at age 20 be cause I am concerned with the consequences of an adolescent birth and it is unlikely there are public health risks for having a first birth between age 20 and 34. Sensitivity analyses showed no substantial differences in results when right censoring at age 22. When right censoring at age of last interview, many of the relationships became insignificant, likely because there are less consistent predictors for later ages at first birth. I included age and a piecewise polynomial spline approximating the shape of the relationship between age and age at first birth because the relationship between age and age at first birth is positive and linear and then began to flatten around age 18.5. In further sensitivity analyses, I also stratified for males and females an d found no substantial differences. I included all adolescents in the risk set, regardless of whether or not they had ever had sex because pregnancy adolescents are at risk to have sex at any time. Excluding those who had not yet had sex would be excluding those who perhaps delayed sex in order to avoid consequences such as an early birth which would bias results. I first examined the main effect of each family structure and family SES predictor variable and then interacted immigrant generation with each pr edictor. In all models, I controlled for gender because there was not a significant interaction between immigrant generation and gender and stratifying only reduced power in the models. I controlled for marriage because births are more likely to occur with in marriage (although this does not assume the partners were married prior to the pregnancy), because traditional cultural

PAGE 135

119 values include the importance of motherhood within marriage, 109,145,153 and because adolescent childbearing even within marriage is of concern. 209 Because grandparents are more often present in single parent families, 101 I controlled for two biological parents in in the home were included; 24% of participants had no resident father and 5% no resident education and employment. Because higher educated people tend to have more skilled or professional employment, I controlled fo r education in models that included parental employment and for employment in parental education models, as well as family income to reduce confounding. There w ere no missing data for the outcome variable. For independent variables, family income had the highest missing at 37%, parental education and employment less than 5% missing, and all other variables less than 1% missing. I used ten multiple imputed data sets created by chained equations to address missing data. 173 I also adjusted for the complex sampling frame of Add Health. Results Table 3. 20 shows weighted proportions, means, and standard deviations for the study variables. About 16% of the sample had a birth prior to age 20. Those who had an adolescent bi rth, were less likely than those who delayed their first birth to have lived with two biological or adoptive parents and more likely to have lived with a grandparent in the home. Parental socioeconomic status is fairly similar in both groups. Not surprisin gly, adolescents who experienced a birth, had an earlier age at first sex and were

PAGE 136

120 more likely to be married at the time of their birth than those who delayed their first birth. Second generation adolescents were more likely than first and third generation adolescents to live with two biological parents and with a grandparent and first and second generation adolescents had more siblings than third generation adolescents. Socioeconomic status improved with each immigrant generation as parents become more edu cated and more likely to be in a skilled or professional occupation although the proportion of adolescents who experienced a birth is highest in the third generation. Table 3 20 : Weighted Means (Standard Devi ations) and P roportions of Select V ariables Adolescent birth Did not have a birth First generation Second generation Third generation N=1636 15.7% 84.3% 20.4% 35.1% 44.5% First birth <20 years old 13.4% 12.3% 19.4% Male 38.6% 55.3% 49.8% 51.6% 54.9 % Family structure Two biological or adoptive parents 39.0% 58.2% 49.5% 70.8% 45.5% Presence of a grandparent 6.5% 4.9% 1.3% 6.9% 5.6% Number of siblings 1.9 (1.5) 2 (1.5) 2.4 (1.3) 2.3 (1.6) 1.6 (1.4) Older sibling 58.2% 61.0% 70.0% 61.6 % 55.4% Family SES Mother's years of education 11 (2.8) 10.9 (3) 8.8 (3) 10 (2.8) 12.6 (2.1) Father's years of education 10.6 (3.2) 11 (3.2) 8.7 (3.1) 10.2 (2.8) 12.9 (3.2) Mother not working 40.3% 42.2% 52.7% 46.7% 33.3% Mother unskilled 25.5% 27.9% 38.7% 33.5% 17.7% Mother skilled/professional 34.2% 29.9% 8.6% 19.8% 49.0% Father not working 9.4% 12.4% 13.2% 13.4% 10.0% Father unskilled 51.6% 46.4% 62.0% 49.3% 36.8% Father skilled/professional 39.0% 41.2% 24.7% 37.2% 53.2% Ot her control variables Family income $27,500 (2,790) $30,100 (3,060) $17,400 (1,550) $27,600 (3,570) $37,100 (2,830) Resident father in home 64.6% 76.0% 76.1% 84.3% 65.4% Wave I age 15.3 (1.8) 16.2 (19) 16.7 (1.8) 16.1 (2.2) 15.7 (1.8) Age at first sex* 14.9 (1.6) 16.9 (2.8) 16.9 (2.9) 17 (2.5) 16 (3.0) Married before 20 yo 31.1% 6.4% 12.5% 13.3% 7.0% Number of observations 257 1379 334 574 728 Number of person months 20996 123823 25308 62509 57002 *For those who had sex by Wave IV

PAGE 137

121 Family structure: Parents and grandparents Table 3. 2 1 Model 1, shows that the hazard of an adolescent birth for first generation adolescents was 42% (1 0.58) lower than the hazard for third generation adolescents (p=0.083). The hazard of an adolescent bi rth for second generation adolescents was 50% (1 0.50) lower than the hazard for third generation adolescents (p<0.01). A post estimation t test found the difference between the hazard of an adolescent birth between the first and second generation was sign ificant at p<0.01. Males had a lower hazard of an adolescent birth than females. For family structure, Model 2 shows that having two biological parents in the home reduced the hazard of an adolescent birth by 46% (1 0.54) and the significance for first and second immigrant generation was reduced suggesting living with two biological parents is potentially a reason first and second generation adolescents have a lower hazard of an adolescent birth than the third generation. Model 3 shows there were no signifi cant interactions between two biological parents and immigrant generation. Models 4 and 5 show there was not a significant main effect nor interaction for the presence of a grandparent in the home. Table 3 21 : Hazard R atios (95% C onfidence I ntervals) for A dolescent B irth and F amily S tructure Model 1 Model 2 Model 3 Model 4 Model 5 First generation 0.61 0.62 0.62 (0.32 1.07) (0.34 1.11) (0.35 1.07) (0.35 1.10) (0.34 1.13) Second gen eration 0.50*** 0.59* 0.62 0.59* (0.34 0.75) (0.39 0.90) (0.34 1.12) (0.39 0.90) (0.40 1.01) Third generation Ref. Ref. Ref. Ref. Ref. Two biological parents 0.54** 0.54** 0.53** (0.37 0.79) (0.30 1.00) (0.37 0.79 ) (0.36 0.79) First gen X two biological parents 0.99 (0.46 2.15) Second gen X two biological parents 0.92 (0.37 2.30)

PAGE 138

122 Table 3 21 : Hazard R atios (95% C onfidence I ntervals) for A dolescent B irth and F amily S tructure Third gen X two biological parents Ref. Presence of a grandparent 1.16 1.41 (0.59 2.28 ) (0.58 3.43) First gen X grandparent 1.82 (0.19 17.31) Second gen X grandparent 0.41 (0.06 2.56) Third gen X grandparent Ref. Age 1.06*** 1.06*** 1.06*** 1.06*** 1.06*** (1.05 1.07) (1.05 1.07) (1.05 1.07) (1.05 1.07) (1.05 1.07) Age spline 0.89*** 0.89*** 0.89*** 0.89*** 0.89*** (0.85 0.93) (0.85 0.93) (0.85 0.93) (0.85 0.93) (0.85 0.93) Male 0.54* 0.54* 0.54* 0.55* 0.54** (0.34 0.86) (0.34 0.86) (0.34 0.86) (0.34 0.87) (0.34 0.86) Married 8.13*** 7.57*** 7.57*** 7.60*** 7.38*** (4.67 14.13) (4.19 13.67) (4.18 13.73) (4.22 13.71) (4.05 13.45) Constant 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) Person months 144,819 144,819 144,819 144,819 144,819 Family structure: Siblings Table 3. 2 2 Models 1 and 2 show there was no significant main effect nor interaction for the number of siblings in the home. Model s 3 and 4 show there was no significant main effect nor interaction for the presence of an older sibling in the home although the hazard ratio 0.63 approached significance (p=0.193) in the second generation. Table 3 22 : Hazard R atios (95% C onfidence I ntervals) for A dolescent B irth and F amily S tructure Model 1 Model 2 Model 3 Model 4 First generation 0.60 0.66 0.59 0.71 (0.34 1.07) (0.33 1.31) (0.32 1.09) (0.29 1.74) Second generation 0.52** 0.65 0.51** 0.67 (0.35 0.78) (0.34 1.21) (0.34 0.76) (0.40 1.12) Third generation Ref. Ref. Ref. Ref.

PAGE 139

123 Table 3 22 : Hazard R atios (95% C onfidence I ntervals) for A dolescent B irth and F amily S tructure Number of siblings 0.96 1.01 (0.84 1.10) (0.83 1.23) First gen X siblings 0.94 (0.75 1.18) Second gen X siblings 0.90 (0.68 1.19) Third gen X siblings Ref. Older biological sibling 0.91 1.09 (0.69 1.20) (0.71 1.68) First gen X older sibling 0.72 (0.32 1.62) Second gen X older sibling 0.63 (0.31 1.27) Third gen X older sibl ing Ref. Age 1.06*** 1.06*** 1.06*** 1.06*** (1.05 1.07) (1.05 1.07) (1.05 1.07) (1.05 1.07) Age*Age 0.89*** 0.89*** 0.89*** 0.89*** (0.85 0.93) (0.85 0.93) (0.85 0.93) (0.85 0.93) Male 0.54* 0.54** 0.54* 0.54** (0.34 0. 86) (0.34 0.86) (0.34 0.86) (0.34 0.86) Constant 8.04*** 8.20*** 8.11*** 8.09*** (4.60 14.06) (4.59 14.66) (4.70 13.99) (4.68 13.99) Observations 144,819 144,819 144,819 144,819 Family SES: Pare ntal education Table 3. 2 3 shows results from the parental education models. Model 1 and 2 show and 4 show that there was no main effect or significant interaction f and immigrant generation.

PAGE 140

124 Table 3 23 : Hazard R atios (95% C onfidence I ntervals) for Adolescent Birth and Parental E ducation Model 1 Model 2 Model 3 Model 4 First generation 0.47 0.2 4 0.41* 0.34 (0.21 1.08) (0.03 1.71) (0.19 0.88) (0.05 2.34) Second generation 0.51* 0.43 0.40** 0.15 (0.30 0.85) (0.07 2.54) (0.21 0.76) (0.01 2.14) Third generation Ref. Ref. Ref. Ref. Mother's education 0.97 0.94 (0.89 1.04) (0.83 1.07) First gen X mother's education 1.07 (0.85 1.36) Second gen X mother's education 1.01 (0.86 1.19) Third gen X mother's education Ref. Father's education 0.93 0.90 (0.83 1.04) (0.77 1.06) Fir st gen X father's education 1.01 (0.83 1.21) Second gen X father's education 1.09 (0.87 1.38) Third gen X father's education Ref. Not working mother 0.73 0.71 (0.45 1.17) (0.43 1.16) Unskilled mother 0.80 0.76 (0.39 1.64) (0.38 1.54) Skilled/professional mother Ref. Ref. Not working father 0.93 0.94 (0.43 2.02) (0.43 2.02) Unskilled father 1.02 1.03 (0.55 1.90) (0.55 1.91) Skilled/professional father Ref. Ref. Age 1.06*** 1.06*** 1.05*** 1.05*** (1.05 1.07) (1.05 1.07) (1.04 1.07) (1.04 1.07) Age spline 0.89*** 0.89*** 0.91*** 0.91*** (0.86 0.93) (0.86 0.93) (0.86 0.96) (0.86 0.96) Male 0.56* 0.56* 0.54 0.54 (0.34 0.90) (0.34 0.90) (0.2 8 1.02) (0.28 1.03) Family Income 0.79 0.79 0.88 0.88 (0.62 1.02) (0.61 1.03) (0.63 1.24) (0.62 1.24)

PAGE 141

125 Table 3 23 : Hazard R atios (95% C onfidence I ntervals) for Adolescent Birth and Parental E ducation Resident Father 0.82 0.82 (0.52 1.32) (0.52 1.32) Married 7.33*** 7.38*** 9.84*** 9.56*** (3.77 14.25) (3.79 14.35) (4. 50 21.50) (4.33 21.08) Constant 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) Person months 136,324 136,324 108,410 108,410 Family SES: Parental employment Tabl e 3. 2 4 shows the results from the parental employment models. Model 1 adolescent birth. Model 2 shows that the hazard of an adolescent childbirth was reduced by 90% (1 0.1) fo adolescents with professional mothers for the first generation only. Post estimation tests found a significant difference (p=0.007) between first generation adolescents with a mother not working an d third generation adolescents with a mother not working and significant difference (p=0.004) between first generation adolescents with a mother not working and second generation adolescents with a mother not working. The hazard of an adolescent birth was reduced by 81% (1 0.19) for adolescents with unskilled mothers compared to adolescents with professional mothers for the first generation only. Post estimation tests found a significant difference (p=0.085) between first generation adolescents with an unsk illed mother and third generation adolescents with an unskilled mother and a significant difference (p=0.013) between first generation adolescents with an unskilled mother and second generation adolescents with an unskilled mother. Figure 1 provides a visu al interpretation of this result.

PAGE 142

126 Model 3 different than 1. Model 4 shows that the hazard of an adolescent childbirth was reduced by 93% (1 0.07) for adolescents with fathers who aren adolescents with professional fathers for the first generation only (p=0.088). Post estimation tests found no significant difference between first generation adolescents with a father not working and third generation adolescents with a father not working and significant difference (p=0.023) between first generation adolescents with a father not working and second generation adolescents with a father not working. Table 3 24 : Hazard R atios (95% C onfidence I ntervals) for Adolescent Birth and Parental Employment Model 1 Model 2 Model 3 Model 4 First generation 0.47 2.31 0.41* 0.68 (0.21 1.08) (0.74 7.25) (0.19 0.88) (0.18 2.49) Second generation 0.51* 0.60 0.40** 0.41* (0.30 0.85) (0.23 1.59) (0.21 0.76) (0.18 0.93) Third generation Ref. Ref. Ref. Ref. Not working mother 0.73 0.94 (0.45 1.17) (0.49 1.83) Unskilled mother 0.80 0.97 (0.39 1.64) (0.40 2.38) Skilled/professional mother Ref. Ref First gen X not working mother 0.10*** (0.03 0.34) First gen X unskilled mother 0.19** (0.06 0.63) First gen X skilled/professional mother Ref. Second gen X not working mother 0.76 (0.25 2.30) Second g en X unskilled mother 0.66 (0.19 2.35) Second gen X skilled/professional mother Ref.

PAGE 143

127 Table 3 24 : Hazard R atios (95% C onfidence I ntervals) for Adolescent Birth and Parental Employment Not working father 0.93 1.33 (0.43 2.02) (0.43 4.09) Unskilled father 1.02 1.14 (0.55 1.90) (0.43 3.03) Professional father Re f. Ref. First gen X unemployed father 0.07 (0.00 1.49) First gen X unskilled father 0.59 (0.15 2.37) First gen X skilled/professional father 1.00 (1.00 1.00) Second gen X unemployed father 0.85 (0.18 4.05) Second gen X unskilled father 0.97 (0.25 3.79) Second gen X skilled/professional father Ref. Mother's education 0.97 0.95 (0.89 1.04) (0.88 1.03) Father's education 0.93 0.94 (0.83 1.04) (0.84 1.04) Age 1.06*** 1.06*** 1.06*** 1.06*** (1.04 1.07) (1.04 1.07) (1.04 1.08) (1.04 1.08) Age spline 0.91*** 0.91*** 0.90** 0.90** (0.86 0.95) (0.86 0.95) (0.84 0.96) (0.84 0.96) Male 0.56* 0.56* 0.54 0.52* (0.34 0.90) (0.35 0.88) (0.28 1.02) (0.27 0.99) Family Income 0.79 0.80 0.88 0.89 (0.62 1.02) (0.62 1.03) (0.63 1.24) (0.62 1.28) Resident Father 0.82 0.82 (0.52 1.32) (0.51 1.30) Married 7.33*** 7.59*** 9.84*** 9.63*** (3.77 14.25) (3.77 15.27) (4.50 21.50 ) (4.44 20.89) Constant 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) Person months 136,324 136,324 108,410 108,410

PAGE 144

128 Figure 3 10 : Immigrant Generation by Mother's E mployment Figure 3 11 : Immigrant Generation by Father's Employment Age at first sex and the hazard of an adolescent birth Because of the relationship betwee n early sexual initiation and an increased risk of adolescent birth, I controlled for age at first sex in the all models but only discuss models with substantial changes. Table 3. 2 5 shows the two models that changed substantially after controlling for age at first sex. Model 1 show the hazard ratio for two biological parents before controlling for age at first sex. Model 2 shows that after controlling for age at first sex, the hazard ratio for two biological parents became closer to 1 and lost significance. Using 0.20 0.60 1.00 1.40 1.80 2.20 1st gen 2nd gen 3rd gen Not working Unskilled Skilled/Professional 0.20 0.60 1.00 1.40 1st gen 2nd gen 3rd gen Not working Unskilled Skilled/Professional

PAGE 145

129 a Stobel Goodman mediation test, 212 I found that age at first sex mediated about 28.4% of the effect of two biological parents on the hazard of an adolescent birth. Model 3 shows the hazard of an older sibling in the second generation prior to controlling for age at first sex. Model 4 shows that after controlling for age at first sex, the interaction for the presence of an older sibling in the second generation became significant and reduced the hazard of an adolescent birth by 59% for second generation adolescents as compared to third generation adolescents (p<0.05). Thus, age at first sex suppressed the relationship between the interaction of an older sibling and the second generation and the hazard of an adolescent birth. Although not shown in the table, age at first sex partially mediated the and the hazard of an adolescent birth in the first generation, but the relationships remained statistically significant. A Stobel Goodman mediation test showed that age at first sex mediated 16.5% of the effect of a mother not working in the first generat ion and 51% of the effect of an unskilled mother in the first generation. Table 3 25 : Hazard of an A dolescent B irth Before and A fter C ontrolling for A ge at F irst S ex Model 1 Model 2 Model 3 Model 4 Two biol ogical parents 0.54*** (0.37 0.79) 0.75 (0.48 1.17) Age at first sex 0.77*** (0.70 0.86) Second gen X older sibling 0.63 (0.31 1.27) 0.41* (0.17 0.97) Age at first sex 0.75*** (0.69 0.83)

PAGE 146

130 Discussion This is the first study, to my knowledge, to examine differences in risk for adolescent birth by immigrant generation in Mexican origin adolescents using a nationally representative longitudinal sample. Understanding how family structure and family SES influence adolescents differently by immigrant generation is a first step to reducing reproductive health disparities in Mexican origin adolescents. Advancing prior research, I found that first and second generation adolescents have a lower risk for adolescent births than third generation adolescents, with the lowest hazard ratio found in the second generation. Although this finding is similar to research showing a higher risk for adolescent birth in native born Latinas versus foreign born L atinas, 19,28,203 it elucidates a non linear relationship between adolescent childbearing and time in the U.S. which has not yet been documented. Because a higher family SES reduces risk for adolescent childbearin g and protective tradition al values may reduce risk for adolescent childbearing, as described by segmented assimilation theory, perhaps second generation adolescents have the best of both of these worlds. Families of second generation adolescents have a hi gher SES than first generation families but are more likely to maintain protective traditional values compared to the third generation. 55,87 Table 3. 2 6 summarizes the family structure models and support of segmented assimilation theory. Supporting segmented assimi lation theory, having two biological parents in the home reduced the hazard of an adolescent birth, but contrary to my hypothesis, this protective influence was equally protective in all immigrant generations. Two biological parents partially mediated the difference in hazard of an adolescent birth between the second and third generation which means one of the reasons the second

PAGE 147

131 generation has a lower risk for an adolescent birth is because they are more likely to live with two biological parents. The resul ts of the Stobel Goodman mediation test suggest that delaying sexual initiation is one of the ways in which having two biological parents in the home reduces the hazard of an adolescent birth, which is supported in the literature. 95,96 The lack of a significant relationship between the presence of a grandparent and risk for adolescent birth could be due to competing positive and negative effects; perhaps the encouragement of traditional values of motherhood and the lower SES often found in homes with grandparents, cancel out the benefits of added supervision and values that delay sexua l initiation. Similarly, the lack of significance for the number of siblings could also be due to the combination of protective factors, such as protective traditional values found in larger families and risks of less parental supervision when there is a l arge number of children in the home. Interestingly, an older sibling reduced the hazard of an adolescent birth in the second generation as compared to the third generation and only approached significance A ge at first sex suppressed the relationship betw een an older sibling and the second generation and the hazard of an adolescent birth. Prior analyses found that an older sibling reduced the hazard of sexual initiation in all immigrant generations 205 and increased the odds of contraceptive use in the second generation as compared to the third generation. 211 These results su ggest that sibling dynamics work differently in Mexican origin families, particularly in second generation families, because older siblings appear to protect younger siblings from risky reproductive health behaviors.

PAGE 148

132 Table 3 26 : Summary of F amily S tructure and the H azard of an A dolescent B irth Variable OR Supported hypothesis? Supported SAT? Two bio parents 0.54*** (0.37 0.79) Partial: Having two biological parents in the home reduced the hazard of an a dolescent birth across all generations. Two parents partially mediated the hazard between the second and third generation. Supports SAT Grandparents 1.16 (0.59 2.28) No: There was no significant main effect or interaction for the presence of a grandpare nt in the home. Neither supports nor rejects. Number of siblings 0.96 (0.84 1.10) No: No significant main effect or interaction Neither supports nor rejects. Siblings not mentioned in SAT. Older sibling 1.09 (0.71 1.68) Partial: No significant interac tion nor main effects for older siblings but after controlling for sexual initiation, older siblings reduced the hazard of an adolescent birth in the second generation as compared to the third generation. Extends SAT. Siblings in migrant families may have closer relationships and older siblings may discourage early sexual initiation which may decrease risk for an adolescent birth. First gen x older sibling 0.72 (0.32 1.62) Second gen x older sibling 0.63 (0.31 1.27) Third gen x older sibling Ref. *** p<0.001, ** p<0.01, p<0.05, p<0.10 Table 3. 2 7 summarizes the parental education models and support for segmented assimilation theory. Extending prior research, I examined education and employment of both parents instead of creating a variable education alone. 21,22,28,186 education wer e not significantly different than one. 2 Table 3. 2 8 summarizes the parental employment models and support for segmented assimilation theory. However, I found that having a mother who does not work outside the home and having an unskilled mother protects ag ainst an adolescent birth in first generation adolescents as compared to professional mothers. Similarly, my prior research found that in the first generation, 2 In the list an adolescent birth in the first generation as compared to the third generation (p<0.05) although this was lost in the imputed models.

PAGE 149

133 having a mother not working decreased the hazard of sexual initiation as compared to having a sk illed or professional mother 205 and having a mother not work ing increased the odds of contraceptive use among all three generations. 211 Mothers who are not working spend more time on child care 141 and are likely providing supervision. However, unskilled mothers work outside the home and for lower wages than professional mothers suggesting that reinforceme nt of protective traditional values such as the importance of virginity may be one of the reasons first generation adolescents with an unskilled mother have a lower hazard of an adolescent birth than those with professional mothers. Age at first sex partia the hazard of sexual initiation is one of the ways that not working and unskilled models reduce the hazard of an adolescent birth. First generation adolescents with a father not working also reduced the hazard of an adolescent birth. Perhaps the added supervision and encouragement of traditional values in not working fathers outweighs economic Table 3 27 : Summary of Parental Education and the H azard of an A dolescent B irth Variable HR Supported hypothesis? Supported SAT? Mother's education 0.94 (0.83 1.07) No: There was no significant main effect nor interaction Neith er supports nor rejects SAT. First gen X mother's education 1.07 (0.85 1.36) Second gen X mother's education 1.01 (0.86 1.19) Third gen X mother's education Ref. Father's education 0.90 (0.77 1.06) No: There was no significant main effect nor interaction Neither supports nor rejects SAT. First gen X father's education 1.01 (0.83 1.21) Second gen X father's education 1.09 (0.87 1.38) Third gen X father's education Ref. *** p<0.001, ** p<0.01, p<0.05, p<0. 10

PAGE 150

134 Table 3 28 : Summary of Parental Employment and the H azard of an A dolescent B irth Variable HR Supported hypothesis? Supported SAT? Not working Mother 0.94 (0.49 1.83) Partial: In the first generation, a mother not working reduced the hazard of an adolescent birth compared to a professional mother. This interaction was partially mediated by age at first sex. This interaction was not significant in the second generation. Rejects SAT that asserts skill ed and professional parents improve second generation outcomes. Extends SAT by showing skilled and professional parents are not always protective for all outcomes. Unskilled Mother 0.97 (0.40 2.38) Professional Mother Ref. First gen X not working mother 0.10*** (0.03 0.34) First generation X unskilled Mother 0.19** (0.06 0.63) First generation X skilled/profess mother Ref. Second generation X unemployed mother 0.76 (0.25 2.30) Second gen X unskilled mother 0.66 (0.19 2.35) Second gen X skilled/profess mother Ref. Not working father 1.33 (0.43 4.09) Partial: In the first generation, a father not working reduced the hazard of an adolescent birth as compared to professionally employed fathers. Rejects SAT that asserts s killed and professional parents improve second generation outcomes. Extends SAT by showing skilled and professional parents are not always protective for all outcomes. Unskilled father 1.14 (0.43 3.03) Professional father Ref. First gen X not worki ng father 0.07 (0.00 1.49) First generation X unskilled father 0.59 (0.15 2.37) First generation X skilled/profess father Ref. Second generation X unemployed father 0.85 (0.18 4.05) Second gen X unskilled father 0.97 (0.25 3.79) Sec ond gen X skilled/profess father Ref. *** p<0.001, ** p<0.01, p<0.05, p<0.10 Cultural values are often used as a reason Latinas have higher adolescent birth rates than non Latino white and black adolescents 21,22,31 macho may not be protective fo r every adolescent outcome, my findings are consistent with other studies that have suggested that traditional cultural values found in Mexican origin

PAGE 151

135 migrant families may reduce the risk of an adolescent birth. Similarly, Landale and colleagues 26 found that third generation Mexican origin women were more likely to have a birth as a first family formation event than non Latino white women and that first generation Mexican origin women were more likely than all race/ethnic groups studied to have marriage as a first event. These findings suggest that traditional values encouraging married childbearing erodes with each immigrant generation. Future research and intervention planning should find ways to encourage families to maintain traditional values that protect adolescents from pregnancy and birth despite increasing exposure to less protective U.S. values. S ocial and economic disadvantage are certainly the fundamental cause of reproductive health disparities. 117 However, I concur with researchers who found an increased risk of HIV with higher SES in East Africa, 201 that we should not define adolescent birth as a problem of low income groups alone and public health experts and clinicians should not ignore risk in highly educated and employed migrant families. Furthermore, because of the protection that stay at ho me mothers provide in first generation families, we might also look for ways to improve supervision in families where both parents must work such as advocating for a living wage. There are several limitations to my study. Due to sample size I was unable t o test for differences in the hazard of adolescent birth by age at migration for first generation adolescents. However, I found the hazard for an adolescent birth increased as age of migration increased using a six category age at migration variable sugges ted by Rumbaut. 169 ( see Appendix ) Furthermore, the small sample size for particular variables such as the number of first generation adolescents with skilled/professiona l mothers and

PAGE 152

136 not working fathers, decreases stability of models, although all hazard ratios remained appropriate in magnitude. Recall bias influences survey data although using a longitudinal study with multiple waves reduces some recall bias. I did not h ave the ability to measure traditional patriarchal, or conservative sexual values directly and I advocate for future research to improve understanding and measurement of Mexican origin values and their influence on health behaviors. However, the use of b oth segmented assimilation theory and previous research on traditional values in Mexican origin and Latino families, supports the conclusions of my study. Conclusion This study is the first to use a longitudinal nationally represented data set to examine the hazard of an adolescent birth by immigrant generation. My findings advance prior research and segmented assimilation theory by showing that the second generation has the lowest hazard of an adolescent birth, suggesting that immigrant adaptation is not necessarily linear. Scholars and public health experts must take into consideration that although socioeconomic status of migrant families improves with each generation in the U.S., this does not necessarily confer protection against adolescent childbeari ng. Finding ways for families to maintain protective traditional values while they acculturate and assimilate in the U.S. will have a significant impact on the reduction of births in Mexican origin adolescents. C hapter 3 C onclusion In this chapter, I ha ve examined sexual initiation, contraceptive use and risk of adolescent birth in Mexican origin adolescents, and I close by summarizing my findings and support for segmented assimilation theory within family structure, family SES, and

PAGE 153

137 cultural values. Tabl e 3. 29 summarizes the influence of family structure and family SES on reproductive health outcomes. Table 3. 30 summarizes the findings by support for segmented assimilation theory by reproductive health outcome. I am the first to examine whether family str ucture influenced reproductive health outcomes differently across immigrant generations. Table s 3. 2 9 and 3. 30 show that I partially supported segmented assimilation theory by finding that two biological parents provided protection for all reproductive heal th outcomes studied. There were no significant interactions between immigrant generation and two biological parents for any of the outcomes suggesting that two biological parents are no more protective for the first and second generation than they are for the third. Two biological parents partially mediated the relationship between the second and third generation and the hazard of sexual initiation and the hazard of an adolescent birth and suppressed the relationship between the second and third generation and the odds of contraceptive use. Although having two biological parents may not be any more important in the second generation than the first or second, it is one of the reasons the second generation has a later age at sexual initiation, similar odds of contraceptive use, and a lower hazard of an adolescent birth than the third generation. The second generation is more likely than the first or third to have two biological parents. Table s 3.29 and 3.30 show that I supported segmented assimilation theory an d advanced literature by finding that grandparents in the home were associated with a delay in sexual initiation and I extended segmented assimilation theory by finding that contrary to my hypothesis, grandparents were associated with an increase in contra ceptive use. There were no significant interactions which challenge and extend segmented assimilation theory because grandparents are no more protective in the first

PAGE 154

138 and second generation than they are in the third. Table s 3.29 and 3.30 show that the lack of significant main effects and interactions in all three outcomes for the number of siblings in the home did not support segmented assimilation theory. Table s 3.29 and 3.30 show that I advanced the literature and extend segmented assimilation theory by sh owing that an older sibling in the home reduced the hazard of sexual initiation in all three generations, increased the odds for contraceptive use for the second generation as compared to the third, and may reduce the risk for an adolescent birth in the se cond assertion that family and culture are protective in second generation adolescents, my results differ from previous literature that found older siblings increased risk y behaviors in younger siblings, 111,112 probably because family dynamics are different in Mexican origin families than non Latino white and non Latino black families. Table 3 29 : Variable and I nteraction I nfluence on O utcomes Variables Risk of sexual initiation Odds of contraceptive use Risk of adolescent birth Two biological parents Interaction Presence of a grandparent Interaction Number of siblings Interaction Presence of an older sibling First generation Second generation Third generation First generation Second generation Third generation First generation Second generation Third generation First generation (not working) (not working) (not working) (unskilled)

PAGE 155

139 Table 3 29 : Variable and I nteraction I nfluence on O utcomes Second generation (not working) Third generation (not working) First generation (not working) Second generation (unskilled) Third generation (unskil led) Table 3 30 : Variables, T heir I nteraction with I mmigrant G eneration, and S upport for H ypotheses Variables Sexua l initiation Contraceptive use Adolescent birth Two biological parents + + + Interaction Presence of a grandparent + Interaction Number of siblings Interaction Presence of an older sibling + Interacti on + ~ Interaction + ~ Interaction + ~ + Interaction + + ~ Interaction + Key: + supported hypothesis; did not support h ypothesis; ~ HR or OR approached significance for support Second, I advance the literature and segmented assimilation theory by showing that parental SES influenced reproductive health outcomes differently by immigrant generation. Segmented assimilation theory asserts that higher educated and more skilled/professionally employed migrant parents can improve the economic assimilation of their children into a U.S. middle class but it also asserts that traditional cultural values can protect adolescents from a deviant subculture. Table 3. 28 shows that results from parental education models found that parental education in first and second generation adolescents was associated with an increase in sexual initiation which provides support

PAGE 156

140 for protection through traditional values and not through necessarily educational education in the second generation may increase contraceptive use and an increase in n all three education may increase contraceptive use, providing support that education supports contraception use and traditional values may not be protective for all outcomes. Table 3.29 and 3.30 mo st consistent influence on all three outcomes. Having a mother not working protected against early sexual initiation in the first and second generation, increased contraceptive use in all generations, and protected against an adolescent birth in first gene ration. Unskilled second generation fathers were associated with a decrease in sexual initiation whereas unskilled first generation fathers were associated with an increase in sexual k and white approach to traditional values and socioeconomic status. While I do not underestimate socioeconomic status and impoverishment as a fundamental cause of disease, 117 my results challenge the assumption that higher SES is always protective for adolescents and we should take into account the complex relationship be tween family, SES, and traditional patriarchal or conservative values While this is easier said than done, a first step is to conduct qualitative research that explores the meaning of cultural values I also acknowledge that negative selection is also oc curring with migration. There are significantly less educated and professionally employed migrants and thus not a lot of socioeconomic diversity. Differing outcomes by generations could be due to an increase in variation in socioeconomic status of families

PAGE 157

141 and not necessarily segmented assimilation. In conclusion, I found support for particular aspects of segmented assimilation but I extended the theory by showing that socioeconomic status and traditional cultural values are more complex than the theory ass erts. Finally, I hypothesize, along with prior research, that traditional cultural values primarily conservative sexual views, are one of the reasons for reproductive health differences among Mexican origin adolescents of different immigrant generations, and may be a mediating factor that explains how family structure and SES influences Mexican origin adolescents by immigrant generation differently. In this chapter I have only speculated about the influence of cultural values and quantitative analyses alon e will not improve our understanding of how traditional cultural values, specifically gender and sexuality, influence reproductive health outcomes in immigrant families. Qualitative data on the lived experiences on Mexican origin young people can elucidate how traditional cultural values, gender, and sexuality, are constructed, reproduced, and negotiated in migrant families and how they influence of reproductive health outcomes in adolescents. This data will ultimately lead to a better understanding of repr oductive health disparities in Mexican origin adolescents and improve interventions and policy planning. 21,26,83,11 5,140,213,214 It is important to note that many of the quantitative associations were similar in both the first and second generation, even when not significant in the second, the direction of the hazard ratio or odds ratio was consistent with the first g eneration. This means that the largest quantitative difference was between the native families, third generation, and the migrant families, first and second generation. I did not interview third generation immigrants in the qualitative phase and

PAGE 158

142 thus I wil l not be able to provide context for the main quantitative associations. In the next chapter, I describe how gender and sexuality are constructed, reproduced, negotiated and how they influence reproductive health outcomes through qualitative data from life history interviews in Mexican origin young women in Denver.

PAGE 159

143 CHAPTER IV MIGRATION PATTERNS, FAMILY STRUCTURE, AND GENDER I begin this chapter by highlighting the life stories of two women and a pair of sisters whose families are typical of each of the three types of migration classified by Hondagneu Sotelo 3 to exemplify how family migration patterns, family structure, gendered discourses, and the role of parents in the education of their children, ar e directly related to reproductive health behaviors in first and second generation Mexican origin women. 4 N arratives Jocelyn: Independent migration. Growing up with a single mother who migrated alone at age 16 in Chihuahua trying to support her migrated to the U.S. to work, encouraged Edna to migrate too, telling her she could make more money to support their mother and siblings, a common reason for independent migration. Edna snuck across the river and moved in with her sister in the U.S. at age 16. Not uncommon to single migrant women, her relationship with her sister became rocky ew partner se juntaron [They got together, common law married/cohabitated]. 134 3 Hondagneu Sotelo 134 defines three types of migration: independent migration, when an unmarried adult migrates alone or with childre n; family stage migration, when the husband or father migrates before the wife and children; and family unit migration, when the family migrates together at the same time 4 As discussed in Chapter 2, if the interview was conducted in Spanish, I provided th e Spanish quote first followed by the English translation. I also edited the quotes for

PAGE 160

144 18 but separated from her common law husband a few years later due to his drug and alcohol abuse. Within a se juntaron and father made her mother take the children to Mexico to live with his family. While living in Mexico, Edna found out her common law husband had another wife and family in the U.S., and so Edna took her two sons back to the U.S. to live with her sister and left Jocelyn in Mexico with her grandmother. Jocelyn became sick with depression and so her mother went ba ck to Mexico to get her and left her brother in Mexico, not being able to father, and se juntaron when Jo celyn was five years old. This third partnership for Edna was again an unhealthy relationship. Jocelyn described her relationship with her mother and her stepfather growing up, ts. that way. I attend school stuff and do things like that and my parents never did so that. He was violent for like a few years because he was an alcoholic and then he him a lot. But the first few years with my mom he was really violent towards her. In addition to displayin g his power, Jocelyn described machismo relationship. Her stepfather made both financial and parenting decisions, never did any yet her mother always wo rked inside and outside. Jocelyn described how she was primarily responsible for the house and her younger siblings, including getting her sister ready for school, bathing her and feeding her, while her brothers could play outside.

PAGE 161

145 produced gendered roles and discourses in the home. Edna chastised her daughter for dressing like a boy when she was younger and then getting a tattoo, which is not proper for a woman, when Jocelyn was a teenager. Jocelyn described expectations, would never say anything. They never helped with school. If you did your homework you were on your older. Later in the interview Jocelyn recalled wha t it was like growing up with migrant parents, It was probably hard because they were never around. They had to work so much getting assistance so it was not a fun... not lik e we had a great childhood. Vasquez, 215 may occur in poor first genera tion families where parents are primarily focused on providing basic necessities like food and shelter to their children and do not have the time to instill cultural traditions and teachings. This type of parenting was also described by Lareau 116 and is a direct outcome of low SES though it is often used to value marriage or education. 216 of unstable and unhealthy relationships causing her children a lot of turmoil and law marriage was rooted in economic necessity; she could no longer stay with her sister so she found a way out.

PAGE 162

146 Alt hough Edna and her partner did not have time to devote to cultural or academic teaching or emotional support and Jocelyn experienced family turmoil, a lack of parental guidance and initial downward assimilation in adolescence, the family did not fit well i Jocelyn did not experience a lot of conflict with her parents over identity and she did not describe being embarrassed or ashamed of her Mexican family and thus does not exemplify dissonant acculturation. 86,88 asked why she chose Mexican and not Mexican American she said, e filling out paperwork and it says white and I know as Caucasians and I feel like Americans Consistent with Rumbaut, 58 Jocelyn, who did not feel American, experienced discriminati on as a child and was made fun of by non Latino white children. She also grew up very poor experiencing family turmoil and violence, not the stereotypical picture of middle class white Americans and the white picket fence. Due to a lack of parental attenti skipping school, smoking, and drinking. It is not surprising that she experienced an early pregnancy and dropped out of school. Despite the fact that Edna did not provide h er children with a stable nuclear family nor was she a housewife, she reproduced gendered discourses on her children that place is in the home, taking care of the home an d the children, although her mother did

PAGE 163

147 not actually do this. Jocelyn recalled that her mother did not talk to her about her period No vayas a salir con tu domingo though not for religious reasons. Jocelyn was not allowed to date, even though her dated without her parents knowing did not use contraception and told me that she was more afraid of her mom finding out she had sex than she was ab out getting pregnant. She became pregnant at 15 and had a molar pregnancy 5 and started taking oral contraceptive pills prescribed by a clinic. She that time. She ran out of pills and became pregnant again at 16. She chose to parent and her mother and stepfather kicked her out of the house, the decision driven by her beat her up. Inst ead of reproducing the gendered division of power in the home she grew up in, she took her son and moved back in with her mother, never looking back. Although her mother eventually let her move back in, Jocelyn had to drop out of school in order to work an d support her son. GED, has been in an equal and healthy relationship with her common law partner for 11 years, raising three more children with him, and she makes deliberat e parenting decisions 5 A molar pregnancy is a conception that results in a tumor, usually benign, and not a fetus

PAGE 164

148 to provide her children with the childhood she did not have. She talks with her 15 year old son about sex and pushes him to succeed in school to prevent an early pregnancy so to two other young women whose mothers migrated independently and exemplify how migration directly related to a common law marriage out of economic convenience that was then associated with unhealthy relationships and family turmoil, survival parenting, a nd extremely gendered discourses. However, independent migration of women does not always end up in adverse outcomes for adolescents. Although structural barriers such as poverty and a lack of documentation predispose single migrant women to economic depen dence on unhealthy partners and survival parenting, 134,217,218 there are ways that families can still promote upward mobility in their children which will be discussed in Chapters 5 and 6. Isabel: Family stage. Migrating at age three with her mother and siblings t o meet her father in the U.S. 6 mother received a phone call from him saying he had been picked up by immigration and six month old and so she turned herself in and was deported as well. Isabel and her younger brother were born in Mexico while her fathe r continued to go between the U.S. and Mexico as a temporary migrant worker and send money to his family in Mexico. When Isabel was three years old, he sent for them. 6 Isab which was family stage.

PAGE 165

149 and a four year old. She grabbed us all, went on a 42 hour bus ride and crossed the with a four, three, and two year old. She was so brave and such a woman [her Isabel believed her father had the idea to migrate but her mother must have supported the idea, ma the U.S. making two children citizens and two undocumented. Her family returned to Mexico to visit extended family and crossed the border again when Isabel was seven. This time Isabel re membered the scary experience and knew that she, her brother, and her Her mot her stayed home until the youngest was in school and then began to work though both parents worked, they always had dinner together and plenty of mom. They were very verbal and you could just see the love. And so I think it was nice to grow up with that because every day my dad would come home and kiss my mom and be very affectionate in front of us and just the sweet words that while. My mom would dress up and they got ready for each other. We were very fortunate to grow up with having parents that really loved each other and th at you could visually [her emphasis] see it. It was fun. It really set the bar high for us as finding our significant other. We knew what it was to have a healthy good relationship and how important it is to grow up as kids and see that.

PAGE 166

150 Isabel recalled that her mothered challenged gendered expectations in Mexico because she wore pants and learned to drive. Although Isabel recalled that her mother primarily cared for the children and the home while she was working, she said her dad plate and put on the apron and help in any way that he could. But responsibilities and she never remembered her brother washing a dish, although this was recently her mother has been making baby blankets to prepare for desired grandchildren. Despite able to reinforce their cultural traditions and identity. Isabel was not allowed to speak English at home and she was involved in traditional dance. Multiple times in the interview Isa bel pointed out that her parents migrated and risked everything to give them a better life and that is why she and her siblings were encouraged to succeed. When I asked if her parents had expectations regarding her schooling, she replied, I think they did, get really excited when we got good grades. So when we got honor roll they would be so proud of us. I think knowing how proud they were pushed us. It just try to do our best at school and overall. We knew how excited and happy and proud they would be of us if we brought hom e good grades or if we brought home just a good report from a teacher or were at one of those conference meetings and heard good stuff. classrooms even though they c

PAGE 167

151 not date until they were 15 and then t hey had to formally present a boyfriend/girlfriend to their parents. She reported that dating was not her priority in high school and despite attending an urban and relatively poor high crowd as Jocelyn describ ed. Isabel recalled, When I was in elementary school people used to get high and so the influences came from a really good family. I give all this (points to her work office) to my mom and dad who raised me and my siblings, in such a good household. Trust me there of girl s getting pregnant. home, parental involvement and reinforcement of their cultural identity, buffered Isabel pa rents did not provide sexual or reproductive health siblings to do the right thing and take care of themselves, Ay mija, no ms cudese Ud. sabe lo que hace mija when it came to school and everything. They knew if we wanted to do drugs we were going to do it regardless if they knew about it. They knew if we wanted to have sex we could do it during school, after school, it could happen anywhere. think that was just the way we grew up. Isabel believed sex was saved for marriage and having ch ildren was acceptable only in marriage. Although she did have premarital sex, she delayed sexual initiation until 19 and was in a caring long was not able to see her graduate from high scho ol and college. Today she is 30, works at

PAGE 168

152 a non youngest sister was only 10 when their father passed away and she became a young mother. Isabel believed this was due to the loss of their father. Similar to prior research 58,88 that found young adults who experienced selective Mexican American, when I asked her to family, their union was healthy and established prior to migration. Despite working long hours as unskilled laborers, a loving partnership between her parents allowed them to provide a loving environment with moderate gendered discourses, active involvement in promoted upward economic assim ilation in their children. Cristina and Arely: Family unit migration. Growing up with two biological parents who migrated to the U.S. together prior to the birth of their children. hen their mother was pregnant with Cristina. Both women reported their parents migrated for a better life and they wanted their children born in the U.S. Their father bought a house in El Paso and he and their mother crossed the border with visas. His cous in bought cars in Mexico and resold them in the U.S. Their dad went with him to Denver and he liked the city a lot. El Paso was still too close to the border so their parents moved to metro Denver when Cristina was three years old. Their brother was born and few years later and younger than her sister. Neither Cristina nor Arely described migration nor settlement

PAGE 169

153 struggles of their parents who appeared to be working class and not as poor as Jocelyn or Sotelo 134 found, family unit migration requires more resources and occurred in families in Mexico who were middle class or at least not had bought a house in El Paso and they grew up in a small single family home in metro Denver. Their father was one of three fathers in my sample with a high school diploma in Mexico and although he worked in a variety of unskilled occupations he became ski lled and self employed as a handy man and a locksmith. Their mother was a stay at home mother, which required sufficient economic at home. Their SES was also related to their documentation status. Cristina mentioned they migrated legally and both women recalled that the family moved freely between Mexico and the U.S. to visit relatives. Unlike Jocelyn who grew up in a very poor family filled with turmoil, Cristina a nd Arely grew up in intact working class family, free from violence. Although Cristina did not describe a very loving relationship with her mother that was complicated by guilt ther and both women described very loving and supportive relationships with their father and were children but unlike Jocelyn, their dad did help out around the house at time s. Arely believed there were some elements of machismo in her father because he did want their mother to stay home with the family, but Arely believed if her mom wanted to work, he king power

PAGE 170

154 the financial decisions. Although their parents had a more egalitarian relationship than nd Arely remembered Similar to Isabel, Cristina and Arely recalled that their brother had less domestic responsibility although it did not bother them very much either. Cristina helped her si blings with their homework but was not responsible for taking care of them as Jocelyn f learn and she did. She was not allowed to play soccer or boy sports because she was a girl and Cristina felt frustrated because her mother seemed to favor her brot her, even though she was the oldest. This family provides a more moderate and nuanced example of gendered arents promoted education over motherhood although they had a nuclear family and a stay at home mom. Despite the moderate gendered division of labor in the home and differential treatment of reminded their children that they scarified for them by coming to the U.S. and they had no choice but to succeed. to this country so you could achieve something so you have to get good grades. And you have to get good grades if you want to go out, or get good grades if you want me to buy

PAGE 171

155 actively involved in their education by provid ing incentives for good grades and Cristina recalled that even when her father came home from work at night very tired, he would try to help them with their homework. Cristina and their brother graduated from college right from high school and their brothe r eventually received his PhD. Arely, although pushed by her siblings and her parents, did not want to go to college and reluctantly received an was getting married at 21. When I was getting married we went to my dad first. You could tell in his face he was kind of sad because I was the youngest and I was the first one to get married really marriage and childbearing, providing them with an alternative option than marriage and childbearing as their mother had done. supervision, and active parental involvement in education. Like Jocel yn and Isabel, Cristina and Arely did not receive any sexual or reproductive health knowledge from their parents. However, Cristina and Arely learned that they were to save sex for marriage and even their father told them to care for themselves and not get pregnant early; their education came first. Their parents did not want them to date but reluctantly allowed it. Arely described her first date, kind of bought me a ticket to go into the club and it

PAGE 172

156 lasted two weeks. This quote exemplifies open communication about dating, which Jocelyn did not experience. Further, this conversation was with her father, when most women reported that their mothers were the ones who talked to them about dating. Both y oung women delayed sex until 17, the average age of sexual initiation in the Add Health participants. Arely did not use contraception the first time but Cristina did. After her first sexual experience, Cristina taught Arely about sex and contraception. Are ly also learned about and her husband wanted to spend time together and delayed their first pregnancy until Arely was 28. She described a happy and equal relationship wi th her husband and she fixes things around the house, like she ha s always wanted to. overweight and often bullied. She felt her mom was very hard on her and not necessarily very s upportive. Cristina described very low self esteem and a long term emotionally unhealthy relationship that resulted in an unplanned pregnancy from an IUD failure at age 21. To please her partner, she had an abortion and she struggles with regret to this da y. Today, Cristina is married to a man who treats her well and they have seven year old twin boys. Although a stable household, economic resources, parental support and supervision, active parental involvement with school does not guarantee children a pa in free life and economic assimilation, it certainly lays a foundation for success. Economic

PAGE 173

157 reinforce their cultural identity, taking them to visit relatives in Mexico most years and not allowing them to speak Spanish at home. Both women immediately responded, as tern, like economic mobility. 87,88 What segmented assimilation theory has lef t out, is how the process of migration is directly related to family structure, family SES, and acculturation. With Cristina and Arely, family unit migration was possible due to improved SES in Mexico. Staying together as a couple and more economic resourc es provided a stable family in the U.S. with the ability to provide children with supervision, emotional support, and cultural teachings. N arrative Summary Migration and family structure Consistent with prior literature, I found that migration caused si gnificant family disruption. 150,219 221 All first generation women experienced some separation from one or both of their parents during the migration process, either with a migrant father working in the U.S. during their childhoods or during family stage or independent migration. 7 Suarez Orozco found that 85% of first generation children surveyed in Boston were separated from one or both of their parents during migration, and these children were more likely to experience d epression than children who were not separated from their parents. 221 As shown in the literature, the young women described that their family 7 Two first generation women migrated with their partners during adolescence. Mercedes new husband migrated to the U.S. and lived with her fath er who was a temporary migrant worker. Alma migrated with her new husband and her parents never migrated to the U.S.

PAGE 174

158 separation was often d ue to cost and risk of crossing the border together. Some parents chose to migrate in order to find work and settle prior to the arrival their children. 134,221 Unlike the first generation, where 100% of them experienced separation from one or both of their parents during their childhoods, only 30% of the second generation experienced separation from a parent. 8 The three second generation young wome n who had been separated from one or both parents due directly to migration or an unstable relationship during the migration process, all experienced an adolescent birth. In addition nts migrated as a family unit, but her father was deported when she was 11 and her mother stayed with the a Mexican man who was married to another woman and had a family in the U. S. It appears that separation due to migration itself may be detrimental to child outcomes, but as seen children. 221 Four first generation women migrated without their parents during their adolescence. Mercedes married at 19 and her new husband immediately migrated to the U.S. Missing him, she followed. Beatriz, became pregnant at 18, married quickly after finding out she was pregnant, and migrated with her new husband prior to the birth of her twins. Inez migrated at age 14, independently with a family from her hometown and Catalina migrated at age 16 to live with her father who had left her mother and siblings when she was a very young child. Inez and Catalina intended to continue school in the U.S. but both young women experienced difficulty during settlement and experienced an 8

PAGE 175

159 unmarried pregnancy within the first year after mo ving to the U.S. It is likely that migrating during adolescence without a parent or to be with a previously estranged parent presents even more difficulty due to a lack of social support in the U.S. Especially with the recent migration waves of children an d adolescents migrating independently from Mexico and Central America, there is a need for research on the experiences of youth Family structure and ethnic identity Supporting segmented assimilation theory, in addition to Isabel, Cristina and higher, one had an Assoc 58 found that hyphenated Americans were more likely to receive homework help from their parents, women who were able to take pride in their ethnic identity and list the many Mexican traditions they valued as well as be proud of their American identity, citing opportunity, values of education, and rights as reasons they felt American, were able to advance their education and delay childbearing while keeping their ethnic identity. The few women who did not define themselves as Mexican American, but related their connection to the U.S. as well as Mexico, also delayed their first births. Eva, a married 31 year old who is unsure if she and her husband des because her identification with American was obvious due to her fluency of English with no accent and her advanced degree and occupation in higher education. Similarly, Teresa, who delayed her first birth until 24 and is currently half way through a combined BA MSW degree, defined herself as Chicano, which she believed was tied to an American

PAGE 176

160 who felt Amer Except for one women, all of the women who identified as Mexican American or discussed ways they felt Mexican and American experienced selective acculturation largely because they grew up in two biological parent families that provided support, resources and reinforcement of ethnic teachings. These families also experienced family stage or family unit migration. This underscores the importance of considering the migration process and family structure when studying acculturation patterns in Mexican origin adolescents. Migration and gender Cons istent with prior literature, gender played an important role in the migration and settlement process. 126,134,138 As Hirsch described, single mothers may migrate because they are less likely to experience stigma in the U.S. than in their home towns in Mexico. 138 Although none of the participants described this as a reason for their mothers their partners or their partner had left them. All three young women believed thei r mothers migrated to give them a better life, often due to the difficulty of single women finding employment in Mexico. Additionally, three other mothers migrated to the U.S. independently during their adolescence to help support their families in Mexico, and one

PAGE 177

161 mother from migrating, and they remained married. While these seven mothers made the migration decision on their own, like Isabel, Cristina, and Arely, all of the young women decision to move suggesting women have greater decision making power in the migration process when they are on their own or have chosen to leave a partner. 134 divorced and never married, had trouble forming stable long term unions. Carla and but become ill within a few years of migration and actually passed away less than ten years after migrating to the U.S. The other three young women whose mothers who migrated as se juntaron with men after ar riving in the U.S. and then either separated from multiple partners, often very unhealthy relationships, or formed a union with a married man. Single women who migrate, especially with children, may be more dependent on men for economic support and more li kely to enter into unhealthy or short term relationships than single men are when they migrate. 134,218 Consistent with my findings, Hondagneu Sotelo 134 described the economic difficulties and lack of kin support that single migrant women encountered once in the became economically dependent on them. At the time of Hondagneu

PAGE 178

162 two of the six women who migrated independently in their childbearing years were still living with a partner and did not legally marry. Hondagneu Sotelo reported the women had not reproduced gendered roles of subordinated wives and had more egalitarian relationships in the U.S. The women in her study were still with the same partners, unlike the women in mine. Hondagneu Sotelo found that single women who migrated independently did not primari ly migrate to help their families in Mexico, as mine did, but migrated for curiosity, to make more money, and were recruited by single migrant women. All three women who migrated independently in my study migrated specifically to help their families in Mex ico, thus the women in my family may have been more vulnerable to economic dependence on men once in the U.S. causing family turmoil and partner changes. I also did not interview these women directly, as Hondagneu Sotelo did; their daughters told me their stories. Gender also played a role in family reunification. The three young women who described being separated from both parents during the migration process stated that their brothers migrated before they did, regardless of birth order. For example, J sister. There was a pattern of sons migrating to reunite with parents before daughters. This suggests some degree of male preference, although none of these young women remarked about being left behind or last to migrate because she was female. I have not yet seen this discussed in the literature and it appears to be an are a of future research.

PAGE 179

163 Migration and gendered discourses I found that households with instability and those that experienced multiple family transitions, such as parental separations and remarriages, were more likely to exemplify and reinforce extreme gendered discourses. Supporting this finding, prior research found a negative association between egalitarian attitudes and poverty. 222 ided a great example of how family instability and the associated that was caused by poverty in Mexico and unfortunately left her economically dependent upon men in the U.S. Of the ten women who experienced extremely gendered households, five of them had parents who migrated independently, four of them were mothers, and one woman migrated independently herself as a young adult. Only two of the eleven families with more m oderate gendered discourses, experienced independent migration. Gendered discourses were more likely to be found when the mother worked outside the home which was interpreted by all young women as economically necessary and occurred more often when single that mothers and fathers reproduced highly gendered norms in their children that they themselves were not able to maintain. Parreas 220 found that when mothers migrated and became the breadwinners in Filipino families, gender norms in the home were not reshaped to encourage equality for women; fathers and children reinforced norms that included mothers as pri mary care takers and fathers as bread winners. Among the women I spoke with, their fathers and mothers clung to the notion of mother as caretaker and homemaker even when she was working outside the home or when he was actively helping inside the home. Whil e the young women in my sample reported they tried to

PAGE 180

164 negotiate or challenge these gender roles, I found evidence that parents and sometimes brothers reproduced the gendered roles. In contrast, families like Isabel, Cristina and rate gendered discourses, even when mothers stayed at home and took care of the children. I found that the young women who experienced more extreme division of labor and power in the household and gendered discourses regarding female expectations, were m migration as a struggle if first generation. Interestingly, these families were also less likely to have migrated legally and were less likely to form stable families b efore or after migration. Thus, it is difficult to untangle the complex relationship between the migration and settlement processes, SES, family structure, and extremely gendered discourses. Migration, gendered discourses, and reproductive health outcomes Women who experienced an adolescent birth were more likely to have families who experienced independent migration, were more likely to grow up in unstable homes, and were more likely to experienced extreme gendered discourses. Of the nine women who experi enced an adolescent birth, 22% lived with both of her parents through 67% of the 12 women who delayed their first birth, lived with both parents throughout their child hoods. No t surprisingly, 33% of women who experienced an adolescent birth had a mom at home during adolescence compared to 58% of women who delayed birth. poverty, a mom working outside the home, reinforcement of extreme gendered

PAGE 181

165 and working cla ss resources, moderate gendered discourses with a parental active involvement in education, and both young women delayed their first sex and their first birth. In the next chapter, I describe these gendered discourses within families.

PAGE 182

166 CHAPTER V FAMILY STRUCTURE AND GENDER In Chapter 3 I found that aspects of family structure and family SES influenced age at sexual initiation, contraceptive use and the hazard of an adolescent birth differently by immigrant generation. I speculated that cultural constructions of gender and sexuality was one of the reasons for these differences. In Chapter 4 I discussed the association between migration patterns, family structure, ethnic identity, and reproductive health outcomes. Segmented assimilation theory does not address the role of gender in the acculturation or assimilation process. However, a more comprehensive exploration of gender is needed to understand how migration and family influence rep roductive health outcomes of adolescents. Instead of speculating about gender, in this chapter I begin to unravel the complex relationship between migration, family structure, family SES, parenting, and reproductive health outcomes by exploring gendered di scourses in the lives of my qualitative participants. Using the theory of gender and power, I examined gender and sexuality through the gendered division of labor in the household and the gendered division of power between mother and father and between br others and sisters. I categorized cathexis, or affective attachments and cultural norms, into discourses of female expectations and expectations of female sexuality along with sex education and dating rules. Overall, first and second generation young women described similar gendered divisions of labor and power and discourses in the household. However, as discussed in Chapter 4, extreme gendered discourses were associated with family instability and turmoil. Gendered discourses were often ambiguous and some times conflicted what was really happening in

PAGE 183

167 the household. In this chapter, I discuss real verses ideal and how it impacted young G Gendered division of labor Before comparing gender in two cultural or social groups, it is important to understand diversity within each group. It is easy to get caught up in the inequality of women in other cultures by incorrectly assuming that the U.S. has a completely gender equal socie 135 and women are still largely responsible for household activities and childcare. 136 Not dissimilar from the U.S., all young women in my sample described at least some gendered division of labor in the household where mothers were primarily responsible for the home and fathers were responsible to economically provide for the family by working outside the home or caring for outside the home (lawn mowing, automobile maintenance). Consistent with previous research, the gendered division of labor was highly nuanced and there were varyin g degrees of gender differences in each family. 134,137,138 To simplify, I classified the gendered division of labor along a spectrum with three points: men and women moving back and forth between working ou tside the home and performing housework; women mainly being relegated to work inside the home although men may contribute to some housework and families; and an extremely gender would fall between moderate and extreme because although her stepdad did not

PAGE 184

168 participate in any household work, her mot her worked outside the home. Households could be on different places on the spectrum for labor and power, as well as discourses. Gendered division of power As Hirsch 138 argued, women are rarely truly powerless and while there are power imbalances between men and women and women more often than men are subordinated, women are also able to negotiate and gain power in a variety of ways. On one end of the spectrum, women and men have completely equal decision maki ng power and influence over the household. Somewhere in the middle, men and women compromise on say (or believe they have the final say), as described in Isabel, Cristina families. On the extreme end, men exert their dominance through violence and restrict Similar to the gendered division of labor in the household, the gen dered division of and power in the household, I also paid attention to how the women felt and their intonation when describing the labor and power. The gendered division o f labor and power was examined between parents and between siblings. Descriptions of households The egalitarian end of the spectrum Only Teresa described a family where the mother and father moved between gender roles and actually reversed traditional ro les. Her father cared for the home and the children and her mother worked outside the home mothers and no women described mothers mowing the lawn or fixing things. First

PAGE 185

169 generation sisters, Maria and Mercedes were the only women to deny any gendered differential treatment of siblings in household labor. One second generation young woman and Maria and Mercedes, first generation sisters, described decision making in their and Mercedes actually described the most gender equal household in the whole sample and the least amount of childhood conflict. This serves as a reminder that not all Mexican families are patriarchal and no t all U.S. families are egalitarian. Both sisters denied differences in rules and discipline between sisters and brothers and actually believed their mother had slightly more power than their father. Maria stated, tomaba la decisin. Yo pienso no los Si, de hecho hasta la fecha mi papa no es que lo mande mi mama pero ella toma las decisiones. Como ahorita estn pensando en irse a Mxico pero mi papa le dice que depende de mi mama. [With my paren controls my father but she makes all the decisions. Like right now for example they are thinking about going b ack to Mexico but my dad says that it depends on my mother so it is up to her.] As discussed in Chapter 4, SES is intertwined with family structure and gender. Teresa, Mercedes and Maria grew up in two parent households. Moderately gendered households E ven in moderately gendered households, la casa la calle 138,142 144 C ontrary to stereotypes that Mexican men are macho women described a more mode rate gendered division of labor. Mothers were mainly responsible for the home and children, fathers were the primary breadwinners and took care of the outside of the home, mowing lawns and fixing cars. Many mothers worked

PAGE 186

170 outside the home at some point dur helped in the home such as making meals when the mother worked or helped bathe the children. These women generally reflected upon this gendered division of labor with a rather neutral tone or positively ac knowledged their fathers contributions to the home. As and Arely also mentioned their father would help out when he could. Differential treatment of siblings by ge nder was common in the majority of families, as seen in previous studies with Mexican and Mexican American families, with daughters, especially older daughters, having more domestic responsiblity. 108,144,145 In almost all families, women described in inside/outside difference between sisters and bro thers. The young women helped cook, clean the house, and care for younger siblings. Brothers had fewer responsibilities and were often expected to perform outside work such as mowing the lawn. The inside/outside division was not too bothersome for most you ng women, as Isabel described in Chapter 4, she never saw her brother wash a dish but it was normal to her. Because sisters spent more time doing housework, brothers generally enjoyed more freedom and privileges than their sisters which caused varying amou nts of frustration for the young women. The majority of women described a relationship in which decisions were discussed and their mothers had a say but they believed their fathers had final say Even when the women described their fathers as head of th e household, their mothers were not asked how she felt about it, household but I make the dec isions. Like my father is the head of the household

PAGE 187

171 and my mom does everything for him but at the end if she wants a window in that wall she gets a window in that wall [my emphasis] and I feel that is kind of what anything. While many highly educated American women cringe when they hear a woman and women may have different roles or seemin necessarily mean that women are constrained and helpless. Hirsch 138 found that young Mexican women in Mexico and the U.S. described more shared decision making in the marriages t han their mothers did and men described a range of power in the household, although older women were not powerless in their relationships. Extremely gendered households While not all extremely gendered households were the same, eight households fell betw een moderate and extremely gendered. In a few families, there was a more extreme gendered division of labor where the father did not permit the mother to work outside the home or did not help with housework in the home, ever. In addition to Jocelyn, who de scribed in Chapter 4 that her stepfather did not perform any household chores in the home, four other women recalled that their fathers would not help with household chores. When asked to describe her relationship with her Ella hacia t odo el que hacer de la casa y mi papa, en los ratos que estaba all, tena carcter muy fuerte. Quera que todo estuviera muy bien, deca que quien dejaba eso y nos deca que lo recogiramos [She did everything, all the housework and my father, the times he was home, had a temper. He wanted everything perfect, he would ask us who left things on the floor and to pick them up.] When I asked No, mi papa no parece nada atter.]

PAGE 188

172 In addition to Jocelyn, six other women described a patriarchal gendered division of power between their parents, most often reinforced through physical violence. I asked day. And he would go out with other women and bring them around, but my mom always stayed Consistent with prior literature on Mexican households with migrant male worke rs, 134,138,139 sustainability while her father was working in the U.S. But, when her father returned to their family in Mexico, he made her mother close her store and return to working in the home only. their brothers, women described more frustration when they were ultimately responsible for their sibli being or when their brothers received more privileges. As Jocelyn described getting her sister ready for school, bathing and dressing her, Eva also expressed frustration with being the primary caretaker for her younger sisters even though she had an older brother. Dating aside, seven women described they were not allowed out of the house alone without their brothers, or were not allowed to stay home alone, without their brothers. Birth order also played a role here. Older brothers were more likely to have power over their younger sisters. Juana and Carla, both the youngest child and the youngest girl, recalled that their older brothers had the power to discipline them. These extremely gendered households shared commonalities in addition to the gendered division of labor and power. Only two of the eight households contained two biological parents, which meant six of the eight moms worked outside the home. Almost

PAGE 189

173 two thirds of these families experienced independent migration and over half migrated illegally and settled without access to resources. Acknowledging the role of SES allows us to move away from defining patriarchy as a cultural phenomenon alone. C Cultural values including marianismo and machi smo female virginity, and the importance of motherhood, which have been supported in qualitative and quantitative literature, are often used to explain why Latinas have different reproductive health outcomes than non Latina white and black adolescents, ye t few studies have actually explored this. 79,109,143,151 154,223 I described three categories of gendered discourses of ideal female expectation la casa y la calle domain, expectations of marriage and motherhood, and proper female and male behavior, marianismo and machismo Teresa: Real versus ideal ld life and ideal expecting Teresa to serve her brother. Her mother had clear expectations for women to be in home while men could be outside. Teresa described her responsib responsibilities, she replied, He was more free to do other things. I think at one point my mom wanted me to cook f his own hands and feet he could do it himself. She wanted us to be very traditional just like the way it is in Mexico but I When Teresa stood up to her mother, she recalled that her mother was upset.

PAGE 190

174 ior and I working outside the home be do it. ll What is most interesting about the context of this gendered discourse is that after her mother started to work outside the home a nd her father cooked and cleaned and took care of the home. She described their relationship, I think their relationship is what I would envy to have nowadays because nobody get became sick and she had to work, he did everything at home like she would do difference on his manho od, as you would think being Hispanic and our culture the going to do anything for her and my Her mom, who was the primary enforcer of gender roles in the home, reversed roles to care for her husband without complaint. She did it because she saw that he was really sick for a while. He told her help me out for just a little bit and when I get back on my feet you can stop working. It just never got better. He just got sicker and sicker until he passed. My mom took it as, W hen asked, Teresa stated she was not aware of the contradiction between the gendered divor ce from an unhealthy partner, and enrollment in college despite these extremely gendered discourses suggest what was really occurring in her household also influenced her future decision making.

PAGE 191

175 The home is the Although my discussion of the gendered division of labor and power showed that la casa y la calle divide still exists, it played out in a more moderate way, with dads bathing children and making a mess of the kitchen when cooking and mothers going to work every day. D espite what was actually occurring in the household and the seven mothers who migrated independently, most of the young women were expected to live at home until they were married. Of all 21 young women, 9 only Eva and Jessica reported that they lived on th eir own or with roommates prior to marriage and Jessica was the only one who did not experience conflict with her parents when living on campus during her to live a t home during college which she did (although interestingly her father for four months after she moved out of the house in order to better focus on her studies. Althou gh parents supported and encouraged higher education for their daughters, their independence from their parents was still viewed as gained through marriage alone. Marriage, motherhood, migr ation In some households, gendered discourses were more extreme because parents had greater expectations of marriage and motherhood for their daughters than their sons and a few placed a higher value on marriage and motherhood than education and economic autonomy. Juana learned that her only option was to marry and hav e children while her 9

PAGE 192

176 expectations for marriage on her than her brother. With the exception of Juana, a first generation young woman who migrated at age 17 and who learned from verbal messages only expectations, the rest of the young women experienced more ambiguous messages that included being expected to marry and have children, but also to get an education and be independent. For example, Mercedes exemplified how female respectability is intertwined with sexual control and motherhood and education, M: Todas nos casamos bi en nunca nos casamos porque estbamos embarazadas. Primero nos casamos y con el tiempo tuvimos. Todas nos casamos mayores de edad nadie de las cuatro nos casamos chiquitas. Salimos bien de la casa. K: Y sus padres le hablaron de un orden que ellos queran para su vida como por ejemplo que saliera de la escuela y despus se casara y despus hubieran nios? I: Si, por eso todas hicimos lo correcto y salimos bien ellos nos decan primero estudien despus tengan novios despus conzcanlo y si se entienden a se casan. Las cuatro estamos casadas, felizmente casadas. M: We all married well We first got married and with time [had children]. We all got married at an older age, none of us four got married young. We left our home in a good manner. K: Did your parents ever talk to you about an order in which you had to follow in life such as first finishing school, then getting married and afterwards having children? M: Yes, we all did the correct thing and left h ome in a good manner. They told us to first study, then have boyfriends, then get to know them, then if we connected, got married. The four of us are married, happily married. Although most parents actively encouraged marriage and made it clear that unma rried childbearing was not acceptable, nine parents experienced divorce and/or had children out of wedlock. As discussed earlier, migration itself causes disruption of families which may not benefit children in the long run. As prior research has found, co mmon law marriage was more common in U.S. migrants than in Mexico. 149,217,218 I found that the foreign born participants and foreign born pa rents were more likely to se juntaron, common law marry, than native born participants. Se juntaron was often described as

PAGE 193

177 occurring once a woman found out she was pregnant but was often rationalized as an acceptable birth within marriage. Prior research h as found that migrant men are more likely to be unfaithful during long periods of time away from their wives and sometimes have multiple families. 134,149 Jocelyn and Alicia, second generation women who grew up in very unstable households, both had mothers who se juntaron, had children with different men, and both of their biological fathers had oth er families in the U.S. Inez, a first generation woman raised by her grandmother, also said that her estranged migrant Mexico and formed a new family in the U.S. an controlling and jealous man in the U.S., when her father was deported. As described in economic dependence upon men, and then the instabilit y contributes to low SES, where fathers and mothers are often working outside the home for long hours and low wages. These families, who often were survival parenting, were the ones to reinforce ideal gendered discourses of traditional marriages, perhaps i n the hope that their children would be better off than they were. This discussion on the reshaping of marriage through migration that may increase improved power and opportunities for women with Mexico U.S. migration as well as losses in power and opportunities for women with Mexico U.S. migration. 134,138,139,149 her father stopped hitting her mother because it was illegal in the U.S. and her mother left the house more often, stopped making tortillas by hand, and worked outside the home on

PAGE 194

178 occasion. Most young women believed that working outside the home resulted in more decision making power for women and expectations for fathers to h elp inside the home. Maria and Mercedes, the first generation sisters, are currently stay at home moms and Mercedes wed her stayed at home and if they were in Mexico, she would have to work. Although both of these women chose to be stay at home mothers, migration provided them with more opportunities; they both desire to learn English, get their GEDs. As Mercedes said, Porque no quiero estar en la casa todo el tiempo toda la vida no in the house all the time for my whole life.] Moving away from marianismo and machismo Mariansimo. The literature has conceptualized marianismo as abstinence in w omen, caring for family, deferring decisions, like condom use, to men, and being self sacrificing, 109,21,223 which has contributed to a negative depiction of Latina women and stereotypes them as vulnerab le to men and weak. Ehlers 224 explained that Evelyn Stevens created the term marianismo subo rdinate posture in Latin America, and described the idealized belief that women are semidivine, morally superior, spiritually strong beings who manifest these attributes in se of marianismo arguing that subordination is not a choice, as Stevens suggested, but rather a and capitalism. Further, marianismo is part of the marianismo/machismo dicho tomy and assumes that gender roles and expectations are fixed in time and place. I argue that the

PAGE 195

179 first definition of marianismo as well as the adapted version in sexual and reproductive literature should be removed from our description of Latina women. Women learned that proper female behavior included gendered play, being such as not complaining and caring for others. Indeed, self sacrifice was present in some of the ways young women described their mothers, 138,142 particularly when describing how their mothers provided for them. Eva gives everything she has to everybody, she always has. She will always go without and was taught that women are subservient a nd self sacrificing; her mother stayed with her 138 And Teresa an unhealthy partner just because she was pregnant. In addition to descriptions of self sacrificing, loving, and caring, quiet women, the Marianismo implies a quiet strength by enduring pain and su ffering caused by men, however the women I interviewed and most of their mothers did not tolerate pain caused by men. In particular, first generation young women who grew up with single mothers often respected how their mothers left their fathers and migra ted alone to give them a better life. For example,

PAGE 196

180 I asked first generation Catalina, how her mother felt after her father left the family and migrated to the U.S. She replied, Pues realmente nunca nos demuestro ella que lo necesitaba. O sea siempre ella sido muy fuerte y nunca hubo eso de que no necesitaba o que necesitaba a un papa o no [Well, really she never demonstrated that she needed him. So she always been very strong and there was never that I did not need or needed a dad or not.] Beatriz, whose single mother migrated to the U.S. by crossing the river pregnant and with though men are a lot stronger than women in certain ways, como fsico, but we are stronger t single mothers at some point in their lives. This should caution us against assuming traditional patriarchy in all Mexican migrant families or assuming that young women saw their mothers as self sacrificing in a way that made tolerate abuse from men. The definition of self sacrificing has changed shape through migration and over time, wom en sacrificed for their children but no longer felt compelled to sacrifice and stay with abusive husbands, calling into question the validity of marianismo I do not believe that marianismo was central to discourses on female expectations and is not a u they performed their roles were often at odds. Second, while three women reported their mothers stayed in an abusive relationship, three others chose to leave their unhealthy relati onships instead of choosing abnegation. Finally, although not all of the young women interviewed described healthy and egalitarian relationships with their current

PAGE 197

181 partners, none of the young women believed in nor practiced a sacrificial and subservient fe male role at the time of the interview. Similar to Hirsch 138 and Gonzalez Lopez, 217 I did not bring up the term machismo in interviews, because I did not want to bring up a concept so often used to perpetuate negative stereotypes about Mexicans and Latinos and I wanted to get an idea of whether women used this term in their own thoughts and conversations. In general, the women believed machismo hey described what their husbands/partners were not machismo in Mexican men, most of the women described varying degrees of machismo that they saw in their fathers or extended family members, with the most extreme de finition including men who refused to let their wives out of the house or work without their permission. In because women now work outside the home. Unlike many of the wome n who described machismo as a thing of the past, Eva saw machismo at work in the U.S. today and described how she felt about machismo earned that women really are the head of the household. The women really dictate and do a lot more than the men do. It could from one generation to the other, Because my dad worked and took care of the family he was entitled to go off and entitlement to go and hang out with friends or do whatever. In that sense, I find it very arrogant because sometimes you have to work. Especially here [in the U.S.].

PAGE 198

182 it very demeaning with my dad sometimes. We have conversations and he just ominent within my family, with my uncles, with my male cousins, it just gets transferred from one generation to the can shift. Responsibilities ed machismo was not much of a problem in the U.S. today, thus there is not one definition of machismo 137 While men enjoy privilege in Mexico and the U.S. and women may be subordinate or economically dependent upon men at times, gender roles are shaped and reshaped by ti me, place, circumstance and are acted upon differently by different people. Echoing Ehlers 224 and Gonzalez Lopez, 217 scholars and the public should move away from the mariansmo/machismo dichotomy. Ambi guity in discourses Gendered discourses were often ambiguous and while this ambiguity can be confusing for adolescents, ambiguity provides alternative discourses for adolescents to draw upon. Ayala also described ways in which young Latinas drew upon cont radictions 142 Inez described a contradictory message from her grandmother who raised her, K: Y su abuela, que quera por su futuro? I: Ella nada mas siempre me deca que lograra tener una casita donde iba estar. como pidindole a mi mama o andara en la calle sola. K: What did your grandmother want for your future? I: She would only tell me to have a house where I could be a t. She also told me that she wanted me to get married, be independent [give the orders in her home], not ask for help from my mother or be in the street alone.

PAGE 199

183 encourag ed her granddaughter to be independent and not ask for any help. As briefly discussed earlier, the young women who struggled to navigate two cultural identities as they achieved higher education pointed out the ambiguities in their parents push for educa three women, Eva, Jessica, and Ana felt pressure to have children despite their academic according to cultural pressure to have children. Ana had finished college and was married before her first pregnancy and Teresa stayed in an unhealthy relationship due to cultural pressures of both marriage and children. Here she discussed how she felt about her first and unplanned pregnancy at age 22, Later in the interview she described her relationship with her partner, It was retarded. I was with him just because I wanted to have a child because I felt like I was getting old. He knows that, because I told him. We did end up getting I found that while the female expectations and discourses of the home being the sacrificing, were intertwined with messages of female st rength, independence and the importance of education, giving the women another discourse to drawn upon to achieve their goals and providing nuance to complicate gendered dichotomies such as marianismo/machismo

PAGE 200

184 Gender, immigrant generation, and reproducti ve health outcomes Contrary to my hypothesis, I found very little difference between the gendered division of labor, power and discourses in first and second generation families. If cultural and religious encouragement of virginity, marriage and early mot herhood were the primary reason first generation young women had earlier births than the second, then I would have expected to see more married births in the first generation. In fact, I found just the opposite. Only one first generation woman out of nine who had a birth was legally married prior to her first pregnancy, as compared to four second generation women out of eight who have been pregnant. First generation women were also more likely se juntaron and never legally marry or legally marry well after the birth of the child than the second generation. None of the first generation women described shame or any disappointment from family members, even when they se juntaron prior to even getting pregnant. A few women reported their parents were sad because they were so young, and one was afraid her father would be mad when her older sister got pregnant out of wedlock, but respectability or shame on the family was not brought up. In fact, as discussed earlier, first generation sisters Maria and Mercedes repor se juntaron at age 19 prior to her first pregnancy and legally married her husband four years later, Yo pienso que tuve nios muy chica pero no lo hice c tuve a mi hijo fuera. No nos casamos bien por la iglesia pero no era con cualquier hombre era con el que pensaba estar siempre [I believe I had children too young but I

PAGE 201

185 man.] Somehow, common law marriage has become an acceptable form of marriage. As prior research has shown, although plenty of pregnancies occur out of wedlock in and se juntan is uncommon. 217,218 It is interesting that this is found in the first generation and does not extend to the second generation, suggesting this a relative phenomenon migrants. Also refuting a claim of cultural and religious reasons for early pregnancies and births in the first generation, is the fact that second generati on were more likely to still be with the partner of their first child. Five of the first generation women, some who eventually legally married and others who did not, had divorced or separated from the partner of their first child. This could be due to the earlier age at birth and marriage, legal or common law, due to a pregnancy, in the first generation or improved SES in the second generation. Even so, divorce and separation is frowned upon in the Catholic Church and marianismo encourages submissiveness t o husband. As Hirsch 138 found, older women, versus younger women, in Mexico were more likely to see marriage as generati onal changes with age and cohort, both first and second generation young women often left unhealthy or abusive relationships. Although often painful for the women to discuss, they believed it was best for them and their children to leave an abusive partner Sadly, verbal and physical abuse was not uncommon with the women, one first generation woman reported abuse by a partner and four second generation reported abusive partners. Two first generation women left their partners when they were

PAGE 202

186 unfaithful and on e, who se juntaron with her partner at 14, felt like she just wanted to be on her own and grow up and left him. Providing even more support for my argument, six out of the 11 first generation women migrated during adolescence, meaning their formative years were even spent in Mexico. None of these woman believed that marriage was their cross as stereotypes would have predicted. Of note, only two women who experienced an adolescent birth were still with their partners which is consistent with prior literature that showed adolescent mothers are more likely to raise their children on In Chapter 3 I explained that quantitative analyses found that first generation adolescents were less likely to use contraception at first sex than the second generation. My qualitative data support this finding, 45.5% of first generation women used contraception at their first sex compared to 70% of second generation women. Prior literature also supports that cont raceptive use increases as generation or acculturation increases and cultural and religious values that discourage contraception is usually speculated as the reason behind this difference. Although contraceptive use at sexual initiation was one of my main outcomes, contraceptive histories, including reasons for stopping and starting contraception, and contraceptive attitudes were beyond the scope of my dissertation I do have a brief contraceptive history of each qualitat ive participant and a ll but one woma n used hormonal contraception at some point in her life. The one woman who did not ever use hormonal contraception was a second generation woman who use condoms prior to her marriage and could not ever conceive. As I discuss in more detail in Chapter 7, al though many of the qualitative participants stated they were Catholic, they did not believe it influenced the way they made decisions in their lives. My qualitative

PAGE 203

187 results do not support the assumption that cultural religious discouragement of contracepti ve use the reason that first generation women are less likely to use contraception than their second (and third) generation peers. Although I did not specifically analyze barriers to contraceptive use, my data suggests that lack of access due to a lack of health insurance, lower SES, and a language barrier may be an important reason first generation migrant adolescents are less likely to use contraception than their native born peers. All of the second generation women in my sample were bilingual while five first generation women did not speak English. Migrants are less likely to have health insurance than native born persons in the U.S. 1 In the future, I could analyz e the attitudes and patterns of contraceptive use in my qualitative sample. I do not argue that motherhood is not valued in Mexican origin families, however, scholars continue to encouarge. 71 In fact, more second generation women reported feeling pressure to marry and have children, especially after they completed their education or married. However, this could also be because fir st generation women were younger at their first births so pressure may not have been necessary. C hapter 5 Conclusion I agree with Gonzalez predominately of gendered dichotomies, such as marianismo/mac hismo without attention to underlying structural and economic influences leads to a negative racialization of Latinos. In this chapter, I have explored the gendered division of labor, power and discourses. Based on my results, I advocate for a more nuance d view of the currently accepted cultural framework. First, gendered discourses were often at odds to what was

PAGE 204

188 actually occurring in the household. Second, it is a moderate view of gendered division of labor, power and female expectations that decrease ris k for sexual initiation and unmarried adolescent childbearing. Third, contrary to common stereotypes, I found that the most extreme gendered division of labor and power was related to earlier sexual initiation and unmarried and unplanned adolescent childbe aring. Finally, socioeconomic status and family structure were directly related to the gendered division of labor, power and discourses in the household. I did not interview third generation women and I did not have enough detail or variation on parental education and employment in the qualitative sample thus I am unable to support or refute the quantitative findings in Chapter 3. In fact, my qualitative findings that more extreme gendered division of labor and power was found in poor, unstable families a nd increased risky sexual behavior do not contradict my quantitative findings that lower parental education reduced the risk of sexual initiation in the first and second generation because the gendered division of labor and power in the household are not t he same as conservative sexual views that may encourage virginity. I describe how in addition to gendered discourses, parents play a role in the sex education and academic education of their children which directly influence reproductive health outcomes.

PAGE 205

189 CHAPTER VI THE ROLE OF PARENTS IN THE EDUCATION OF YOUNG WOMEN In Chapter 3, I found that second generation adolescents had a lower hazard of an adolescent birth than the first and third generations. Supporting my quantitative findings, the mean age at first birth for second generation women in my qualit ative sample was 23.3, almost five years older than the mean for first generation women, 18.4. Although in Chapter 5 I explained that gendered roles and discourses were related to family structure and SES and reproductive health outcomes as I hypothesized, I found that first and second generation households actually had very similar gendered division of labor, power and gendered discourses and gender does not completely account for differences in reproductive health behaviors between the first and second ge neration, such as age at first birth. In this chapter, I discuss how parents played an active role sex education and academic education of their adolescents in order to further explore why first and second generation have such different reproductive health outcomes. I describe the role of parents in sex education, how sex education influenced the first sexual experiences, the role of parents in academic education, and how they intersected to influence reproductive health outcomes of first and second generat ion Mexican origin young women. S ex Education Mothers played the dominant role in sexual education and communication, as previously reported in the literature. 151,217,225 Mothers are often the parent responsible for the moral and sexual development and actions of their children. Although fathers may blame for any unacceptable behavior of their daughters and sons. 151,217 Interestingl y,

PAGE 206

190 some women mentioned their fathers encouraged them to respect and care for themselves, as Arely described in Chapter 4. Other fathers conveyed another frequent Gonza lez Lopez 217 discussed, the loss of virginity may be a private affair, but pregnancy out of wedlock is a family affair, and fathers are trying to preserve their daughters and their own respectabilit ies That is not to say that fathers of the women I interviewed did and how premarital sex and pregnancy could hurt them. As Gonzalez Lopez 195 found, fathers cared more about protecting their daughters from harm and increasing chances of economic success than virginity in and of itself. Five of the second generation women specifically highlighted that their fathers placed high academic expectations upon them and truly wanted them to succeed. For example, Carla, whose brother was an adolescent father, said that her mother and father colloquial Spanish p No vas a salir con un domingo siete Exploring sexuality and sex education with young women, I found four parental messages that can then be linked to first sexual experiences: a m essage of avoidance or cuidate parents in sexual education: almost all the women felt first sexual experiences were often described as a negative or at best, a neutral experience; and women generally felt unprepared and rarely discussed having sex or preventing pregnancy or STIs prior to the event.

PAGE 207

191 Avoidan ce and discomfort Before comparing Mexican parents to native born U.S. parents, we need to remember the diversity of groups within the U.S. and that parents in the U.S. of any race/ethnic group are uncomfortable talking to their children about sex. 148,226 Almost all menstruation. Hirsch 138 found that Mexican young women gr owing up believed it was disrespectful to talk about sex or even ask their mothers about sex. Consistent with prior literature 138,217 that found Mexican women did not often receive reproductive health knowledge from their families, only four of the 21 women rec eived any sexual knowledge, such as what sexual intercourse actually is, pregnancy and STI prevention, or on about sex was detrimental to them. While I did not ask about menstruation or learning about it, five women specifically pointed out that they received no information about menstruation from their mothers. Some women found they were unprepared for sexual intercourse because they painful experience. They also reported they were not given the knowledge about using contraception, how to get it and use it, or how to bring it u p with partners. Eva discussed The first time I talked to my mom about it [contraception] she was mortified. She was mortified when she found out I tried on my first tampon for going to fail. My friend Emily took me to the bathroom and she got on the stall next to me and she walked me through it. It was the most awkward experience

PAGE 208

192 ever. The whole hour an d a half through swimming class I thought it was just going to float out. I was so nervous about it! These are things that culturally you use it. No one said when you have sex for the first time this is how awkward or how weird it is or these are the things you should ask. None of it. So all of it was awkward but I have a discussion with them anywa y. And the good thing is now I cret and tell them more about it because I know how weird and awkward it was for ta sexual and reproductive health, eight young women recalled that they hid relationsh ips and activities with young men from their parents. This is a form of negotiation and hiding relationships may also reflect a lack of communication and overall discomfort ab out sex and dating found in many of these families. Reasons for avoidance. Some women believed the lack of parental communication about sex was due to embarrassment and discomfort. First generation Alma, an adolescent mother herself, stated, Pienso que e s los padres que no nos hablan les da vergenza al hablarnos yo creo que si nos hablaran o nos llevaran y nos dieran anticonceptivos as como lo hacen aqu creo que seran menos porque es lo que nos falta que nos hablen ms. Porque no basta lo de la escuel a porque dicen muy poquito los papas necesitan decir ms o llevarlo con los doctores y que miren lo que pasa o va pasar. the topic. But if they took their children to clini cs and get them on birth control and spoke to them about it might help, there would be fewer [pregnancies]. They little. The parents need to talk more or bring them to the d octor and see what happens or will happen.]

PAGE 209

193 First generation Guadalupe, who experienced an adolescent birth, wanted her children to women reported a general lack of pare ntal trust, communication and involvement in their Alicia, also an adolescent mom said, ow not to say parents are to blame for, but I think a lot of it has to do with not talking to them about sex or birth control, or just not having that trust. And I think that it could be prevented a little better if teens felt they could talk to their pare nts. K: And when you say talk to them do you mean sex in general or talk to them about anything? al ways tell me [about sex]. And first generation Inez, also an adolescent mom said, A veces es por miedo. O porque no se siente uno con el apoyo del papa o de los papas a los hijos y de los hijos al papa. O sea la mama y papa. Yo creo que es eso porque yo tengo varias nias y varias amistades que han salido de muy temprana que no las deja uno salir para afuera pero es falta de confianza yo siempre digo que la confianza que t le des a tus nia o nio. from children to parents or the mom and dad. I think it is that because I know many girls and friends that have gotten pregnant at an early age the y have left that it is the trust you give your daughter or son.] Three women brought up a la ck of parent adolescents trust when they talked about sneak out to do it. I asked first generation Mercedes, one of the few women planned all three of her children, ho w she felt about the Latina adolescent birth rate. She said, Que muy mal. S, porque como le dio todas las amiguitas de mi hijo tienen 15 aos y ya andan embarazadas y yo pienso que es falta de comunicacin de los padres que no hablan con sus hijas. Tambi n muchas veces les prohben que no

PAGE 210

194 anden con ese nio y entre ms lo prohben ms van. Yo pienso que el base de ese problema es la comunicacin y que no ay comunicacin con sus hijas. Entre ms les niegan una cosas ellas lo van a ser. Por eso yo pienso qu e hay tantas nias faltas de comunicacin que por eso salen embarazadas. Por eso yo hablo mucho con mi hijo de 15 aos porque no quiero que vaya salir as. Tengo miedo aunque sea hombre pierde lo mismo tanto el hombre como la mujer. Yo no quiero que mi hij o vaya hacer eso por eso siempre, siempre hablo con l. Le doy libertad pero no le doy libertinaje. Trato de ser su amiga para que me tenga confianza. pregnant. I think there is a l ack of communication between the parents and their kids. Many times the parents limit their kids and ban them from seeing or going out with a certain boy. The more you tell your kid not to do something the more they want to do it. I think the base of the p roblem is communication and not enough communication with daughters. The more they are denied things the more that to happen to him. I am scared even if he is a boy, he will l ose the same as a him. I give him freedom but I am not too liberal with him. I try to be his friend and gain his trust.] In addition to trusting her son and giving him som e freedom, she also encourages open communication with him. Although Mercedes initially talked about a lack of communication with daughters, as many women did, she clarified here that commun ication is important for sons women generally focused on communication and interventions for daughters, while they are the ones who ultimately become pregnant, it reflect s the blame of adolescent pregnancy on young women, as the (ir)responsible party. Parental avoidance or discomfort with sexual topics precluded them from providing their adolescents with sexual and reproductive health knowledge. Although public discourse and a few women mentioned that communicating about sex with adolescents may be encouraging them to ha ve sex, my qualitative data suggest that a lack of sexual and reproductive health knowledge was related to more negative first sexual experiences and inability to talk with partners about sex.

PAGE 211

195 Lack of preparation for first sexual experiences A lack of sexual knowledge was related to a lack of preparation, for what was going to happen during sex and being prepared for the possible outcomes of sex for their first sexual experience, and was associated with general ly negative descriptions of first sexual experiences. Younger women, particularly those younger than 17, fearful emotions. Women at younger and older ages also described t some sexual knowledge reduced the fear and more negative emotions. The four women who reported that their mothers provided them with any sexual knowled ge, what sex was, how to prevent pregnancy, did not describe feeling unprepared for sex and two described positive experiences. These four women all had their first sex at age 18 or older, which does not support the concern that talking about sex with adol escents encourages them to have sex. Two other women sought out information on sex and contraception prior to having sex and also did not describe feelings of being unprepared or fearful. All six of these women who reported sexual health knowledge used con dom or hormonal contraception the first time they had sex. Feeling unprepared was the only difference I found between descriptions of first sexual experiences between first and second generation young women. First generation were more likely to specificall y report feeling unprepared, perhaps because they were less likely to have received sexual education in school in Mexico and may have been less likely to seek out sexual and reproductive health information on their own, having not grown up in the U.S. wher e it is somewhat more permissible to discuss sex.

PAGE 212

196 Partner discussions. Only six women, all 17 and older, reported that they had previously talked with their partners about having sex and the possibility of a pregnancy. All but one of them used some form o f contraception at their first sex and the one who did not stated that she and her boyfriend agreed to support each other if they got pregnant and would have been happy about it. Only one of these women reported she still felt unprepared, perhaps because h er discussion about sex and pregnancy revolved around the idea that if she became pregnant, she could stay in Mexico with her boyfriend and not return to the U.S. where she had recently moved. Three of the women still reported negative experiences in addit ion to three women who had talked about sex with their boyfriends prior to having sex but it was because their boyfriends had asked them to have sex but they had declined in the past. A few women did describe talking with their partners after the first tim e it happened, discussing the possibility of pregnancy and the emotional component of sex, one making sure it was both of their first times and one who general, women gener ally did not learn the skills or have the self efficacy to discuss sex with partners prior to it happening, which increased risk of non contraceptive use. Virginity Parental messages. While the majority of women expressed a lack of open and useful dialog ue with their mothers about sex, supporting Garcia, 145 this did not mean were expected to remain virgins until they were married, either through conve rsations with their parents or their religious values, which is not surprising because parents of all

PAGE 213

197 race/ethnic groups commonly encourage abstinence until marriage, even when they know it is unlikely. 226 There is a lot of controversy on abstinence only education, and the majorit y of research does not support it, particularly because adolescents who take an abstinence pledge are less likely to use contraception when they do have sex. 191 Although only one woman followed this advice and waited to have sex until she was legally married, I found that expectations of abstinence until married, however convey ed, was associated with a delay sexual initiation and a delay in first birth. Looking at the six young women who had their first sex prior to the age of 17 (the average age at first sex for Mexican origin young women in the Add Health study), only one had reported an expectation of remaining a virgin until married and none expressed desiring this for themselves compared to 11 out of 15 who delayed their first sex until after age 17. Looking at the nine women who had an adolescent birth, two of them reported expectations of virginity while nine of the 12 who delayed their first birth reported messages of virginity. Consistent with prior research, some of the women believed their brothers were not rong for them. 217 I do not necessarily advocate for en couraging adolescents to remain virgins until married because I am unsure if the expectation of virginity helped delay sexual initiation because of the actual message of waiting or because it reflected greater parent child communication. I found that the re were no clear differences in messages of virginity between first and second generation young women, but complicating cultural frameworks, the four women who stated they desired to remain virgins, were all second generation young women. We cannot assume that foreign born Mexicans are always more traditional than

PAGE 214

198 native born Mexicans. Further, two first generation sisters reported they both had sex virginity and marriage are constructed somewhat fluidly. It is not premarital sex that is concerning, but the lack of preparation for sex and the lack of discussion with partners that has been associated with negative outcomes such as unplanned pregnancy, STI and emotional regret or depression. 191,227,228 Parental actions. The women in my sample learned that women need to be protected from dangers, from men, and from themselves and doing anything that cou ld dishonor or ruin their respectability and that of their families. In many families, native and migrant, there are stricter controls for daughters than sons. Prior research in Mexican and Filipino families suggests that one of the reasons women and girls are relegated to the home is to protect their sexual respectability and virginity, 138,142,144,151,229 as well as family respectability. 137,138,217 For example, Mercedes a first generation young woman [We all left the house in a good manner] as Gonzalez Lopez found that some Mexican women dis cussed the importance of caerse bien or marrying as a virgin and not shaming the family. 217(p40) However, the importance of virginity in migrant families may be more about safety than respectability. 151,195 Migrant parents worry about their daughters virginity because there is more risk in the U.S. than in their home countries and they want their daughters to be safe from rape, abusive relationships, STIs and unplanned, unmarried pregnancies. 145,195 Migrant families generally settle in high poverty, high crime areas and view U.S. society as more corrupt and sexua lly immoral placing their daughters at risk. 87,150,195,229

PAGE 215

199 The young women described supervision and rules related to boys and dating including, not being allowed to play with boys when their brothers could play with girls, no riding in cars with boys, formal presentations of boyfriends and the boyfriend needing to ask permission from the father, dating in groups when their brothers could date alone, boyfriends not being allowed in the house, and some were not allowed to date at all. Second generation Carla described her dating experience have three dads. Boys were never allowed to call the house. Growing up my communication with a boy would only be at school. If a boy was to call the house my brothers would h quinceaera 10 my brothers watched carefully as to who I would dance with and if they saw that I danced too much with one of them they would go and take him away. Once in involved with like my extracurricular activities like for MECHA, the Hispanic/Chicano groups he would help out and so he knew the guys that wer e in that circle and he was fine with those guys. It was okay for me to go to dinner with them, to go to lunch with them, for them to call the house, but my brother pretty much had to know who they were before any of that could happen. And because of that More second than first generation women complained that their parents treated their brothers differently than they did, in regards to privileges and dating, and more second generation negotiated gendered dating expectations, usually in th e form of sneaking out. There are two plausible reasons for this. First, reflecting change with migration, second generation parents who are raising their children in the U.S. may have more concerns ts raising their children in 10 A quinceaera is a celebration and ri te of passage for young women on their 15 th birthday, often observed in Latin American countries.

PAGE 216

200 Mexico, as discussed above. Second, second generation women who grew up in the U.S. may have been more frustrated with differential sibling treatment because they were hile first generation women may not have viewed differential treatment as concerning because it was so commonplace in Mexico. Both possible reasons demonstrate a change in both the meaning of virginity and treatment of daughters through the process of migr ation and provide a reminder that families may not necessarily be more egalitarian once in the U.S. I do not expect parents to have no controls or rules regarding dating but the double standard and the messages of vulnerability may not be helpful to young women in the long run. 151 While I advocate that adolescents do need to learn responsibility and self subjectivity can be defined as feeling and acting as a sexual agent instead of a sexual object. 145 that they were responsible to discussed multiple messages rolled into one saying from her mother, I was engrained not to have sex before marriage for sure (laughs) and t hat I needed to be a virgin when I got married. Not to go out with a bunch of guys and just give it up everywhere. Just to respect myself more than anything After a respect yourse lf, for you. No dejes que te manocen, tu aceptan de como t eres, y si te van amar, te van amar sin que t les dando nada. T eres mujer y te tienes que respetar who you are and if they are going to love you, they are going to love you without you giving them anything.] Was the main thing she would always say. grope/touch her, Hirsch 138 found that Mexican young women were expected to demand

PAGE 217

201 respect from young men and not allow them to ruin their respectability through allowing 145 be protected from men but are yet responsible for the behavior of the men. This contradiction of vulnerability and responsibility was also present when pare nts confined their daughters to the home to protect them from dangers of the street but also expected them behave like grown women and be responsible for the home and their siblings. 144 experiences, and lack of sexual subjectivity. The four women who learned that sex is rst sexual experiences (although this was common in many helpful of message as waiting or delaying is. Cudate 145 an other theme of parental messages about sex was to take care of yourself, cudate, and to respect yourself, darte respetar Some women believed cudate was not necessarily conveyed in a direct sexual way but was used to encourage proper behavior and making good decisions. Second generation Arely recalled

PAGE 218

202 Ten cuidado mija. Nada ms es what he meant by cudate, she replied cudate as taking care that they used protection and did not become pregnant. When I asked her what her parents told her about sex, she said, Mi papa no, mi mama era la que nos deca que cuando llegramos a tener relaciones sexuales con nuestros novios que nos cuidramos que usramos proteccin que no lo hiciramos chicas porque no quera que tuviramos un embarazo pronto. as the one that told us that when we were going to have sex with our boyfriends that we had to take care of ourselves and use Cudate appears to reflect better parent child communic ation in the home and a way for parents to show their love and concern for the daughters. Vulnerability, responsibility, and first sexual experiences Garcia 145 found that the adolescent Latina girls she intervi their lack of sexual subjectivity or sexual agency. Similarly, almost all the women I was the one who initiated it and most reported their partners had asked, gestured, p ressured or forced them. While a few women reported they had discussed the possibility of sex prior to it actually happening, only one woman described a planned experience; a romantic evening planned by her boyfriend the night he proposed to her.

PAGE 219

203 This la ck of sexual subjectivity is an embodiment of the vulnerability that the Mexican or Latina women. U.S. culture and media present aggressive men with an insatiable drive for sex while there is less importance on the sexuality of women and more warnings for them to be careful. 230 on the good girl side of the good girl/bad girl dichotomy and allows her to maintain respectability. Interestingly, some women maintained a lack of sexual subjectivity as they got older. When asked if she used contraception or protection the first time she had sex, Yesinia said, my God. always happen like that. I always tell her every choice has its consequences. Here, she described sex as something you have no control over. This type of mindset predisposes women to a lack of agency and control over their bodies and sexual experiences. They may be less likely to refuse sex and less likely to actively prevent a pregnancy or STI if the y believe sex is out of their control. Wanting it. Although the women made it clear that their partners were the one who initiated or asked, with the exception of a few women, they made it clear that they were not victims. I asked Mariela, who snuck out of her house at age 14 on the night she

PAGE 220

204 Wanting it at the time did not mean they wanted it to happen prior or afterward. Three women had wanted to wait until they were married and one until she was out of high school but all four said they wante d it at the time it happened. This shows a separation from goals and actions and how fluid they are. Not all women wanted to have sex the first time. Four women described feelin g pressured by their boyfriends and one was forced. Three of the four women wer e younger than 17 and three lived with a single mother or grandmother for part of their childhoods. Supporting prior literature showing that sexual abuse and sexual assault increase risk for adolescent and unplanned pregnancies, 231 all four women had unplanned pregnancies and three were younger than 18. their first sexual experiences as negative or at best, a neutral experience. Such important moments in life are rarely black and white, and many women described positive and neg ative emotions associated with the experience, however, only four women actually who married in Mexico at 19 and migrated to follow her new husband described her firs t experiences, as Isabel, who had sex at 19 with her boyfriend, described, So we did talk about it, I was okay with it happening, and it was someone that I However, most of the women described more negative emotions and thoughts after their

PAGE 221

205 wanted to have return to the U.S. said, yo A few women reported that they actual stopped during sex due to pain and then were unsure if they actually had sex. Eva, who doing something they knew they weren Lopez also found. 217 For example, Yesinia described her first sexual experience that occurred with her finance a week before she turned 19, Y: I think it was just unplanned and he kind of gestured it and I said yes. We we nt he was jus done instea K: Afterwards, how did you feel about it? Y: Very dirty and naughty. Kind of like everybody knew I had just did it. Even walking I was like, they can tell because of sitting down, I pretended like I was on my period. As expected, the majority of women who were forced, or even pressured, described very negative a nd traumatizing experiences.

PAGE 222

206 Although I did not specifically ask, none of the women offered adjectives related to pleasure or enjoyment. As Garcia 145 discussed, one way to stay on the good girl line of the dichotomy, is to deny that sex is pleasurable. Gonzalez Lopez 217 also found tha t the majority of women reported negative first sexual experiences, even when voluntary. However, she found that some specifically indicated a pleasurable and enjoyable experience, which I did not find. I did not probe much into details of the first sexual experience because it was not directly related to my research question while Gonzalez Lopez studied the sex lives of immigrant women and men. Further, it appear that the women she interviewed were generally older at their first sex because they were all b orn and raised in Mexico. Older women may be more likely to experience and to describe pleasure than very young women, such as 14 year olds. Contraception and first sex Nine of the 21 women did not use contraception at their first sex. Not surprisingly, they were younger (average age 16.3 vers u s 18). They also were more likely than those who used contraception to have reported being unprepared for sex, described sexual heal th knowledge from their families nor did they actively seek it prior to having sex. It is important to note that of the 12 women who used contraception, nine of them used a condom and three were using OCPs. While condoms are the best way to protect against STIs, they are not very effective at preventing pregnancy. Also of concern is that not ask more relied upon the men to protect them and I wonder if he did not present the condom, if the

PAGE 223

207 Garcia 145 found that the adolescents she interviewed were reluct ant to talk to their or demanding the use of a condom. Clearly, more research is needed on how young people discuss sex with partners. The majority of women who repor negative first sexual experiences, and felt unprepared. While researchers and public health planners should address preparing women about sexual experiences due to the association with a lack of contraceptive use, which increases risk for unpl anned pregnancy and STIs, women and men should have the right to a pleasurable and positive first sexual experience. The number of women who described negative first sexual experiences is not acceptable, whether or not we believe young women should be having bjectivity, especially young experiences. Public discourse that continues to promote objectification of women of color, a good girl/bad girl dichotomy, and the aggressive predatory male image must begin to change. A cademic Education My mom did have expectations about school for my brother and I. I do remember that clearly. She did you to make me proud and I need you to do explain to kids that growing up, sometimes it feels like your parents are not

PAGE 224

208 how to do something. It hits on your self esteem and kind of messes with you. I really think it has to fuck with you on some level to be a parent and not be able to Hisp My mom was very into that but c ould never attend my parent teacher conferences, could never attend my track meets, could never come to any of my dance meets, or any of that because she was busy working. But I knew she supported me. Why? Because when I came home, dinner would be there. S he never missed a day of work. We always had a place to live, I always had clothes, how to navigate it and use it. Instead we just assume something and blame someone and then our kids are just lost because no one helps them and their her and I had to reach out and ask our mentors. First generation, college educated, Eva Education was a theme that emerged in vivo. I expected to look at educational outcomes of the young women as they related to timing of reproductive health behaviors but emerged as a powerful influence on reproductive health decision making in almost all of the interviews. In fact, I told my research assistant, a second generation Guatemal an who is a recent college graduate, that I was surprised that even young women who experienced an adolescent birth and did not complete high school said their parents wanted them to do well in school, finish high school, and attend college. She replied, 232 This prompted me to explore the actually occurring.

PAGE 225

209 In the passage above, Eva ex plained the paradox of education in Mexican communities and families. She begins by pointing out that her mother wanted her to do well in school but her mother did not know how to support her education. She believes that children may interpret their parent as not caring about their education which perpetuates a stereotype of Latino families not supporting education. Eva goes on to describe survival parenting and economic barriers that prevented her mother fro m attending PTA or being active in her education. In this section, I argue that Mexican migrant parents want their daughters to succeed academically and do not promote motherhood before education, yet due to many barriers are often unable to put their desi academic success. Eva mentioned one of two ways I found that patterns could promote academic success in their daughters despite immense barriers; messages of sacrifice and active involvement in educa tion. Academic achievement of Mexican and Latino adolescents Prior literature and public discourse has suggested that Latino families, in particular Mexican origin families, do not value education. 71 The lower aca demic aspirations and outcomes of Mexican origin adolescents as compared to other race/ethnic and immigrant groups, 58,88 coupled with the high adolescent birth rate, paints a negative picture of Mexican origin families. As d iscussed in Chapter 5, the gendered division of labor, power, and discourses in the Mexican origin women I interviewed were not nearly as patriarchal as stereotypes would have us believe; there is a lot of nuance. Similarly, how Mexican origin families con vey academic expectations to their children, is also highly nuanced and does not support current stereotypes.

PAGE 226

210 Despite the notions that Mexican families do not value education and the barriers that Mexican parents face to helping their children with acade mics, many first and second generation Mexican adolescents experience academic success and attain higher education. In fact, a greater proportion of second generation Latinos are college educated than the Latino population as whole. 233 In addition to segmented assimilation, research line generation achieves higher education than the first and the third generation. Despite some controversy on what happe ns to academic achievement between the second and third generations, there is general consensus that the second generation achieves higher academic achievement than the first. 125,198,234 Regardless of the theoretical framework, all Mexican families are not ruining educational prospects for their children. Consistent with prior research, 109,142,145,195,235,236 every single woman in my sample reported her parents wanted her to do well school. Prior research found that Latino youn g women and their mothers believed that motherhood should come after education and establishment of a career 109 and multiple studies have found that Latino families view adolescent pregnancy as a threat to completing higher education. 109,142,145,195 That being said, the young women described differences between what their parents wanted for them and what their parents actually did to help them achieve it. As discussed in Chapter 5, there was a divide between the real and the ideal and parental actions ofte n undermined their intentions. Although Mexican migrant parents do value education and want their daughters to achieve higher education, barriers often prevented them from promoting academic success in their

PAGE 227

211 children. As Swidler 237 than values, allows us to understand how migrant parents want their children to succeed academically but learning new strategies of action may be too costly in their current socioeconomic situation. There is room for improvement in the way Mexican migrant families promote education in their daughters but it is important t o frame their limitations within a structural context to avoid perpetuating harmful stereotypes. acting on their desires to promote educational achievement and how par ents and the Mexican community at times undermined education of young women. Next, I explain the two ways in which some parents overcame these barriers to promote educational achievement; through messages of parental sacrifice and active parental involveme nt in Parent barriers to promoting academic success Research has shown a large race/ethnic achievement gap in the U.S. and Latino and Mexican origin adolescents are more likely to drop out of school than their non Latino white black, and Asian peers. 1,235,238,239 Latino children are less likely to be read to at home and their parents are less likely to talk to them about their education than other race/ethnic groups. However, these differences are largely related to low SES. 240 242 Further, some Latino families believe that education begins at school and not at home as often occurred in their countries of origin where attending secondary sc hool is often a costly privilege. The young women discussed numerous ways they felt their parents had difficulty directly supporting their education. A language barrier prevented parents from helping

PAGE 228

212 with homework, attending teacher conferences or talkin g with teachers, and created a power dynamic when children were responsible for translating for their parents. Six of the seven women who said they translated for their parents described translating as a negative experience, feeling like they were an assis tant rather than a child or linking it with a lack of guidance parents provided. Foner and Dreby 229 listed multiple studies that found translating was a sourc e of conflict in Mexican immigrant families. Jessica explained possible consequences of this power conflict, It would always be great to find a way to get the parents involved and to empower the parents because I think if the parents are empowered then t he child will be as well. I think especially in Mexican American communities you have parents that ow are be doing more for the parents and a focus on the parents would make a difference. First and second generation parents of women in the qualitative sample averaged between six and nine years of education and only one parent had a college degree. Some of the women explained that their parents could not afford to attend secondary school in Mexico and there is less opportunity for higher education in Mexico and Central and Sou th America, especially rural areas, than there is in the U.S. For rural women especially, marriage and motherhood may have been their only option to obtain status as an adult. 138 Consistent with prior literatur it difficult for them to help their children with homework. 235 L ow levels of educational attainment and a lack of understanding of how the school system works in the U.S. presented difficulty for the young women, especially in high school when it came to taking advanced classes or applying for college and financial aid 235 Alicia and Alma, both adolescent mothers, specifically stated they wished their parents had pushed them

PAGE 229

213 mo re and Mercedes, who delayed her first birth but did not graduate from high school Siempre nos apoyaron para estudiar no mas que cada quien tomo sus propias decisiones pero siempre nos apoyaron para seguir estudiando suppo rted us to study but everyone made their own decisions but they always supported us to keep studying.] Many first generation women, especially those who migrated during adolescence believed their parents supported school but did not push them nor really qu estion ed lack of experience being pushed to excel academically themselves, is one reason they were sometimes unable to actively promote academic success in their daughters. Some socioeconomic situations could be described as working class and quite a few of the women verbalized that th hard hours to support them and put food on the table. Many low income parents in the U.S. may face long hours and low wages, but it may be worse for migrants. Migrants, especially those that are undocumented are at risk for exploitation, very low wages, long hours, or unsafe working environments. 195 Levitt found that in Dominican migrant families, parents provided less supervision and ha d less control over their children in the U.S. due to their long work hours. 66,243 The lack of time and energy migrant parents have presents a barrier to attending parent teacher conferences or helping with homework. Jocelyn whose story opened Chapter 4, believed that her parents did not guage barrier and because her mother only went

PAGE 230

214 as far as 3 rd grade. Jocelyn and Alicia felt their mothers cared more about chores and helping in the house than they did about homework and grades probably because their working mothers depended upon their da and survival parented. When migrant parents struggle to put food on the table, helping with homework may not be the priori ty, especially with language barriers and low educational attainment. As others have argued, the devaluing of education may be more directly related to impoverishment and discrimination of poor minority communities than cultural values of education. 87,215 Before judging parents and an entire culture that they do not value education, we should seek to understand the structural barriers to p revent blame and stereotypes. Navigating two cultures Five women, four college educated and one currently enrolled in college, believ ed the Hispanic/Mexican culture does not value education the way that it is valued in the native U.S. (white) culture. They also believed a lack of educational opportunities and cultural support that encourages motherhood more than education is a reason La tina adolescents are more likely to give birth than non Latina adolescents. Carla believed gives their kids m ore free time. Like I said, I would have wanted to go out more but yet I graduated with high honors. I knew what my goal in life was. I knew push their kids to do things beca culture, and I believe that is one of the reasons why [Latinas have a higher birth rate] too. Similarly, Ana stated,

PAGE 231

215 I think, and is no emphasis on education so there is a void there and so you have to fill it maybe you would want to have a baby instead. That sounds rude, huh? the challenges of navigating two different cultures. Obtaining higher education and delaying childbearing made them feel more separate d from their Mexican roots and family and friends, although they were deeply rooted and proud of their ethnicity. Vasquez 215 discussed how college may be seen as class roots in some poor minority communities. Some women in my sample felt disassociated from their families and communities because they were higher educated but they were really rude to me described that her education made her feel separated from her family and community, were in my situation. Growing up, for instance, families would come over with kids my age and I would never associate with them. I always felt like there was a And then you reach that point and feel like you are no longer part of that in limbo not really knowing where you belong or where you should be or what you can eventually put something together for yourself. I think a lot of people who do go beyond wh connection to your culture and you start exposing to yourself things that your and put it next to the wedding pictures to do like a comparison to how important it is. I think a big issue is that a lot of people get scared because then you start to detach yourself from what you really are.

PAGE 232

216 Jessi ca verbalized how some of the higher educated women I spoke with felt; as they become closer to completing their education, even in households where parents actively pushed education and held high educational expectations for their children, when the young women completed their education, motherhood suddenly trumped education. Eva stated, In my family I find it interesting because education is really important for my mom but kids will always trump that. I feel like she brags more about her kids than she do es about what I do, that I went to school, and that kind of stuff. But Similarly, Ana believed her parents encouraged and expected her to graduate from college but after t he birth of her two children, could not understand why she was considering a graduate degree. While her own family supported her delay in childbearing understanding how a ma rriage between an educated woman and an uneducated man was going to work. This does not mean that Mexican parents and communities do not support education or value motherhood above education. First, the women who believed the Mexican/Latino culture does not value or know how to value education had college degrees or were currently enrolled in college at the time of interview. Despite the separation they felt from their families and communities, it did not prevent them from academic success and they all re ported that their parents expected them to succeed in school and discouraged adolescent childbearing. Second, when Jessica was discussing masters she [mother] was lightin g candles with all these saints around the house and my

PAGE 233

217 dad would say your mom is going to burn the house down! ucation in her own way. Perhaps some parents, like parents of other race/ethnic groups, may not be great in communicating their pride in their children. Third, U.S. white women who delay marriage and childbearing and obtain college degrees and professional employment have a very different life course than women who marry in their teens in rural Mexico and become stay at home mothers. It is possible that Mexican migrant parents have no idea what higher education looks like and any life they have ever seen, particularly when they come to the realization that their college educated daughters are not yet married or mothers. Messages of sacrifice Despite all of the barriers mi grant families face, two of the 11 first generation women and four of the 10 second generation had college degrees by the age of 25 and none of these women experienced an adolescent birth. When looking at patterns between those who experienced an adolescen t birth and dropped out of school and those who delayed birth and finished high school or c ollege, two patterns explained this difference. U.S. and young women whose paren ts had an active involvement in the education had higher academic achievement and were more likely to delay their first sex and first birth In addition to the migration narratives provided by the first generation young women, all of the second generatio n women reported that their parents told them how and why they migrated. Three second generation women stated their parents migrated during

PAGE 234

218 their adolescence to support their parents and siblings in Mexico. The other seven second generation young women and the three first generation who migrated before age 10, believed that their parents migrated to the U.S. for a better life, often for their children. Beatriz, whose mother migrated while pregnant and with two children under three, passed away when Beatriz she came here. Como para mantenernos no 145 found that all 40 second generation adolescents she interviewed learned about their their parents sacrifice and migration struggles. Another study found that girl advance their education was guided by values of familismo and that they would be able to better care for their families with education. 236 linking their migration sacrifices to expectations. Eight young women recalled that their parents told them to work hard so they would not have to struggle or told them that they wanted more for them. Mariela, who moved around to different states during her with working long hours they may not have been able to provide a lot of educational assistance. Mariela did well in school as a child but after her father was deported and her

PAGE 235

219 mother remarried a controlling and jealous stepfather, she married at 14 to get out of the of her father. Knowing that their parents wanted more from them may have helped early educational success but it did guarantee academic success. A more pow erful message was when parents described their migration as a sacrifice and directly linked to educational success. All five 11 of the young women who received this message of sacrifice delayed their first sex to after age 17 and delayed their first birth to adulthood, and four completed college. Eva: Jessica : "We sacrificed for you to be here so you need to excel." Cristina: "We went to this country so you guys could achieve something so you have to get good grades." Julieta: "He would always tell us those stories [about his migration] because he wanted us to be better and he wanted us to do things better and just thank God for what we had here because he didn't have that out there and stuff like that." Isabel: mom had and the courage that she had and how she fought for her family. Both hink they came here necessarily 11 specifically but did believe her parents migrated to give her and h er siblings a better life. It is likely that she heard the same message as Cristina and although Arely did not want

PAGE 236

220 need to succeed in one simple sentence, Julieta and Isabel related their parents sacrif ices to why their parents pushed them to succeed. When parents specifically linked their of education and making their desires for their children expectations. I am the first to document the relationship between parental messages of sacrifice and reproductive health outcomes in Mexican migrant families in the U.S. However, others have found that migrant parents who leave their children in their home country view thei r decision as a sacrifice and hope it will encourage academic achievement and economic mobility in their children. 220,244 Single migrant mothers from the Philippines de scribed their struggles and sacrifice to their children to motivate them to do well in school 220 and single migrant mothers were often able to buffer the emot ional distress of their absence by reminding their children in Mexico of their sacrifice. 244 Additionally, ss will justify the 235(pp13 0 & 176) Knowing that failure to succeed would family turmoil and transitions. The final way that parents encouraged educational s uccess was through their actions. I found that in addition to messages of sacrifice, parents who acted on their success, with or without a

PAGE 237

221 education, helping with homework, attending parent teacher conferences, checking on grades and school attendance, bringing them to school and picking them up, punishing them for not getting good grades or using positive reinforcement for good grades, not allowing them to work outside the home while they were in school or financially helping with tuition or providing cars or a rent free home duri ng college. While the young women who had an adolescent birth said their parents told them to study and would sometimes pack their lunches or pick them up from school, none of them described consistent investment directly related to education such as check ing on grades or attendance. In fact, these women described their parents specifically did not help them. When I asked Alma, Uno solo, llegaba y pona hacer las tareas y siempre est aba en mi casa siempre yo no tena ayuda de mi mama y mi papa [I did. I got home and did my homework and was always in my She later said, Mi mama si nos deca que estudiarnos pero tampoco nos puso muchas re glas o as para que fuera ms estricta no fue tanto as no nos deca si quera que estudiemos pero llegramos a una cierta parte yo no los voy a esforzar. Y mi papa tambin nada ms quiero que estudien pero no nos decan nada. Si falto que fueran ms estri ctos en ese sentido. regarding school. She told us that she was not going to force us to complete a certain level of education and to complete however much we wanted to. My father While all of the parents had a language barrier and most had a time barrier due to working long ho urs for low wages, the ones that managed to show some active excelled academically and delayed their first birth. With the exception of Eva, the other

PAGE 238

222 four women whose paren ts placed high expectations on their children because of their sacrifice also had active parental involvement. In Chapter 4, I discussed how Cristina and the evening exhaus education, and a lack of understanding of the U.S. school system, she reported that she and her sister did not have to do any housework as long as they were doing well in school. So what is interesting about my mom is that she never let us do laundry, we never did any ch ores growing up. She made sure we knew how to do everything but she would never make us do anything because she would tell us our main focus is school and we were to focus on that and to do well in school. She would say, My mom did laundry by washing it outside on the washboard, everything by at us then she now that I look back. But we knew how to do everything. Despite similar economi c and language barriers that Mexican migrant families faced, Gibson 235 found that Sikh children in California ou tperformed their white and Mexican origin peers. Sikh parents made it clear that schoolwork came before housework or any other activities and strictly monitored attendance and time spent on homework, even though they rarely stepped foot inside the school. My qualitative data have shown that Mexican migrant parents do in fact value education and want their children to succeed because that is one of the reasons they migrated to the U.S. in the first place; to give their children a better life. But parents fac e many barriers when it comes to be involved in their

PAGE 239

223 able to provide ano ther discourse for their daughters to draw upon that valued education before motherhood. Aside from difficult to change structural implications, such as immigration reform and increasing the minimum wage, we can help teach parents that they can dramaticall to them about their sacrifice, why they need education and back up their messages with actions such as monitoring attendance and grades. Supporting my conclusions, when I asked the women what they thought about the Latina adolescent birth rate, eight women specifically talked about the need for academic children. Ana and Carla were quoted at the begin ning of the section describing that the Hispanic culture needs to shift emphasis from motherhood to education then motherhood. Mercedes said more parent child communication was needed. I asked her to clarify what she meant by communication, En general pero tambin hablarles de que no deben tener relaciones tan chiquitas porque para todo hay tiempo y el tiempo de la adolescencia no es para tener relaciones. Ese tiempo es para estudiar y estudiar [Communication in general but also to talk to them about not having sex at an sex. It is the time to study and study.] Many of these women talked about how they talk to their own children about the importance of education and are acti vely involved their education. Immigrant generation and education Second generation women were more likely to have parents who placed high academic expectations on them, through messages of sacrifice and active investment in

PAGE 240

224 their education. None of the first generation women who migrated after age 10 described any messages of sacrifice, perhaps because they migrated during their adolescence and their parents did not have much left time to parent. Educational attainment was higher in second generation wom en than first (mean years of education 14.6 and 11.8, respectively); all of the second generation had their GED or high school diploma by the time of interview while over half of the first generation women had not received a GED or high school diploma. Twi ce as many second generation women had graduated from college and the two first generation women who graduated from college, migrated prior to age 10. Second generation women finished high school and college prior to their first pregnancy, whether or not i t was unplanned. Many second generation women specifically stated that they wanted to finish their education prior to getting married and having a child, thus migration and being born in the U.S. with both exposure to U.S. society and having parents who we health decisions, placing educational goals above the goals of motherhood. All three of the second generation women who experienced an adolescent birth eventually received their GED or diploma and only one of the six first generation women who experienced an adolescent birth received their GED or diploma which could represent greater opportunity due to documentation status in the second generation or that the first generation young moms chose t o mother their children prior going back to school as first generation Mercedes described in Chapter 5. Gender and education Because education is not part of my main research question, I only briefly discuss the role of gender in the education of Mexica n origin adolescents. Migrant parents often

PAGE 241

225 have stricter controls over their daughters, fearing U.S. culture will corrupt or harm them and they will lose their families respectabilities. 144,145,150,229,235 While some parents strongl y encouraged and expected college for their daughters, as discussed in Chapter 5, living away for college was not as acceptable for daughters as it was for sons. Gibson mak ing her less valuable for marriage and more likely to buck traditional roles of mothering and housework, and they would not allow their daughters to live away from home during college. Vasquez 215 Mexican migrant families placed more importance on their sons completing education than daughters. However, the majority of women in my sample believed their parents had the same academic expectations for their brothers and themselves; only first generation education more because they were male. Two women recalled that their brothers were lazier which caused their parents to push their brothers more. Future work can provide ion. Education as a potential mediator Throughout Chapter 4, 5, and 6, I have argued that risk for adolescent birth is related to migration patterns, family disruption, extreme gendered discourses, sex education promoting vulnerability and a lack of prep aration. However, messages of sacrifice and the importance of academic success, allowed young women to drawn upon a discourse other than motherhood as the only option. Eva is a case and point. She experienced significant family turmoil; she was left in Mex ico for a few years with her

PAGE 242

226 grandparents before being reunited with her single mother who remarried and divorced multiple times. She also received extreme gendered discourses that included the vulnerability of women. When Eva wanted to move out of the hou se to better focus on tu eres una mujer machismo education for her brother more than her. As the oldest daughte r, Eva was also responsible for caring for her mother and her younger siblings. Despite these hardships and excel, making her mother proud, showing her mother what women can do, and providing an example to her three young degree, delayed marriage until 25 and is unsure if she w experience highlights the nuance in gendered discourses in migrants and provides an example of how parents can encourage upward economic mobility in their children, despite economic hardship and family instability. In addition expectations, Eva sought out mentors whom she believed were crucial to her success. I mmigrant Generation, Adolescent Births, and Gendered Households I examined the qualitative data for associations between migration, family structure and and academic education and reproductive health outcomes, in three different ways. I compared differences between immigrant generation (RQ3), differences between those who had an ado lescent birth and those who delayed, and differences between extremely

PAGE 243

2 27 gendered and moderately gendered households. 12 I conclude this chapter by highlighting my conclusions for each of the three differences. Differences by immigrant generation Table 6.1 sh ows differences in main concepts by immigrant generation. Similar numbers of first and second generation women came from extremely gendered households. A greater proportion of second generation women lived with two biological parents than the first (70% vs 36%) and more first generation families experienced independent migration (57% vs 40%) and described their migration as a struggle (86% vs 37%), thus the first generation experienced more turmoil and was likely of lower SES than the second generation. Tab le 6 1 : Migration, F amily S tructure, S ex and A cademic E ducation, and R eproductive O utcomes by I mmigrant G eneration First generation Second generation Extremely gendered households 36% 40% Independent migra tion* 57% 40% Migration struggle* 86% 37% Two bio parents 36% 70% Parental education action 9% 60% Parental sacrifice 18% 30% Virginity messages 54% 60% Mean years of education 11.8 14.6 Sex is bad message 18% 30% Age at first sex 17 17.6 Used con traception at first sex 46% 70% Experienced an adolescent birth 54% 30% Identified as Mexican American 9% 50% *I did not include the four women who migrated independently or with a partner although two were from extremely gendered households and two fro m moderately gendered households. 12 It was very difficult to find patterns and dra w a dividing line between moderate and extremely gendered households. Originally, I defined 10 households as extremely gendered because they were extreme in one of three categories: labor, power, cathexis. I analyzed the data a second time and decided that only labor and power should be information to categorize cathexis and because two families only had extreme cathexis and not extreme labor and power meaning they were differ ent than the other eight. Making this change did not substantially change my conclusions.

PAGE 244

228 In Chapter 4 I argued that migration was related to family structure and often caused family disruption. Migration also causes a disruption in schooling for children of all ages. Those who migrated during childhood but after starting school, reported difficulties with language and being made fun of. Migrating during adolescence makes it much more difficult for adolescents to stay in school and graduate which also increases risk for an adolescent pregnancy. 31,19 8 Four of the women who migrated during adolescence had planned to continue school, only two even attended school and both dropped out. None of the women mentioned that documentation was a reason they had an early pregnancy nor did they mention a lack of options during their adolescents although Isabel believed a lack of options related to documentation status was a reason for Latina adolescent childbearing. Interestingly, while fairly equal numbers of first and second generation women received messages of virginity, more second generation women than first learned that sex age at first sex was similar for first and second generation women and consistent with my quantitative findings and prior research, a greater proportion of second generation used contraception at their first sex than the first generation. As discussed earlier in the chapter, second generation women were more likely to have parents who placed high academic expectations on them, through messages of sacrifice and activ e investment in their education. Educational attainment was higher in second generation women than first (mean years of education 14.6 and 11.8, respectively); all of the second generation had their GED or high school diploma by the time of interview while over half of the first generation women had not received a GED or high school diploma.

PAGE 245

229 As stated at the beginning of the chapter, the average age at first birth for first generation women in my qualitative sample was 18.4, all whom had their first birth prior to 21. For second generation women, the average age at first birth was 23.3. Although second generation women were older with their first pregnancies and births, they were not any better than first generation women about planning their first births. As women women in the sample who have ever been pregnant, only three of them, roughly 18%, planned their first pregnancies. However, of those who have had a pregnan cy and birth, more second generation young women had previously discussed the possibility and timing of pregnancy with their partners. This could be because second generation were generally older and were more likely to be married prior to their first preg nancy or it could represent a change with immigrant generation. Hirsch 138 found that both migration and generational age contributed to a delay in first birth after marriage. Both younger women, and those who h ad migrated, were more likely to purposely delay or desire to delay their first births in order to enjoy being married. Migration increases exposure to U.S. women who are more likely to delay their first birth after marriage than they are in Mexico and are more likely to seek education and careers prior to having a child. Consistent with segmented assimilation theory and prior research, a greater proportion of second generation self identified as Mexican American, than the first generation because the sec ond generation is U.S. born. Differences by adolescent birth Table 6.2 shows differences by adolescent birth. Advancing segmented assimilation theory and prior literature, I found that a greater proportion of women who

PAGE 246

230 experienced an adolescent birth des migration and a struggle. This demonstrates the importance in understanding the migration processes and may also be a proxy for SES. Consistent with my quantitative results and previous literature, a greater n umber of women who experienced an adolescent birth experienced family turmoil and did not live with two biological parents their whole life probably because these families experienced independent migration and were of lower SES. The lower SES and family tu rmoil made it difficult for parents to be parental sacrifice nor active involvement in education and not surprisingly, they had three years less education than women who del ayed their first birth. Table 6 2 : Migration, F amily S tructure, S ex and A cademic E ducation, and R eproductive O utcomes by Adolescent Birth Adolescent birth Delayed first birth First generation 67% 45% Inde pendent migration* 50% 36% Migration struggle* 60% 38% Two bio parents 30% 58% Parental education action 0 50% Parental sacrifice 0 27% Mean years of education 11.3 14.5 Virginity messages 22% 83% Sex is bad message 44% 8% Age at first sex 15.9 18. 4 Used contraception at first sex 56% 58% Extremely gendered households 67% 25% Identified as Mexican American 0 54% *I did not include the four women who migrated independently or with a partner although two were from extremely gendered households and two from moderately gendered households. A greater proportion of women who experienced an adolescent birth received messages that sex was bad rather than the importance of virginity. They had sex almost two years earlier than women who delayed their fi rst birth, but they were equally likely to use contraception at their first sex. With such a small sample, it is possible that the young women who used OCPs at her first sex but had missed pills and got pregnant or the few

PAGE 247

231 women who said their first sex wa s the only time they used contraception, accounts for this similarity. Key to my findings, is that a greater percentage of women who had an adolescent birth described an extremely gendered household. Supporting segmented assimilation theory, none of the wo men who experienced an adolescent birth defined their race/ethnicity as Mexican American, perhaps because they did not selectively acculturate having grown up with parents were likely survival parenting. Differences by gendered households Table 6.1 shows differences in family and outcomes by gendered household. In Chapter 5 I described the gendered division of labor, power and discourses and I defined extremely gendered households if they were described as extreme in the gendered division of labor or powe r. 13 In addition to more gendered households experiencing independent migration and experiencing family turmoil, women from extremely gendered households were less likely to receive messages of virginity than women in more moderately gendered households and were also more likely to receive messages that sex is bad. Of the eight women in the most extremely gendered households, they initiated sex about a year earlier than women who grew up in moderately gendered households (16.6 and 17.8, respectively). Half o f them used contraception at first sex while 63% of those growing up in moderately gendered households used contraception at sexual initiation. Almost two thirds (63%) of the eight women experienced an adolescent birth compared 13 home until her dad was deported at 10 and her mom remarried a very controlling and sometimes violent man. She left her home with her husband at 14 to get out of the house. Because the gendered division of power was a factor in her decision to marry so young, I classified her as having an extremely gendered household. I actually recla ssified her after drawing my conclusions so her classification did not change my conclusions.

PAGE 248

232 to 31% who grew up in modera tely gendered households. It is important to note that two of the three women who grew up in an extremely gendered home but delayed their first birth received parental messages of sacrifice and the third, Yesinia, experienced a violent and unstable home un til she migrated to the U.S. at age 10 and moved in with a more loving, supportive and gender equal stepfather. Messages of sacrifice may buffer the influence of family instability and extremely gendered households. Table 6 3 : Migration, F amily S tructure, S ex and A cademic E ducation, and R eproductive O utcomes by Gendered Household Extremely gendered household Moderately gendered household First generation 50% 54% Independent migration* 71% 20% Migration st ruggle* 71% 40% Two bio parents 30% 69% Parental education action 0 46% Parental sacrifice 25% 23% Mean years of education 12.8 13.4 Virginity messages 38% 69% Sex is bad message 38% 15% Age at first sex 16.4 17.8 Used contraception at first sex 50 % 62% Experienced an adolescent birth 63% 31% Identified as Mexican American 0 50% *I did not include the four women who migrated independently or with a partner although two were from extremely gendered households and two from moderately gendered house holds. Conclusion As others have argued, reproductive health decisions and behaviors do not occur in a vacuum. There are social, cultural, structural, economic, and political factors influencing decisions. In these three qualitative chapters, I have atte mpted to complicate the current cultural framework by showing that the Mexican culture is not as straightforward as scholars and public discourse has lead us to believe. There is a wide range in how households construct and reshape gender and sexuality. I challenged segmented assimilation theory because it has left the process of migration out of the theory, presents family structure, family SES, and traditional values as mutually

PAGE 249

233 exclusive and assumes traditional values are always protective. My qualitat ive data show that households with more extreme gendered division of labor, power, and discourses were more likely to be single family households or have experienced multiple family transitions. The young women from these households reported their parents migrated as a struggle perhaps because these families were less likely to have entered the U.S. legally. Thus, socioeconomic status in intertwined with family structur e and extreme gendered discourses. In sum, women who experienced an adolescent birth had parents who experienced independent migration, were of lower SES, experienced more family transitions and turmoil, experienced extremely gendered households, did not r eceive messages of parental sacrifice or active parental involvement in education and more likely to learn that sex is bad rather than being encouraged to delay sex, and fe lt unprepared and engage in sex early. Again, this argument does not contradict nor support the quantitative results because I did not interview third generation women nor did I have the detail or variation in parental education and employment. However, messages of virginity from parents were associated with a delay in sexual initiation. This supports my speculation that first and second generation lower educated migrant parents reduced the hazard of sexual initiation in their adolescents because they are more likely to hold conservative sexual views not necessarily have more extreme gen der and power, than h igher educated migrant parents. In the next chapter, I discuss I mix quantitative data and qualitative data to better understand the arguments in the last three chapters

PAGE 250

234 CHAPTER VII MIXED METHODS I ntegrating Quantitative and Qualitative Results In Chapter 3, I discussed the results of my secondary quantitative analysis, using the National Longitudinal Study of Adolescent Health, a large nationally representative longitudinal sample. This analysis allowed me to: (1) examine event history; (2) examine how family structure and SES moderated the association between immigrant generation and reproductive health outcomes; and (3) generalize results to a national population. The quantitative analysis offe rs limited opportunity to explain how the associations may work within the context of migration, family dynamics, gender and sexuality, acculturation, and assimilation. Much of the quantitative literature and literature reviews on reproductive health in La tino adolescents recommends future qualitative research to better understand how attitudes, norms, and context such as migration history, help explain quantitative relationships. 21,26,28,31,214 In chapters 4, 5 and 6, I discussed the results of my qualitative interview s. The qualitative interviews allowed me to explore highly nuanced gender identities and roles in first and second generation young women and their families. The qualitative interviews also allowed me to understand reproductive health behaviors within a co ntext of migration, family dynamics, gender and sexuality, and acculturation and assimilation. things 14 245(p2) and t hese 14 activities, or situations 245

PAGE 251

235 meanings are derived from human interaction. In fact, West and Zimmerman 129(p129) easily captured in simple survey questions but the qualitative semi structured interviews, where a participant shared her story, included meaning and interaction in addition to a life history of events. I mixed qualitative as well as quantitative methods to combine strengths of both, including the generalizability of the quantitative phase and the in depth context of the qualitative phase, and ensured there are no over lapping weaknesses. 158 Most importantly, using a mixed methods approach, I filled theoretical and methodological the unmaski 246(p235) 145 by providing evidence to support immigration reform a nd refute reproductive health disparities in Mexican origin adolescents. In addition to the iterative process between the quantitative and qualitative phases I descr ibed in Chapter 2, I combined results from both phases to answer my research questions. In this chapter I discuss results of Phase III; I mixed qualitative and quantitative results in two ways. First, I used qualitative data to inform selection of mediator initiation and the hazard of an adolescent birth differently by immigrant generation and immigrant generation. Second, because grandparents and siblings have often been left out in studies examining reproductive health behavior and there are not quantitative variables

PAGE 252

236 on relationships with grandparents and siblings in Add Health, I used qualitative data to explain how grandparents and older siblings may protect young women from early sexual initiation and non contraceptive use. 15 Similarity of the quantitative and qualitative samples In order to have the ability to draw inferences between the nationall y representative quantitative sample and the local qualitative sample, I interviewed women who are the same ages as the Add Health participants are now, so that the women theoretically could have participated in the Add Health study. Because of the singula r location and bias in purposive snowball sampling, they may more like each other and less like women elsewhere. However, my intentions were not to generalize my qualitative findings to Mexican origin women everywhere or to Add Health, but to provide a fir st step in contextualizing reproductive health behaviors and the quantitative relationships between immigrant generation, family structure and family SES. It is important to understand the similarities and differences in the quantitative and qualitative sa mples, before integrating findings. Table 7.1 presents the means and proportions of the Add Health participants and the qualitative participants by immigrant generation. Life history interviews provided family structure and family SES at multiple time poi nts while I measured these variables at Wave I in the Add Health participants. In order to provide a more meaningful comparison, I tried to measure the variables around the time of the qualitative women reported their 15 Add Health does have information on sibling relationships only for par ticipants whose siblings participated in Add Health which is a significantly smaller sample size.

PAGE 253

237 mothers stayed at home when they were young children but entered the unskilled employment as unskilled because they were unskilled workers during the partici adolescence or when she would have been in Wave I of Add Health. Table 7 1 : Descriptive Means and Proportions of Qualitative and Quantitative Participants First generation Add Health First generatio n qualitative Second generation Add Health Second generation qualitative Third generation Add Health Age at interview (mean) 35.6 32.9 34.6 30.7 34.5 Age at migration <6 24.6% 18.2% 6 13 42.6% 27.4% 13 17 32.8% 18.2% >17 0% 36.4% Age at migration (mean) 8.3 12.2 Two biological or adoptive parents 47.4% 36.4% 69.2% 70.0% 44.2% Presence of a grandparent* 1.3% 0.0% 6.6% 0.0% 5.7% Number of siblings 2.4 3.6 2.3 2.2 1.6 Older biological si bling 69% 100% 61% 80% 55.1% Family SES~ Mother's years of education 8.9 6.4 10 7.6 12.6 Father's years of education 8.8 8.2 10.2 9.2 12.8 Mother not working 53.0% 36.4% 47.7% 60.0% 34.4% Mother unskilled 37.5% 54.5% 32.6% 20.0% 17.5% Mother skilled/professional 9.4% 9.0% 19.7% 20.0% 48.1% Father not working 13.5% 0.0% 13.4% 10.0% 10.7% Father unskilled 61.3% 63.6% 49.1% 60.0% 36.3% Father skilled/professional 25.2% 18.2% 37.5% 30.0% 53% Reproductive hea lth outcomes Age at first sex 16.7 17 16.9 17.6 15.8 First birth <20 years old 15.30% 54% 15.2% 30% 23.0% Contraceptive use at first sex 48.6% 45.5% 54.8% 70.0% 64.8% Highest level of education Less than high school 39.1% 54.5% 22.2% 0.0% 25.4%

PAGE 254

238 Table 7 1 : Descriptive Means and Proportions of Qualitative and Quantitative Participants High school degree 60.6% 45.5% 77.4% 100% 74.4% College or higher 7.9% 18.2% 12.6% 40% 11.8% Observations 338 11 559 10 745 Having a grandparent in the home was measured at Wave I. No qualitative participants lived with grandpar ents during their adolescence, but 4 first gen and 3 second gen lived with grandparents at some point in their lives ~Two participants did not know their fathers' education or employment I calculated the mean age for Add Health participants at Wave IV i n 2008 and added six to get the mean age in 2014, when I interviewed qualitative participants. I had widened the age inclusion for qualitative participants by two years above and below the age of the Add Health participants to facilitate recruitment. The A dd Health participants were about 3 4 years older than the qualitative participants, although in both samples, first generation participants were about a year older than second generation participants. The mean age at migration was lower in Add Health part icipants than in the qualitative sample and there were no participants who migrated after age 17 in Add Health although there were three qualitative participants who migrated between 18 and 19. For notable family structure differences, slightly more first generation women in the Add Health sample qualitative sample grew up with intact two biological parent families although in both samples, less first generation women grew up with intact two biological parents than the second generation. A greater proportio n of the qualitative participants had an older sibling in the home than the quantitative participants, but in both samples, >60% had an older sibling in the home. For family SES, parents of the qualitative participants have slightly less education than the Add Health participants, however, in both samples parents averaged less than 12 years of education. For first and second generation Add Health participants, the majority of participants had mothers who were not working at Wave I. This was true for second generation qualitative participants, but for first generation

PAGE 255

239 participants, the majority were employed in unskilled occupations. In both samples, the percent of mothers who were skilled or professionally employed increased from the first to the second gene qualitative sample with less than 15% of fathers not working and the majority being unskilled workers. The qualitative participants and quantitative participants had similar ages at fir st sex although first generation qualitative participants had an earlier age at first sex than the second, probably due to one second generation qualitative participant with a first sex at age 23, an outlier. Contraceptive use at first sex was also similar among the quantitative and qualitative samples with more second generation women using contraception at first sex than the first generation. A higher proportion in my qualitative sample experienced an adolescent birth compared to those in the Add Health s ample (42.9% vs. 23.4%, respectively) which could have been a result of non random and snowball sampling. Because of the small sample size, having a higher number of women who experienced an adolescent birth actually allowed for a better qualitative compar ison between women who was around 15% as it was in the qualitative, I would have only had one or two women in each generation with an adolescent birth making it dif ficult to find patterns. The second generation qualitative participants achieved higher levels of education at the time of interview than the Add Health participants, however, in the five years since Wave IV, some of the Add Health could have achieved more years of education and completion of degrees which could make the education gap between samples smaller. Second generation participants were higher educated than the first generation in both samples

PAGE 256

240 although the qualitative first generation participants h ad lower levels of education than the Add Health participants, which could be due to the later ages of migration of qualitative participants. Mixed methods limitations The largest limitation to the study is that I drew inferences and mixed data between a nationally representative sample and a local purposive sample. With the exception of age at time of qualitative interview and age at migration, there did not appear to be substantial differences in demographics or outcomes that would invalidate use of qual itative data to provide context for the quantitative data. I do not believe that the 3 4 year age difference between sample decreases validity because I focused primarily on childhood and adolescent family and cultural determinants of age at first sex and risk for an adolescent birth and all women in both samples were old enough to have experienced both outcomes. Looking at it from a cohort perspective, Add Health participants were on average born in 1980 and the qualitative participants born in 1984 and th ere were probably not substantial changes in the U.S. between those years that would have caused substantial differences in life experience in childhood and adolescence. Age at migration is a more meaningful difference because my qualitative participants w ere somewhat older when they migrated and spent more of their adolescence in Mexico than the Add Health participants. Because qualitative interviews were to provide context, not to quantitatively measure timing of outcomes, I do not believe the age differe nce substantially impacted conclusions drawn from the qualitative data.

PAGE 257

241 M Background My quantitative analyses discussed in Chapter 3, found that having a mother not working reduced the hazard of sexual initiation in the first and second generation; increased the odds of contraceptive use in all generations; and reduced the hazard of an SES predictor across all three re productive health outcomes. In Chapter 3, I only speculated why having a mother not working may be protective, particularly in the first and second generation. Although I only had a few skilled or professionally employed mothers in my qualitative sample, I had enough working and not working mothers to draw conclusions. As more thoroughly discussed in Chapter 4, a greater number of women in my qualitative sample who delayed sexual initiation until after 17 and delayed their first birth until after 20 had sta y at home mothers. Women with stay at home mothers described more supervision and more positive relationships with their mothers, including support and communication. The seven women with mothers who stayed at home throughout their childhood and adolescenc e all lived in families with two biological parents and a working father. These results extend segmented assimilation theory because the theory treats family structure and family SES as mutually exclusive; supportive two parent families and skilled or prof essional higher educated parents are each associated with upward assimilation in second generation adolescents and there is no mention of how not working or stay at home mothers may be more protective than professionally employed mothers.

PAGE 258

242 To my knowledge, no other studies have used quantitative or qualitative data to reproductive health outcomes in Mexican origin adolescents. A recent report found that on average, stay at home mot hers spend seven hours more per week on child care than working mothers and married stay at home mothers spend nine hours more than married working mothers, 141 thus children with a mother at home spend more time with their mothers than those with working mothers. Research has found a neg ative relationship between maternal monitoring, rules, strictness, control, maternal support and risky sexual behaviors. 25,190,77,82,247,248 The few studies that examined nativit y or acculturation did not find differences in parental processes and sexual risk behavior by nativity or acculturation. 25,190 Fathers are often left out of adolescent reproductive health research and research and public discourse on Mexican origin men has described them as macho possessive men who control their wi ves and daughters and provide economic, not emotional support. 19 5 Although mothers and fathers may parent differently and mothers are often 145,217 fathers are still important to the growth and development of their children and warrant further attention in the literature. In fact, a recent study foun d that a close relationship with fathers, but not mothers, reduced the odds of sexual initiation in adolescents. 249 I chose potential mediators based on qualitative findings that could help explain ent and reproductive health behavior. I then integrated the quantitative and qualitative findings by using qualitative data to explain the quantitative associations. Although I had similar findings

PAGE 259

243 in contraceptive use models, I did not analyze them here d ue to model stability and the number of missing on the outcome variable. Figure 1 shows the conceptual model for my hypotheses. I hypothesized that maternal adolescent communication, maternal adolescent relationship quality, maternal supervision, and mate rnal control could mediate the relationship between having a mother not working and a reduced hazard of sexual initiation and an adolescent birth in the first generation. I also hypothesized that having a resident father in the home will reduce the hazard of sexual initiation and adolescent birth across immigrant generations. Figure 7 1 : mployment Methods Quantitative sample. I used data from all four waves (1994 2008) of the National Longitudinal Study of Adolescent Health (Add Health). The survey began in 1994 with 20,475 7 12 graders and participants were interviewed again in 1996 (Wave II), 2001 2002 (Wave III), and 2007 2008 (Wave IV). 160 I restricted the sample to Mexican origin respondents. I excluded five participants missing on ever having sexual intercourse, six participants missing on the presence of a parent in the home and seven participa nts with an age at first sex prior to 11 due to the unlikelihood that sex was consensual at that age. 165 After excluding respondents who were missing without a sample weight, the final sample size was 1,637 Mexican origin participants. For the Moderator: immigrant generation employment Reproductive health behavior First sex & first birth IV DV M ediators: maternal relationship, communication, supervision, control

PAGE 260

244 adolescent birth models, I included the five participants missing on ever s ex, I excluded six participants missing on the presence of a parent in the home, seven participants with an age at first sex prior to 11 due to the unlikelihood that sex was consensual at that age, and six participants who were missing on the outcome of th eir pregnancy. 165 No first births occurred prior to Wave I amon g Mexican origin participants. After excluding respondents who were missing without a sample weight, the final sample size was 1,636 Mexican origin participants. Variables. Table 7.2 presents the independent variables. All independent variables were measur ed at Wave I except for age and marriage that were time varying. Table 7 2 : Independent Variables Independent variables Measured Type of variable Immigrant generation First=Born in Mexico to Mexican born pa rents Second=Born in the U.S. to at least one Mexican born parent Third=Born in the U.S. to two U.S. born parents and classifies ethnicity as Mexican 167,198 Categorical 1=not working outside of the home 2=unskilled employed 199 3=skill ed/professionally employed* Categorical Control variables Number of years of education: 0= never attended school 8= 8 th grade or less 10= no high school degree but more than 8 th grade 12=GED or diploma e college 16=college graduate Continuous Age Age in years Time varying continuous Age squared Age squared Time varying continuous Gender ~ 1=male 0=female Dichotomous Family income Logged annual household income, adjusted f or inflation and family size at Wave I Continuous Resident father Presence of a father figure (biological, step) in the home Dichotomous Married Age at marriage, marriage end, second marriage, second marriage end Time varying dichotomous Potential medi ators

PAGE 261

245 Table 7 2 : Independent Variables Relationship quality with mother 1=strongly disagree 2=disagree 3=neither agree or disagree 4=agree 5=strongly agree Categorical Communication with mother y your mother and you 16 Scale as above Categorical Parental control (4) amount of television; (5) type of television sho ws; (6) bed time; and (7) food you eat. 7 item factor variable (EI 1.38, loadings <0.3) 27,250,251 Z score of factor variable based on age group 11 14; 15 17; 18 21 Factor at home when you return from 1=always or most of the time 0=sometimes, almost never, or never 252 Index variable taking the sum of the three questions Continuous index *Because of the s mall sample size in first generation mothers for skilled (n=50) and professional employment (n=25), they were combined for all generations. ~I controlled for gender instead of estimating stratified models because there were no significant interactions betw een immigrant generation and gender. Analysis. I conducted two analyses; the hazard of sexual initiation and the hazard of an adolescent birth. I modelled the age at sexual initiation using complementary log log discrete time hazard models. I used a pers on year dataset and participants were 16 I initially used a factored variable for quality of relationship that included the satisfaction with communication variable. For communication I used two va riables about whether or not the adolescent talked to her/his mother about a personal problem or dating. The HR of sexual initiation was increased for adolescents who had spoken to his/her mother about a personal problem or dating which suggests that adole scents who are dating and may be having problems are more likely to have sex and more likely to single satisfaction variable For example, in my qualitative interviews, one first generation young women who delayed sex until 19, hesitated when I asked if she talked to her

PAGE 262

246 censored at the year they experienced the event or the age of their last interview. 172 I right censored all participants at age 20 because I was concerned with the consequences of early sexual initiation and because participant s with a first sex after age 20 are no longer at risk for an adolescent pregnancy. I included age and age squared in our model because there is a concave relationship between age and first sex. I then used complementary log log discrete time hazard models to examine age at first birth. I used a person month dataset and participants were censored at the month and year they experienced the event or the age of their last interview. 172 I right censored all participants at age 20 because I was concerned with the conseque nces of an adolescent birth and it is unlikely there are public health risks for having a first birth between age 20 and 34. I included age and a piecewise polynomial spline approximating the shape of the relationship between age and age at first birth bec ause the relationship between age and age at first birth is positive and linear and then began to flatten around age 18.5. In further sensitivity analyses, I also stratified for males and females and found no substantial differences. I included all adolesc ents in the risk set, regardless of whether or not they had sex because pregnancy and birth are outcomes of sex and adolescents are at risk to have sex at any time. Excluding those who had not yet had sex would be excluding those who perhaps delayed sex in order to avoid consequences such as an early birth which would bias results. Each model included a control variable for gender, age, age squared or age spline, marriag e in adolescent birth models. I entered each covariate to the model using a stepwise approach and performed a Wald test to ensure that each additional variable contributed to the model outcomes and to examine if the covariate mediated the

PAGE 263

247 relationship betw een immigrant generation and a not working mother. The models do not include 5% of participants with no resident mother. Because more educated parents tend to have more skilled or professional employment, I controlled for education. I used ten multiple imp uted data sets created by chained equations to address missing data. 173 Family income had the highest missing at 37%, parental education and employment were missing less than 5%, and all other variables were missi ng less than 3%. I also adjusted for the complex sampling frame of Add Health, including sampling weights to the U.S. population. Qualitative sample. Between March 2013 and April 2014, I conducted qualitative semi structured life history interviews with 1 1 first generation (born in Mexico) and 10 second generation (born in the U.S. to a Mexican born parent) Mexican origin women in the Denver metro area who were 27 41 years old, the same age as the Add Health participants during that time. Women were recrui ted through active and passive snowball sampling at community organizations and health centers in the Metro Denver area. The study was approved by the Colorado Multiple Institutional Review Board. Each semi structure life history interview lasted between o ne and two hours and three women received follow up interviews. All women were compensated for their time. A bilingual bicultural research assistant transcribed the interviews in the language in which it was conducted and then translated the five Spanish i nterviews into English. After verifying the transcriptions and translation, I conducted a content analysis guided by segmented assimilation theory and the theory of gender and power using a five step process developed from Ulin 184 and Strauss and Corbin 185 reading, coding, displaying, redu cing and interpreting. Using segmented assimilation theory, I focused on family

PAGE 264

248 structure, family SES, family dynamics, language, race/ethnic identification, and education. Using the theory of gender and power I focused on the gendered division of labor, p ower, and cathexis. I also paid close attention to migration narratives. First, I read and re read each transcript, in English and in Spanish. Next, I domain coded for migration, family, reproductive health behaviors, and gender in Atlas ti during data col lection. After all interviews were complete, I then subcoded within each domain and allowed for in vivo codes and themes. I displayed the data in different ways, identified broad patterns, drew conclusions and then rechecked my conclusions with the display s, codes, and original transcripts. I used Atlas ti qualitative software to provide organization and transparency. Results Table 7.3 presents the results from the stepwise sexual initiation models. Model 1 shows that the hazard for first generation adoles cents with mothers who are not working was reduced by 43% (0.36*1.2) as compared to first generation adolescents with skilled or professional mothers. Post estimation tests found that the hazard ratio between first generation adolescents with mothers who w ere not working and third generation mothers who are not working was significantly different (p=0.045) but the hazard ration between first and second generation mothers who were not working is not significantly different. See Chapter 3, Figure 3.1. Model 2 shows that the hazard ratio for the first generation maternal communication increased, the hazard of sexual initiation was reduced 8% (1 0.92). Model 3 shows that the ha zard ratio for the first generation not working mothers remained unchanged but adolescents with greater maternal quality relationship had a 16%

PAGE 265

249 (1 0.84) reduction in the hazard of sexual initiation compared to those with lower maternal relationship quality remained significant. Model 4 shows that the hazard ratio for the first generation with mothers not working remained unchanged and the hazard for parental control was not different from 1 alth ough the Wald test was significant. Model 5, the final model, shows that the hazard ratio for the first generation with mothers not working remained unchanged and that the covariates did not mediate the interaction of immigrant generation and a mother not Also of note, eve n after controlling for all the variables in the model, including the quality of the relationship with the mothers, the presence of a resident father was still associated with a 27% (1 0.73) reduction in the hazard of sexual initiation. Table 7 3 : Hazard Ratios for Sexual Initiation, Maternal Employment, and Maternal C haracteristics Model 1 Model 2 Model 3 Model 4 Model 5 First generation 1.61 1.53 1.59 1.61 1.62 (0.72 3.60) (0.67 3.52) (0.70 3.60) (0.72 3.64) (0.71 3.67) Second generation 1.03 1.02 1.01 1.02 1.01 (0.73 1.45) (0.72 1.45) (0.72 1.43) (0.73 1.44) (0.71 1.46) Third generation Ref. Ref. Ref. Ref. Ref. Unemployed Mother 1.21 1.21 1.21 1.21 1.22 (0.81 1.79) ( 0.82 1.78) (0.82 1.79) (0.82 1.78) (0.83 1.81) Unskilled Mother 1.06 1.03 1.04 1.04 1.03 (0.65 1.71) (0.64 1.68) (0.63 1.70) (0.63 1.70) (0.63 1.71) Professional Mother Ref. Ref. Ref. Ref. Ref. First gen X not working mother 0 .36* (0.15 0.87) 0.37* (0.15 0.90) 0.34* (0.14 0.84) 0.34* (0.14 0.83) 0.34* (0.14 0.84) First generation X unskilled mother 0.46 (0.18 1.18) 0.49 (0.18 1.30) 0.46 (0.18 1.20) 0.46 (0.17 1.20) 0.46 (0.17 1.20)

PAGE 266

250 Table 7 3 : Hazard Ratios for Sexual Initiation, Maternal Employment, and Maternal C haracteristics First gene ration X skilled/profess mother Ref. Ref. Ref. Ref. Ref. Second generation X unemployed mother 0.69 (0.41 1.18) 0.69 (0.41 1.17) 0.68 (0.41 1.14) 0.69 (0.41 1.15) 0.69 (0.41 1.16) Second gen X unskilled mother 0.78 (0.43 1.43) 0 .80 (0.44 1.47) 0.80 (0.43 1.47) 0.79 (0.43 1.46) 0.80 (0.42 1.49) Second gen X skilled/profess mother Ref. Ref. Ref. Ref. Ref. Quality of communication with mother 0.92** (0.87 0.98) 1.03 (0.94 1.12) 1.03 (0.94 1.14) 1.04 (0 .94 1.14) Quality of relationship with mother 0.84* (0.72 0.97) 0.83* (0.72 0.97) 0.83* (0.72 0.96) Parental control 0.97 0.97 (0.90 1.05) (0.90 1.05) Supervision 0.98 (0.87 1.11) Age 16.08*** 16.00*** 16 .06*** 16.07*** 16.06*** (9.62 26.90) (9.57 26.74) (9.63 26.79) (9.63 26.82) (9.61 26.83) Age X age 0.92*** 0.92*** 0.92*** 0.92*** 0.92*** (0.91 0.94) (0.91 0.94) (0.91 0.94) (0.91 0.94) (0.91 0.94) Male 1.28** 1.30** 1.32** 1.32 ** 1.32** (1.07 1.54) (1.09 1.56) (1.10 1.58) (1.10 1.58) (1.10 1.58) Mother's education 1.02 1.02 1.02 1.02 1.02 (0.98 1.06) (0.98 1.06) (0.98 1.06) (0.98 1.06) (0.98 1.06) Resident father 0.72* 0.73* 0.73* 0.73* 0.73* (0.56 0.92) (0.56 0.94) (0.56 0.94) (0.56 0.95) (0.56 0.96) Family income 0.97 0.97 0.97 0.96 0.96 (0.84 1.11) (0.85 1.10) (0.84 1.10) (0.84 1.10) (0.83 1.10) Constant 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0 .00 0.00) (0.00 0.00) (0.00 0.00) Wald test Prob> F <0.01 <0.05 <0.05 <0.05 Observations 10,007 10,007 10,007 10,007 10,007 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Table 7.4 presents the results from the stepwise adolescent birth models, ha zards discussed are in bold. Model 1 shows that the hazard for first generation adolescents with mothers who are not working was reduced by 90% (1 0.10) as compared to first

PAGE 267

251 generation adolescents with skilled or professional mothers. Post estimation tests found a significant difference (p=0.007) between first generation adolescents with a mother not working and third generation adolescents with a mother not working and significant difference (p=0.004) between first generation adolescents with a mother not working and second generation adolescents with a mother not working. The hazard of an adolescent birth was reduced by 82% (1 0.18) for adolescents with unskilled mothers compared to adolescents with professional mothers for the first generation only. Post estimation tests found a significant difference (p=0.085) between first generation adolescents with an unskilled mother and third generation adolescents with an unskilled mother and a significant difference (p=0.013) between first generation adolescents wi th an unskilled mother and second generation adolescents with an unskilled mother. See Chapter 3, Figure 3.3. Model 2 shows that the hazard ratio for the first generation with a mother not aternal communication was not significantly different from 1. Model 3 shows that the hazard ratio for the first generation with a mother not working remained unchanged and the erent than 1. However, as the quality of relationship with the mother increased the hazard of an adolescent birth was reduced by 25% (1 0.75) (p=0.06). Model 4 shows that the hazard ratio for the first generation with a mother not working remained unchange d and the hazard for parental control was not different from 1 and the quality of the relationship with the mother lost significance although the hazard ratio remained similar. Model 5, the final model, shows that the hazard ratio for first generation with a mother not working remained unchanged and that the covariates did not mediate the interaction of immigrant

PAGE 268

252 generation and a not working mother nor the interaction of immigrant generation and an zard of an adolescent birth was reduced by 22% (1 0.78) p=0.073. The hazard ratio for having a resident father in the home was not significantly different from 1 in any of the models. Table 7 4 : Hazard Ratios for Adolescent Birth, Maternal Employment, and Maternal C haracteristics Model 1 Model 2 Model 3 Model 4 Model 5 First generation 2.26 2.28 2.28 2.16 2.25 (0.72 7.16) (0.71 7.28) (0.72 7.28) (0.64 7.27) (0.68 7.48) Second generation 0.59 0.59 0.57 0.56 0.54 (0.23 1.56) (0.23 1.56) (0.22 1.53) (0.22 1.45) (0.22 1.34) Third generation Ref. Ref. Ref. Ref. Ref. Not working Mother 0.92 0.92 0.92 0.91 1.12 (0.48 1.76) (0.48 1.75) (0.50 1.72) (0.48 1.72) (0.56 2.2 6) Unskilled Mother 0.95 0.95 0.94 0.94 0.96 (0.39 2.30) (0.39 2.29) (0.40 2.21) (0.40 2.21) (0.41 2.26) Professional Mother Ref. Ref. Ref. Ref. Ref. First gen X not working mother 0.10*** 0.10*** 0.09*** 0.09*** 0.09*** (0.03 0.3 5) (0.03 0.36) (0.02 0.35) (0.02 0.36) (0.02 0.34) First gen X unskilled mother 0.18** 0.18** 0.17** 0.16** 0.16** (0.05 0.62) (0.05 0.61) (0.05 0.56) (0.05 0.56) (0.05 0.53) First gen X skilled/profess mother Ref. Ref. Ref. Ref. Ref. Second gen X not working mother 0.76 0.76 0.77 0.73 0.74 (0.25 2.28) (0.25 2.31) (0.26 2.32) (0.25 2.16) (0.26 2.14) Second gen X unskilled mother 0.69 0.69 0.72 0.71 0.75 (0.20 2.37) (0.20 2.40) (0.21 2.52) (0.20 2.48) (0.2 3 2.49) Second gen X skilled/profess mother Ref. Ref. Ref. Ref. Ref. Quality of maternal communication 1.01 1.25 1.23 1.24 (0.85 1.21) (0.90 1.75) (0.88 1.73) (0.89 1.73)

PAGE 269

253 Table 7 4 : Hazard Ratios for Adolescent Birth, Maternal Employment, and Maternal C haracteristics Quality of maternal relationship 0.75 0.77 0.80 (0. 55 1.01) (0.55 1.06) (0.58 1.10) Parental control 1.12 1.12 (0.95 1.32) (0.95 1.31) Supervision 0.78 (0.60 1.02) Age 1.06*** 1.06*** 1.06*** 1.06*** 1.06*** (1.05 1.07) (1.05 1.07) (1.05 1.07) (1.05 1.07) (1.05 1.07) Age spline 0.90*** 0.90*** 0.89*** 0.89*** 0.90*** (0.86 0.93) (0.86 0.93) (0.86 0.93) (0.86 0.93) (0.86 0.93) Male 0.55* 0.55* 0.56* 0.55* 0.53** (0.35 0.88) (0.35 0.87) (0.36 0.87) (0.35 0.87) (0.34 0.82) Married 7.46*** 7.45*** 7.49*** 7.56*** 8.06*** (3.71 15.01) (3.69 15.04) (3.78 14.87) (3.83 14.93) (4.10 15.86) Mother's education 0.96 0.96 0.95 0.95 0.95 (0.88 1.04) (0.88 1.04) (0.87 1.03) (0.87 1.03) (0.87 1.03) Resident father 0.84 0.84 0. 84 0.82 0.85 (0.52 1.37) (0.52 1.35) (0.52 1.33) (0.51 1.33) (0.53 1.37) Family income 0.77 0.77 0.77 0.77 0.76* (0.59 1.02) (0.59 1.02) (0.59 1.01) (0.59 1.01) (0.58 1.00) Constant 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** (0 .00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) Wald test Prob> F p=0.89 p=0.14 <0.05 <0.05 Person years 136,324 136,324 136,324 136,324 136,324 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Discussion Contrary to my hypothesis, qua lity of the maternal relationship, quality of relationship between having a mother not working in the first generation and sexual initiation or risk for an adolescent bi rth. It is possible that the variables are not measuring what I wanted them to measure. For example, having a mother at home before school, after school, or before bed does not actually mean the mother is providing supervision. It is also possible that the re are other factors that I did not measure that better account for the reasons why having a not working mother is protective against early sexual initiation

PAGE 270

254 in the first generation alone. In future work, I could add in additional variables to test if cons ervative sexual values mediate the relationship between a mother not working and a reduced hazard of sexual initiation and adolescent birth in the first generation. Additionally, structural equation modeling may better capture latent variables such as pare ntal supervision and control. Immigrant generation. Descriptive results (Table 7.1) and prior research 89,92 found that first and second generation families are more likely to consist of two biological parent families than the third generation. According to a recent report, married mothers are more likely to say that they are not working in order to care for their family, as compared to single or cohabitating mothers. 141 Married stay at home mothers are more likely to be foreign born than single or cohabiting mothers and married women with working husbands generally have a higher SES than single women. 141 Cristina and employed and skilled. Perhaps one of the reasons having a mother at home is more beneficial in the first and second g eneration is because migrant mothers who stay home are more likely to be married to a working husband and be of higher SES and may be more likely to choose to stay home and native Mexican origin mothers, particularly of lower SES, may be at home because th ey are unable to find a job due to a disability or an economic downturn, low education or discrimination. Thus, in the first and second generation, migrant mothers with married working husbands may be choosing to stay home because they can afford to and be cause as discussed in Chapter 5, taking care of the family and being in the home was a common gendered discourse in migrant families. Having a stay at home mother in native born Mexican origin families may represent low

PAGE 271

255 education and a lack of employment o ptions because a third of stay at home mothers in the U.S. are impoverished and only a tenth of working mothers are impoverished. 141 Further, mothers who have not chosen to stay at home may not be providing as much supervision and support because of economic circumstances or because they are trying to find work. I controlled for the presence of a father and family income in the models but I may not have captured underlying family dynamics and reasons behind the employment choices of mothers. Relationship quality and communication. Consi stent with prior research, 190,77,82,247,248 maternal relationship quality and maternal communication were associated with a reduced hazard of sexual initiation although only maternal relationship quality was associated with a reduced hazard of an adolescent birth. My qualitative r esults support this finding. All of the six women in my qualitative sample who had their first sex prior to age 17 and six of nine women who had an adolescent birth had mothers working outside the home. Fewer of these women described satisfactory relations hips and communication with their mothers than those who delayed their first sex and birth. Women who delayed their first sex and their first births described more frequent communication and a better relationship quality with their mothers. Second generat ion and talk with her daughter, however, her mother did not work much outsi de the home. Among all women in my qualitative sample, women who believed their mothers were constantly at work or had single mothers described more disappointing relationships with their mothers. A recent demographic survey found that married stay at hom e

PAGE 272

256 mothers were more likely to report they stay at home to care for children than single and cohabitating mothers who said they were at home due to difficulty finding a job. 141 This serves as a reminder that we should not judge these mothers has not valuing time or relationships with the generation, college educated Jessica, who delayed sex and has never been pregnant described survival parentin g well, I thought it [relationship with her parents] was normal but when I talk to people that were born here, Caucasians, it was a little different. I know my parents were more focused on work and providing for the family than other aspects of education us. Looking back now I can see how stressful situation the situation was, we had no money and my p really develop a relationship with us. The qualitative data demonstrated how nuanced relationships are and how difficult they are to measure on a Likert scale. Jocelyn described the limitation s of her housework or stuff like that. She was really absent most of our lives growing up because Jocelyn would probably not have rated her satisfaction with the quality of her relationship with her mother very high, second generation Alicia, who had her first sex at 16 and first birth at was a good mom. She er in the interview I asked Alicia what values or lessons she learned from her mother. She became teary and when I asked

PAGE 273

257 harder when I was younger. I wish she would hav mother did not ask her about her school day but immediately listed off chores for her to do. This is another example of real versus ideal, as discussed in detail in Chapter 5. Alicia of her relationship with her mother call into question what she wanted her relationship with her mother to be like and what it may have actually been like. Quantitative survey questions are not able to captu re this nuance and ambivalence and they may be prone to participants answering in the ideal instead of the real. Supervision. Maternal supervision did not significantly influence the hazard of sexual initiation but it was associated with a reduction in the hazard of an adolescent birth (p=0.073). There has been conflicting evidence on the influence of supervision, some studies have shown no influence of supervision on risky sexual behaviors 190 and others have shown supervision decreased risky sexual behaviors. 77,82 When it came to supervision, four of the six women in my qualitative who experienced sexual initiation prior to age 17, specifically reported that their mothers were not home with them afterschool and sometimes into the evenin g. Two of the four women recalled neighbors or grandparents helping out. Eight of the 15 women who delayed sex reported their father was always home and another young woman said she felt like her parents were home a lot, despite their work schedules, because of their parents made an effort to have dinner together and spend time together as a family. Only one young woman specifically recalled that when her single mother was at work, she was often alone with her siblings.

PAGE 274

258 Similarly, a greater number of women who delayed their first birth reported their parents, supervision did not adequately captured supervision because supervision was related to delayed sexual initiation and adolescent birth in my qualitative sample. Parental control. Parental control was not significantly related to sexual initiation or adole scent birth in the quantitative models. I did not specifically ask qualitative participants about parental control although this usually came up in conversations around dating. Women with parents with very clear dating expectations and rules that were acco mpanied by parental communication and supervision often delayed sex and their first birth. However, very high parental control, when the young women were not allowed out of the house without their brothers or were not allowed to date at all was associated with more extreme gendered discourses, as described in Chapter 5 and often led to sneaking out and hiding relationships. Parental control may not have a linear relationship with sexual risk behavior. In the quantitative models, similar to prior research us ing Add Health, 27,250,251 the variable I used to measure parental control had relatively low loading scores onto the factor variable (<0.3) and thus may also not be validly measuring the complexity of parental control. Lee and Hahm 27 used the same parental control variables as I did but used an index instead of a factor variable and found more parental control reduced the odds of having four or more sexual partners but was not a ssociated with history of STIs, regret of sex after alcohol use or condom use at most recent sex. Perhaps future research can attempt to better measure parental control or monitoring and their influence on adolescent health outcomes.

PAGE 275

259 Fathers. In the final stepwise model for sexual initiation, (Table 7.3, Model 5) after controlling for maternal relationship quality and communication, supervision, control, maternal education, maternal employment and family income, having a resident father in the home still re duced the hazard of sexual initiation. However, having a resident father in the home did not reduce the hazard of an adolescent birth in any of the models (Table 7.6). Prior studies have found that Mexican origin adolescents described different relationshi ps with their mothers and fathers 25,253,254 and a qualitative study 254 found that adolescents described relationships with their mothers as more open, affectionate, and close than with their fathers. Results as these can lead t o a negative portrayal of Mexican origin fathers and may perpetuate a lack of research on fathers. My qualitative results suggest that close supportive relationships with fathers were particularly important to decision making about reproductive health beha viors in Mexican origin daughters. Only two of the six women (33%) who had their first sex prior to age 17 described close relationships with their fathers, compared nine of 15 (60%) of young women who delayed their first sex. Similarly, only about 18% of the nine young women who experienced an adolescent birth described close or supportive relationships with their fathers or stepfathers compared to 67% of 12 young women who delayed their first birth. My qualitative results demonstrated the importance of ha ving a father in the home, a supportive relationship with a father as well as the timing of the loss of a father. Some women did not have a close or supportive relationship with a father because there was not a father in the home. Three women grew up wit hout father figures in their childhoods and another three women described negative relationships with their father figures. For example, Juana, a first generation young woman who had an unplanned

PAGE 276

260 a father figure or who described a negative relationship wi th their father figure during childhood, experienced an adolescent birth. Although recall bias is certainly present, all three women who described bad or negative relationships with father figures were clear that the relationships were not good prior to th eir pregnancies. In fact, Mariela married at 15 specifically to get out of the house because of turmoil with her mother and stepfather. A few women described rather neutral relationships with their fathers, they with disappointment. Eleven women describe positive and supportive relationships with their fathers, nine of whom delayed their first birth. For example, Arely, who planned her first child at 28 years old, des a lot of people. He was a really good father, a very good father. He would always give us about her friends, sh Isabel, 29, a college educated first generation young woman who is not yet married and has never been pregnant, described her father as hard and long hours but was fun and played with them when he got home. She believed the lack of a father figure may be a reason for adolescent births. makes me really think why? Even most of my extended family and good friends, I

PAGE 277

261 were mar do with having a good man in the family. Having that good male role model. My sister, my younges the only thing [her emphasis] that changed in our life and in our family that could have maybe had something to do know when it comes to the other Hispanic girls if it has something to do with different world and you would think people would be more conscious about protection and more open to just feel like it has to do with the mom and dad, so it would be interesting to know what these Latina girls, what their upbringing were. Like if their dads in Mexico were working, or if their dads w ere deported, or if their dads were iced [detained], one of those stats. My quali and having a supportive relationship was associated with more egalitarian gendered roles and was associated with delayed childbirth and improved educational outcomes. Although timin g is difficult to gauge in a cross sectional study, as Isabel suggested, the loss of a father figure during adolescence may be more detrimental than the loss of a father figure at other ages. For example, Isabel stated above that her sister had a supportiv e father until age 10, yet she still had an early unmarried birth. Second generation Mariela, who married at 15 and had her first daughter at 17 He was just the typical d father was deported at age 11. Her mother remarried out of convenience and her stepfather was controlling and jealous. Despite her supportive biological father during her childhood, the loss of him and the presence of controlling stepfather during adolescence

PAGE 278

262 reasons is that I kind of got married early was not just because, obviously because I love Similarly, a supportive stepfather during adolescence may buffer a negative relationship with a biological father. First gen eration Yesinia, who had her first birth at 20 and is currently a medical assistant with a daughter in college, reported her biological mother married when she migrated to the U.S. when Yesinia was 10, was supportive and loving. My [step]dad, I remember 30 minutes of readin g every day until I got used to it. He really did win his role as a dad. Of course in the beginning I would call him always be my dad. father was closely related to her future academic and reproductive health decisions. The quantitative results partially supported these findings. Perhaps ha ving a father in the home is sufficient to reduce early sex, possibly due to more parental monitoring and more resources and stability in two parent families, but relationship quality with a father may be more important to reducing the hazard of an adolesc ent birth than merely the presence

PAGE 279

263 communication and reproductive health outcomes to further test this qualitative conclusion. Future work is also needed on better qualitative un derstanding of father child relationships. Conclusion I did not find quantitative evidence that maternal relationship quality, communication, supervision or control were the main reasons why first generation adolescents with mothers not working had a low er hazard of sexual initiation and adolescent birth than subsequent generations with mothers not working. However, the negative association with maternal relationship and communication quality and the hazard of sexual initiation, did support qualitative da ta that maternal relationship quality and communication were associated with a delay in sexual initiation and first birth. Future quantitative analyses could use structural equation modeling to better capture reasons why a mother at home is more protective in the first generation. My qualitative results suggest finding ways to improve maternal adolescent relationship quality and communication such as providing a living wage so that mothers, and fathers, can afford to spend more time with their children, pro viding economic and parent resources to single mothers, or teaching parents how they might improve their relationships despite time constraints. In addition to understanding mother daughter relationships, highlighting positive father daughter relationshi ps in Mexican origin families can help break down stereotypes of machismo and promote a more positive view of Mexican fathers. This may inspire fathers to improve their relationships with their daughters and take an active and supportive role in their live s if they know how important they are. More research is

PAGE 280

264 understanding how fathers can help their daughters achieve their goals, delay sex and early childbearing, and incre ase contraceptive use, we can better design education programs for fathers. Bringing the importance of fathers into public discourse may provide support for immigration reform that keeps families together and prevents deportation of fathers and other polic ies that may improve working wages, hours, and conditions for workers so that mothers and fathers can spend more time with their families. M ediators of Parental Education Background Prior research has found higher parental education protects against risk y adolescent behaviors and improves health and academic outcomes and segmented assimilation theory asserts that higher parental education is associated with an increased economic mobility in second generation adolescents. 31,88,93,115,118 Extending literature and theory, in Chapter 3, I found that fat increased the hazard of sexual initiation in the first generation. I speculated that higher educated migrant parents may be more liberal in their sexual views and may be more permissive with dating rules. In fact, Gonzalez Lopez 195 has suggested that migrants from rural areas of Mexico are more likely to hold conservative sexual views and practices, such as the importance of female virginity, than migrants from urban areas largely due to higher education and employment options for men and women in urban areas. As I argued in Chapter 4, the migration process, including where migrants are from and why

PAGE 281

265 they migrate is related to family structure, SES and subsequent adolescent outcomes, although seg mented assimilation theory does not address these factors. I did not have the diversity in family socioeconomic status in my qualitative analysis to explore differences in sexuality discourses or dating rules between lower and higher educated families, no r did I have enough detailed description of the area in Mexico where the parents of second generation participants were from. However, as discussed in Chapter 6, clear parental messages of virginity or waiting to have sex was associated with a delay in sex ual initiation. Additionally, all families had some level of gendered discourses with different treatment of daughters and sons, especially when it came to sexuality. I chose variables in Add Health that represented expectations and communication about sex and one variable related to gender. Figure 7.2 shows the conceptual model for my hypotheses. I hypothesized that parental disapproval of sex and contraception, gendered parental expectations, parental communication about sex, and parental religiosity coul d mediate the relationship between the hazard of first sex and parental education in first and second generation adolescents. Figure 7 2 Methods Data. I used data from all four waves (1994 2008) of the National Longitudinal Study of Adolescent Health (Add Health). The survey began in 1994 with 20,475 7 12 Moderator: immigrant generation Parental education Age at sexual initiation IV DV M ediators: parental expectations of sex, communication about sex, religiosity, gendered parental expectations

PAGE 282

266 graders and participants were interviewed again in 1996 (Wave II), 2001 2002 (Wave III), and 2007 2008 (Wave IV) 160 I restricted the sample to Mexican origin respondents. I excluded five participants missing on ever having sexual intercourse, six participants missing on the presence of a parent in the home and seven participants with an age at first sex prior to 11 due to the unlikelihood that sex was consensual at that age. 165 After excluding respondents who were missing without a sample weight, the final sample size was 1,637 Mexican origin participants. Table 7.5 presents independent variables and how they were measured. All variables were measured at Wave I except for a time varying variable for age. Table 7 5 : Independent V ariables Independent variables Measured Type of variable Immigrant generat ion First=Born in Mexico to Mexican born parents Second=Born in the U.S. to at least one Mexican born parent Third=Born in the U.S. to two U.S. born parents and classifies ethnicity as Mexican Categorical Number of years of education: 0= never attended school 8= 8 th grade or less 10= no high school degree but more than 8 th grade 12=GED or diploma 16=college graduate Continuous As above Continuous Control variables Age Age in years Time varying continuous Age squared Age squared Time varying continuous Gender ~ 1=male 0=female Dichotomous Family income Logged annual household income, adjusted for inflation and family size at Wave I Continuous Resident father Presence of a father figure (biological, step) in the home Dichotomous 1=not working outside of the home 2=unskilled employed 3=skilled/professionally employed* Categorical nt As above Categorical

PAGE 283

267 Table 7 5 : Independent V ariables Parental expectations about sex intercourse with someone who was special to you and whom you knew well, like a steady 1=strongly approv e 2=approve 3=neither approve or disapprove 4=disapprove 5=strongly approve Categorical As above Categorical contraception Scale as above. Categorical contraception As above Categorical Equal parental gendered expectations (parent survey) k is the most 1=be well behaved 2=be popular 3=think for himself 4=work hard 5=help others The variable was coded as a 0 if the parent selected a different number for each question and 1 if they selected the same numbe r. Dichotomous Parental communication about sex How much the parent has talked to the child about: (1) the negative things that would happen if she/he got pregnant; (2) the dangers of getting an STI; (3) the bad respect if she/he had sex; (4) moral issues of not having sexual intercourse; (5) birth control; and (6) sex. 1=not at all 2=somewhat 3=a moderate amount 4=a great deal 6 item factor variable (EI 3.8, loadings >0.7) Factor Parental religiosity Higher score, more frequent, important or agreement. (1) You believe the scriptures are the word of God; (2) frequency of attendance of religious services; (3) how important religion is; and (4) how often you pray. 4 item factor variable (EI 2.78, loading s>0.7) Factor *Because of the small sample size in first generation mothers for skilled (n=50) and professional employment (n=25), they were combined for all generations. ~I controlled for gender instead of estimating stratified models because there were no significant interactions between immigrant generation and gender. Analysis. I modelled the age at sexual initiation using complementary log log discrete time hazard models. I used a person year dataset and participants were censored

PAGE 284

268 at the year they e xperienced the event or the age of their last interview. 172 I right censored all participants at age 20 because I was concerned with the consequences of early sexual initiation and because participants with a first sex after age 20 are no longer at risk for an adol escent pregnancy. I included age and age squared in our model because there is a concave relationship between age and first sex. I performed stepwise models beginning with the baseline model including education and immigrant generation and the control variables. I then added each potential mediator into the model and performed a Wald test to ensure the variable added meaning to the model. I used ten multiple imputed data sets created by chained equatio ns to address missing data. 173 Family income had the highest missing at 37%, parental education and employment were missing less than 5%, and all other variables were missing less than 5%. The variables from the p arent survey, parental gendered expectations, parental communication about sex and parental religiosity and had about 20% missing. I also adjusted for the complex sampling frame of Add Health, including sampling weights to the U.S. population. Results Tab le 7.6 hazard of sexual initiation was positive in the first generation (1.07*1.01=1.08, p<0.10), and was not significantly different from one in the second (1*1.01=1.01) and third gene 0.84) hazard of sexual initiation. Model 3 again shows no change in the interaction

PAGE 285

269 between immigrant generati on but equal gendered parental expectations were associated with a 25% (1 0.75) reduction in the hazard of sexual initiation. Model 5 shows that the decreases (p=0.081 to p = 0.104) but the HR remains the same at 1.07. Parental communication about sex was associated with a 15% (1 1.15) increase in the hazard of sexual initiation. Model 6, the final model, shows the HR for the interaction between nt generation remains the same and the significance is p=0.109. The hazard for religiosity is not significantly different from 1. However, about sex remained significant Table 7 6 : Hazard R atios for S exual I nitiation, M Education, and Mat ernal C haracteristics Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 First generation 0.38* (0.16 0.90) 0.33* (0.14 0.78) 0 .32* (0.14 0.77) 0.32* (0.13 0.78) 0.40 (0.16 1.02) 0.41 (0.16 1.05) Second generation 0.76 (0.30 1.94) 0.73 (0.28 1.86) 0.72 (0.28 1.84) 0.72 (0.27 1.88) 0.84 (0.31 2.23) 0.85 (0.31 2.27) Third generation Ref. Ref. Ref. Ref. Ref. Ref. Mother's edu 1.00 (0.94 1.06) 0.99 (0.94 1.05) 0.99 (0.93 1.05) 0.99 (0.93 1.05) 0.99 (0.94 1.06) 1.00 (0.94 1.06) First gen X mom edu 1.07 (0.99 1.16) 1.09* (1.00 1.18) 1.09* (1.00 1.18) 1.09* (1.00 1.18) 1.07 (0.99 1.17) 1.07 (0.98 1.17) Second gen X mom edu 1.01 (0.93 1.09) 1.01 (0.93 1.09) 1.01 (0.93 1.10) 1.01 (0.93 1.10) 1.00 (0.92 1.09) 1.00 (0.92 1.09)

PAGE 286

270 Table 7 6 : Hazard R atios for S exual I nitiation, M Education, and Mat ernal C haracteristics Third gen X mom edu Ref. Ref. Ref. Ref. Ref. Ref. Mom's disapproval of sex 0.84*** (0.77 0.91) 0.85*** (0.78 0.92) 0.85*** (0.79 0.92) 0.85*** (0.78 0.92) 0.85*** (0.78 0.92) Mom's disapproval birth control 0.98 (0.92 1.05) 0.99 (0.92 1.06) 0.99 (0.92 1.06) 0.99 (0.92 1.0 6) Equal gendered parental expectation 0.75* (0.57 0.99) 0.72* (0.56 0.93) 0.72* (0.55 0.93) Parental commun. about sex 1.15* (1.02 1.31) 1.15* (1.02 1.31) Parental religiosity 0.97 (0.87 1.07) Ag e 15.95*** 15.66*** 15.66*** 15.54*** 15.65*** 15.63*** (9.49 26.82) (9.29 26.40) (9.30 26.37) (9.19 26.27) (9.26 26.43) (9.25 26.40) Age x Age 0.92*** 0.92*** 0.92*** 0.92*** 0.92*** 0.92*** (0.91 0.94) (0.91 0.94) (0.91 0.94) (0.91 0.94) (0.91 0.94) (0.91 0.94) Male 1.28** 1.17 1.16 1.17 1.19* 1.20* (1.08 1.52) (0.98 1.40) (0.98 1.38) (0.98 1.39) (1.00 1.41) (1.01 1.42) Not working mother 0.97 (0.75 1.25) 0.95 (0.75 1.21) 0.95 (0.75 1.21) 0.95 (0.75 1.20) 0.95 (0.75 1.21) 0.95 (0.75 1.21) Unskilled mother 0.93 (0.67 1.29) 0.92 (0.68 1.26) 0.92 (0.68 1.26) 0.91 (0.67 1.24) 0.93 (0.68 1.26) 0.93 (0.69 1.26) Skilled/Pro fessional mother Ref. Ref. Ref. Ref. Ref. Ref. Family income 0.97 (0.85 1.11) 0.97 (0.85 1.10) 0.97 (0.85 1.10) 0.99 (0.86 1.13) 0.98 (0.86 1.11) 0.98 (0.86 1.11) Resident father 0.72* (0.56 0.92) 0.75* (0.58 0.97) 0.75* (0.59 0.97) 0.75* (0.58 0.97) 0.76* (0.59 0.99) 0.76* (0.59 0.99)

PAGE 287

271 Table 7 6 : Hazard R atios for S exual I nitiation, M Education, and Mat ernal C haracteristics Constant 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) Wald test Prob> F <0.001 <0.001 <0.001 <0.001 <0.001 Person years 10,007 10,007 10,007 10,007 10,007 10,007 *** p<0.001, ** p<0.01, p<0.05, p<0.10 models. Model 1, hazard for sexu al initiation increased by 1% (1.09*0.92) in the first generation, and by 7% (1.15*0.92) in the second generation, but decreased by 8% (1 0.92) in third generation. immigrant generat 0.84) reduction in the hazard of sexual initiation. Model 3 again shows no change in the isapproval of contraception and sexual initiation was not different from 1. Model 4 again shows no change in the magnitude of the hazard ratio for the interaction cation and immigrant generation E qual gendered parental expectations wer e associated with a 25% (1 0.75) reduction in the hazard of sexual immigrant generation but parental communication about sex was associated with a 13% (1 1.13) increase in the hazard of sexual initiation (p<0.10). Model 6, the final model, and that the hazard ratio for parental religiosity was not significantly different from 1.

PAGE 288

272 However expectations, and parental communication about sex remained significant. Table 7 7 : ducation, and Paternal C haracteristic Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 First generation 0.20** (0.07 0.54) 0.18** (0.06 0.50) 0.18** (0.06 0.50) 0.19** (0.07 0.57) 0.22** (0.08 0.62) 0.23** (0.08 0.63) Second generat ion 0.13*** (0.05 0.32) 0.13*** (0.05 0.33) 0.13*** (0.05 0.33) 0.13*** (0.05 0.33) 0.16*** (0.06 0.39) 0.16*** (0.06 0.41) Third generation Ref. Ref. Ref. Ref. Ref. Ref. Father's education 0.92** (0.87 0.97) 0.92** (0.87 0.98) 0.92** (0.87 0.98) 0.93* (0.87 0.98) 0.93* (0.88 0.99) 0.93* (0.88 0.99) First gen X father's education 1.10* (1.00 1.20) 1.10* (1.01 1.21) 1.10* (1.01 1.21) 1.10 (1.00 1.21) 1.09 (0.99 1.19) 1.09 (1.00 1.19) Second gen X father's education 1.15*** (1.06 1.24) 1.15*** (1.06 1.24) 1.15*** (1.06 1.24) 1.15*** (1.06 1.24) 1.13** (1.05 1.22) 1.13** (1.05 1.22) Third gen X father's education Ref. Ref. Ref. Ref. Ref. Ref. Father's disapp roval of sex 0.84*** (0.77 0.91) 0.86** (0.78 0.95) 0.88** (0.80 0.97) 0.87** (0.79 0.96) 0.87** (0.79 0.96) Father's disapproval birthcontrol 0.95 (0.87 1.04) 0.95 (0.87 1.04) 0.96 (0.88 1.04) 0.96 (0.88 1.05) Equal parental gendered expectation 0.75* (0.56 1.00) 0.72* (0.53 0.97) 0.72* (0.53 0.97) Parental communic ation about sex (0.99 1.30) (0.99 1.30) Parental religiosity 0.96 (0.86 1.07)

PAGE 289

273 Table 7 7 : ducation, and Paternal C haracteristic Age 14.15*** (6.94 28.86) 13.96*** (6.80 28.65) 13.86*** (6.77 28.37) 13.57*** (6.60 27.89) 13.61*** (6.64 27.89) 13.56*** (6.63 27.75) Age x age 0.93*** 0.93*** 0.93*** 0.93*** 0.93*** 0.93*** (0.91 0.95) (0.91 0.95) (0.91 0.95) (0.91 0. 95) (0.91 0.95) (0.91 0.95) Male 1.45*** 1.26* 1.24* 1.26** 1.28** 1.29** (1.23 1.72) (1.05 1.51) (1.04 1.48) (1.06 1.50) (1.07 1.53) (1.08 1.55) Father not working 0.99 (0.66 1.49) 0.97 (0.65 1.45) 0.96 (0.64 1.44) 0.97 (0.65 1 .46) 0.97 (0.65 1.43) 0.96 (0.65 1.42) Father unskilled 1.09 (0.86 1.39) 1.08 (0.86 1.36) 1.08 (0.86 1.36) 1.08 (0.85 1.37) 1.08 (0.85 1.36) 1.08 (0.86 1.36) Father skilled/prof essional Ref. Ref. Ref. Ref. Ref. Ref. Family income 0.97 (0.83 1.13) 0.96 (0.82 1.12) 0.95 (0.81 1.12) 0.96 (0.83 1.12) 0.95 (0.82 1.11) 0.96 (0.82 1.12) Constant 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** (0.00 0.00) (0.00 0.00) (0.00 0.00) (0.00 0.00) ( 0.00 0.00) (0.00 0.00) Wald test Prob> F <0.001 <0.001 <0.001 <0.001 <0.01 Person years 8,039 8,039 8,039 8,039 8,039 8,039 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Discussion Contrary to my hypothesis, parental expectations about sex, contrac eption, gendered expectations, communication about sex, and parental religiosity did not mediate the relationship between parental education and immigrant generation. Again, this does not mean that higher educated first and second generation migrant parent s do not have more liberal sexual views. It could be that the variables are not adequately capturing what I intended for them to capture. For example, parental communication about sex was positively associated with the hazard at sexual initiation. This cou ld be because parents were more likely to talk to their adolescent about sex they found out the adolescent had sex or if their parents were worried about their adolescent having sex and not because the

PAGE 290

274 parents were trying to prevent it from happening. It i s also plausible that there are other reasons for the association between immigrant generation, parental education and sexual initiation that I did not account for in my model. Perhaps there is no difference in parental expectations or disappointment about the adolescent having sex but more educated first and second generation migrant parents may allow dating at an earlier age or allow for a later curfew or are generally providing less supervision. Immigrant generation and gendered parental expectations. I did not have enough variation in parental education in my qualitative sample to explain the positive association between sexual initiation and parental education in the first and second generations. The mean age of parental education of qualitative parti cipants was less than 10 years, placing them on the lower end of parental education. As discussed in detail in Chapters 5 and 6, despite the lower educational level of these families, there was a lot of variation in gendered discourses, roles, and sex educ ation and expectations in families. All of the women interviewed described at least some level of gendered discourses from independence from her parents is gained through marriage. It is interesting that the dichotomous variable I used as a proxy for equal gendered expectations, which only assessed whether the parent had chosen the same important thing for a boy and girl to learn or whether they chose a different thing for a boy and girl, significantly reduced the This suggests that parental gendered expectations that are the same for boys and girls may reduce the risk of sexual init iation in their adolescents. It is important to

PAGE 291

275 acknowledge that a parent who believes the same things are important for girls and boys does not necessarily treat daughters and sons similarly. I concluded in Chapter 5 that extreme degrees of gendered roles and discourses in families was associated with early sexual initiation and adolescent birth. The dichotomous variable can clearly not measure that level of complexity although both quantitative and qualitative results suggest gender plays a role in reprod uctive health outcomes of Mexican origin adolescents. Further, gendered roles and discourses were not necessarily associated with sex education and parental communication about sex. I discussed in Chapter 6 that more second generation women than first gene ration women complained about differential sibling treatment regarding privileges and dating and more second generation recalled negotiating or rejecting this by sneaking out. There are surely other factors, such as exposure to native born non Latino peers that may influence outcomes in addition to gendered roles and discourses. Segmented assimilation theory needs to address the role of gender in the acculturation and assimilation processes of migrant families. Parental expectations and communication abo ut sex. Supporting my was associated with a reduced hazard of sexual initiation. Prior research has found that maternal disapproval of sex has reduced the risk of sexua l initiation. 247,248 I discussed in Chapter 6 that young women who received messages to delay sex, whether for the importance of virginity or because the consequences of sex could deter their futures success, had later ages at sexual initiation than young women who her told mainly not to get pregnant or were not told anything at all about sex. The quantitative variable does not

PAGE 292

276 measure messages of virginity or delaying sex, only that the adolescent believed the parent would disapprove of them having sex with a boyfriend/girlfriend now. However, the qu alitative and quantitative results show that parental expectations do matter and that conveying to adolescents that they should delay sex may result in the adolescent listening and delaying sex. Future qualitative research on what types of parental message s about sex are the most helpful, including providing sexual and reproductive knowledge, will lead to better adolescent and parental education about sex. Parental disapproval of contraceptive use and the hazard of sexual initiation was not significantly different from 1. Parents may disapprove of their adolescent having sex but approve of contraceptive use if they do engage in sex. Dittus and Jaccard 247 found that adolescents were more likely to underestimate maternal disapproval of sex if they reported maternal approval of contraception. A three way interaction model by Khurana and Cooksey 255 found that maternal disapproval of contraception moderated the relationship between frequency of maternal sexual communica tion and sexual outcomes and was also moderated by virginity status of the adolescent. It is likely that parental disapproval of contraceptive use depends upon other contextual factors. Consistent with prior research, 248 the frequency of parental communication about sex was associat ed with an increase in the hazard of sexual initiation probably because parental communication about sex increases when parents find out or are worried that their adolescent is sexually active. Testing three way interactions between maternal disapproval of sex, frequency of sexual communication, sexual experience of the adolescent and condom use and STI history, Khurana and Cooksey 255 found that frequent maternal communication about sex and disapprov al of sex for adolescents who are

PAGE 293

277 already sexually experienced increased the risk of inconsistent condom use and STIs. Again, parental communication and expectations on sex are intertwined and related to contextual factors. Religion. Contrary to stereoty the reasons for conservative sexual values and encouragement of virginity for women, parental religiosity did not mediate the relationship between parental education and the hazard of sexual initiation o r and adolescent birth. Consistent with prior research 22,79,120 I did not find a significant association between parental religiosity and the hazard of sexual initiation or an adolescent birth. Gonzalez Lopez 217 argued that Catholic teachings are the family that enf orce these teachings, not the church. Similar to Gonazlez 217 findings, the majority of the women in my study identified as Catholic but stated they did not believe religion played a role in their life decisions. Also supporting Gonzalez protestant, had parents who were pastors and held more fundamental and conservative views on sex than the women who identified as Catholic. Two of the women desired to remain virgins until married and one of them succeeded in that. In the U.S., it is the f undamental protestant that uphold more conservative sexual views, including promotion of abstinence only education, being against abortion, and against same sex marriage. 256 Further analysis on religion is beyond the scope of my dissertation but the lack of a quantitative support for the influence of religion on the hazard of sexual initiation and the qualitative results that suggest the Protestant faith reinforcing more conservative sexual views, should caution us

PAGE 294

278 against using the Catholic religion as a reason for values or behaviors in Mexican adolescents. Conclusion While this analysis did not support my hypothesis that parental expectations about sex, communication about sex, gendered expectations and religiosity did not explain why higher parental education is associated with an increased hazard of sexual initiation, it did support my qualitative findings that parental messages of virginity and delaying sex was associated with delaying sex and that expectations of gender may be associated with delayed sex. Further, this analysis extends prior literature an d segmented assimilation theory by showing that parental education is not always associated with better outcomes and that gender should not be left out of segmented assimilation theory. Future research, including structural equation models that can capture latent variables and qualitative research on how Mexican origin parents talk to their adolescents about sex, is needed. G randparents, Siblings, and Reproductive Health Behaviors There is a dearth of research on the influence of grandparents and sibling s on adolescent reproductive health. There has been almost no research on the influence of grandparents on reproductive health behaviors in adolescents, although grandparents have been shown to improve educational outcomes, particularly in single family ho mes. 103 While there is considerably more research on siblin gs, scholars have yet to come to an agreement on the influence of siblings on reproductive health behaviors or how siblings may influence each other. In the quantitative analyses, I found that the presence of a grandparent in the home reduced the hazard o f sexual initiation and increased the odds of contraceptive use

PAGE 295

279 protecting younger siblings from risky sexual behaviors in qualitative interviews I decided to examin e the influence of siblings in my quantitative models. Consistent with my qualitative research, but challenging prior quantitative research, I found that having an older sibling in the home reduced the hazard of sexual initiation for three immigrant genera tions and increased the odds of contraceptive use in the second generation as compared to the third generation. Although there are no variables in the Add Health data set that may explain why grandparents and older siblings may be protective for particular reproductive health behaviors, I used data from the qualitative interviews to help explain the quantitative relationships. I extended prior literature and segmented assimilation theory by exploring both quantitatively and qualitatively how grandparents an d older siblings may influence reproductive health outcomes in adolescence. Grandparents Prevalence of a grandparent in the home. Table 7.1 above shows that the proportion of adolescents in the Add Health sample who had a grandparent in the home at Wave I ranged from 1.3% in the first generation to 6.6% in the second generation. While no qualitative participants lived with a grandparent during the adolescence, four first generation and three second generation participants lived with a grandparent at some p oint in their lives and almost all participants discussed their relationships with grandparents. Prior research found that when a father or both parents migrate without their children, children are cared for by extended female family members, often grandmo thers. 219,220 In my qualitative sample, two first generation young women, Eva and Yesenia, lived with their grandm others when their mothers migrated to Mexico. One

PAGE 296

280 second generation young woman, Jocelyn, was left with her grandmother in Mexico after later her mother migrated back t o the U.S. with her brothers. Additionally, Juana and Catalina reported their grandparents helped their single mothers in various ways when Research has shown that whi le parental migration causes family turmoil and difficulty, particularly with migration of a mother or both parents, supportive and loving caregivers can often ease the difficulty and children are not as bad off as some scholars and the media portray when blaming migrant parents. 220 While there is a difference in the type of relationship between grandparents and grandchildren who live together with a biological parent in the home, live together without a biological parent in the home, and do not live together at all, I focused this analysis on how grandparents may influence reproductive health behaviors in general, not on specific types of grandparent child relat ionships. Grandparents and reproductive health behaviors. Segmented assimilation theory does not specifically address the influence of grandparents on second generation adolescent outcomes. However, as discussed in Chapter 3, I hypothesized that due to rei nforcement of traditional values, grandparents in the home may be associated with a decreased risk for sexual initiation, an increased risk for non contraceptive use, and a decreased risk in unmarried adolescent births. Reviewing the results from Chapter 3 I partially supported my hypothesis. Table 7.8 shows the quantitative models that included grandparents and each reproductive health outcome. Consistent with my hypothesis, Model 1 shows that having a grandparent in the home was protective and reduced th e

PAGE 297

281 hazard for sexual initiation by 31% (1 0.69) (p<0.10). Contrary to my hypothesis, Model 2 shows that adolescents with a grandparent in the home were twice as likely to use contraception at their first sex compared to those without a grandparent in the ho me (p<0.10). Not supporting or rejecting my hypothesis, Model 3 shows the hazard ratio for a grandparent in the home is greater than one (1.16), meaning a grandparent in the home increased the hazard of an adolescent birth, though not statistically signifi cant at p=0.670. Unlike most quantitative papers that can only speculate on the reasons behind quantitative relationships, my mixed methods study has allowed me to use qualitative data to help understand these relationships. Table 7 8 : The Presence of an Grandparent in the Home and R eproductive Health O utcomes Model 1 Model 2 Model 3 HR of age at first sex OR of contraceptive use HR of an adolescent birth First generation 0.67** 0.42* 0.62 (0.51 0.88 ) (0.19 0.92) (0.35 1.10) Second generation 0.80* 0.64 0.59* (0.66 0.96) (0.40 1.01) (0.39 0.90) Third generation Ref. Ref. Ref. Two bio parents 0.62*** 1.88*** 0.54** (0.53 0.72) (1.40 2.51) (0.37 0.79) Grandparent in the hom e 2.03 1.16 (0.45 1.04) (0.99 4.17) (0.59 2.28) Constant 0.00*** 1.40 0.00*** (0.00 0.00) (0.15 13.12) (0.00 0.00) Observations 10,562 1,425 144,819 Through my qualitative data, I found four main ways in which grandparents may play a role in reproductive health decision making: (1) direct reproductive health advice; (2) providing a link to cultural roots; (3) giving nature of grandparents, providing financial support, gifts, or time and supervision; and (4) providing emotional support and love.

PAGE 298

282 Direct advice. Only one participant described direct advice regarding reproductive health from a grandparent. I asked Inez, who was raised by her grandmother in Mexico, if her grandmother talked to her about dating. She replied, Si, pero ella nada ms me deca que no. Las palabras de ella eran de no salir embarazada porque eso nos iba truncar la vida. Y que no nos dejara que nos pegara (laughs). Porque dice que all todos pegaban, ella tambin tuvo malas experiencias con sus esposos. Entonces nunca ms aparte yo creo que nunca nada de cudate solo nos deca que el novio se respetaba que el novio. Yes, but she would just tell me no. The words she would use were to not get pregnant because that would ruin my life and to not let anyone hit us (laughs). She would say that everyone would get hit, she also had bad experiences with her husbands. She would just tell us to take care of yourself and that you had to be respectful in a relationship with your bo yfriend. Later in the interview I asked Inez what her grandmother wanted for her future. She said, Ella nada mas siempre me deca que lograra tener una casita donde iba estar. com o pidindole a mi mama o andara en la calle sola. She would only tell me to have a house where I could be. She also told me that she wanted me to get married, be independent [give the orders in her home], not ask for help from my mother or be in the stree t alone. While an ambiguous message of being independent and yet responsible for the actions of from early pregnancy, domestic violence, and dependence on men. Inez m igrated independently at age 15 and quickly became pregnant and dependent on a violent partner. Multiple times during the interview she discussed how difficult it was to have left her igrate. Thus, her reproductive health outcomes may be more directly related to the loss of a supportive and loving relationship with her grandmother, in addition to the harsh life of a young migrant, than ambiguous gendered messages.

PAGE 299

283 Grandparents and cult ural roots. ambiguous, it did contain elements of Mexican traditions and female respectability. Drawing on segmented assimilation theory that asserts immigrant parents who maintain and encourage cultural traditions of their home country promote upward mobility in their second generation adolescents, I had initially hypothesized that having a grandparent in the home would be related to an increase in traditional Mexican/Catholic values and result in a lower hazard of sexual initiation, but higher odds of non contraceptive use and higher odds of a married adolescent birth. In support of my hypothesis, one study examining grandparents and reproductive health, found that adolescents who reported receiving advice about se x from grandparents reported attitudes that would likely delay sexual initiation, although they did examine sexual activity as an outcome. 105 However, The quantitative results shown in Table 7.8, only partially supported my hypothesis; a grandparent in the home reduced the hazard of sexual initiation but also incr eased the odds of contraceptive use at first sex. Although the quantitative models only partially support my hypothesis, five qualitative participants did in fact associate their grandparents to traditional or religious values and their cultural roots. Fi rst generation Eva, whose grandparents helped with her and her siblings while her mother was working and raised her for a few years while her mother was in the U.S., experienced family turmoil and disruption and very traditional gendered ideologies from he r mother and father yet graduated from college, delayed marriage until 25 and does not yet have children. She said, development and so growing up with my grandma I think really helped a lot of

PAGE 300

284 differen t, just cultural things. In terms of really traditionally going back my grandparents were more of that. I feel like that kind of helped me with my upbringing. She later attributed how she met and started dating her future husband at 15 to the spirituality believing in signs, energy and the universe, that she learned from her her life. Similarly, Yesenia, who was also raised by her grandmother when her mother migrated to the U.S. associated religion and respect with her grandmother and looked to up Yesenia shows that adolescents may think about what their grandparents would want t hem to do and how they would react, perhaps an example of respeto often reported in the literature. 107,109 Many women listed respect as an important value they learned from their families, particularly in first generation families. Fear of disappointing a grandparent may be a reason for adolescents with a grandparent in the home or a close relationship with a grandmother to delay sex or use contraception to prevent a pregnancy, both preventing a pregnancy and possibly keeping a grandparent from knowing they are having sex. As discuss ed in Chapter 4, a Mexican American ethnicity and maintaining cultural traditions and identification while adapting to the U.S. was associated with higher educational achievement and a delay in childbearing. In addition to Eva and Yesinia, four other women associated their grandparents with cultural practices; Mariela

PAGE 301

285 described how watching her grandmother kill a chicken on their rancho in Mexico made machismo in her parents but perhaps her grandparen rough childhood and was beaten with long sticks by her parents in Mexico, and Jessica Similar to lin king grandparents with cultural roots, even second generation participants who did not have a relationship with their grandparents due to death or distance in Mexico, knew something about how their grandparents lived and struggled in Mexico. Four women des cribed that their mothers or fathers had migrated independently during their adolescence specifically to financially support their parents and siblings in migrated at 1 7 to support his father and 16 siblings after his mother died giving birth to twins. Knowing their parents supported their grandparents may reflect familismo also described in Mexican families in the literature. 107,109 Although none of the women used that word, many of the women, spe cifically the second generation women, described that they learned to value family from their parents. Prior literature also suggests that in for female adolescents, achieving higher education is one way to help their families. 236 Thus, which may be related to delay in sexual initiation to preserve family honor as discussed in Chapter 6 or delay childbearing (through delaying sex or using contraception) to achieve higher education. 142,195 Further, some of these stories were linked to the messages of sacrifice heard from their parents in Chapter 6, thus knowledge of cultural roots and of

PAGE 302

286 how difficult life was in Mexico for their p arents and grandparents may result in adolescents appreciating and taking advantage of educational opportunities in the U.S., not wanting to dishonor their parents sacrifice and hardships of their grandparents. Giving of grandparents. In addition to the cultural roots and guidance grandparents provided, grandparents both living in the household and near the household, often provided financial support, gifts, time and supervision to their children and grandchildren. Prior literature has found that grandpar ents in the home may provide financial resources, if the grandparent is working, and added supervision which can improve child outcomes, especially in single parent families. 103 In Mexico, Juana reported that her grandfather, who struggled financially himself, helped support her mother and siblings while her fat grandparents walked her siblings and her to and from school and gave them chores when they came home from school while their grandmother lived with her and her brothers so her mother could care for her dying father baby sister so her mother could work. This addition of supervision, which has been found to decrease sexual risk behaviors, 77,82 may be another reason my quantitative analyses found that having a grandparent in the home reduced the hazard for sexual initiation and increased the odds of contraceptive use. Young women reported that both in Mexico and in the U.S. they often saw their grandparents and extended families on weekends, often Sunday dinner and holidays. All the women reflected nostalgically on these gatherings; grandparents brought food and

PAGE 303

287 gifts. First generation Beatriz actually described her gran dparents when I asked if there was anything else important to know about her childhood, B: Siempre estbamos cuando vena mis abuelos siempre eran Sbados y Domingos jugando lotera con las cartas y todos los primos jugaron era lo ms bonito hasta que mur ieron en el 2003. Ya cada quien cada hermano se apart y fue diferente ya hasta que nos venimos para ac. K: Y sus memorias de su niez ms o menos fueron buenos? B: Si eran buenas porque me acuerdo de all que bamos con mi abuelito cuando cumpla anos s iempre me hacan una rosca en Enero bamos all con mi abuelita o ellos venan. Vivamos cerca del rancho subamos con mi mama y all estbamos. B: When my grandparents would visit weekends we would play games, lotera (like bingo), and all the cousins pla yed, it was the nicest thing up until when they passed away in 2003. The family sort of separated after my grandparents died and everything changed until we moved away. K: So do you think your childhood memories were good? B: Yes. Because I remember my gr andparents coming to my birthday parties they would always make food, rosca (bread made during catholic winter holiday, family makes and invites others) for us. We lived close to them so we would always go and visit. Similarly, first generation Mercedes recalled how her grandparents would take them toys and money when they visited and second generation Jessica fondly described how her grandmother in Mexico would crochet for her and her sister. While gifts may have been an effort to help financially with the family, it was also a symbol of love. Love and support of grandparents. Emotional support from parents has been associated with a decrease in risky sexual behaviors 190,77,82,247,248 and grandparents may provide additional emotional support or buffer a lack of support in sing le parent families. While the women described relationships with grandparents differently, some closer than others, only two women described a grandparent negatively. First generation Beatriz, whose single mother passed away when she was a young child, des cribed her

PAGE 304

288 reported a very close and loving relationship with her maternal gr andmother. Seven participants did not know their grandparents or did not provide detail about their relationship with their grandparents. With the exception of Beatriz, the other 13 women described positive relationships with grandparents, many specificall y describing loving and supportive relationships. First generation participants more often described loving relationships with grandparents, probably because they grew up near them in Mexico and grandparents of the second generation participants were away from them in Mexico. First generation Yesenia and her siblings were initially left in the care of her aunt when her mother migrated to the U.S. Yesenia recalled that her grandmother had rescued them when she learned their aunt was neglecting them. My gran dma was the best grandma. It was such a change because we were so used to being with my mom and dad and then she [mom] came to the United States and moved from one place to the other. She left us to an auntie that we had to help her eating and even taking doing God knows what because they were tired of not eating an d not having food and stuff like that so they left school and went to live with their friends and God knows what they were doing. Me and my little brother had to stay there and my grandma saw that we were dirty. I think I had lice. My auntie was talking to her guys a big house and not having all the essentials for somebody to grow to going to my had a shower, we had a b oiler, we had food. She would take us lunch to school. them and removed them from a neglectful li ving situation with their aunt. Her women who did not live with their grandparents or lived across the border from each

PAGE 305

289 hips. Second generation Jessica, who visited her grandmother in Mexico on holidays, cried softly during the interview while explaining that her grandmother had recently passed away and Jessica was not able to attend the funeral because she was visiting her My grandmother just passed away in August and she was completely different. She would like crochet stuff for us and every time we would go Mexico she would be happy to see us and would always ask my mom for pict ures of us, so yeah (crying). My grandma was definitely more involved in everything of all my grandparents, so yeah. She would always ask about us a lot because every time we would go [to Mexico] we never caused any problems so we were her favorites, we we re always helping my mom. She saw how happy my mom was happy with us was really, really poor but somehow she always managed to give us something for Christmas, like little thing sorry. Second generation Jocelyn did not describe a loving or supportive relationship with her mother and stepfather perhaps they were often working and were in an unhealthy relationship. However, she liv ed with her grandparents during the summers and her grandfather lived in the U.S. with them when her youngest sister was a baby and she describe her parents. In this sit uation, grandparents provided a close relationship that she did not experience with her parents. Limitations. Quantitatively, I was only able to measure the presence of a grandparent in the home at Wave I, not if the grandparent had previously lived with the participant which may also influence adolescent behaviors as my qualitative data showed. I asked all qualitative participants to describe their extended families and grandparents, although I did not probe further if participants did not give detail abo ut grandparents

PAGE 306

290 because it was not the focus of my study. Future research qualitative research that focuses on grandparents may uncover more detail than I had the ability to describe. It is interesting that I did not observe a significant quantitative rel ationship between grandparents and risk for adolescent birth, especially when grandparents were associated with a delay in sexual initiation and an increase in contraceptive use, both of which are associated with a reduction in adolescent births. Perhaps t he small proportion with a grandparent in the home at Wave I coupled with the small proportion of adolescent births did not have the power to detect a significant difference. Further, having a grandparent in the home does not measure the quality of the rel ationship with the grandparents or timing of grandparents in the home. Perhaps questions related to the quality of the relationship with a grandparent, as my qualitative results have shown, could elucidate a significant association between a supportive rel ationship with a grandparent and a reduction in risk of an adolescent birth. Conclusion. In sum, grandparents provided the young women in my study with a link to their cultur al roots which was associated with the importance of family and sacrifice; financ ial resources and supervision; and love and emotional support, all which could delay sexual initiation, increase contraceptive use, and decrease risk of adolescent childbearing. It appears that although grandparents were often associated with cultural root s and religion qualitatively, quantitatively grandparents were associated with increased odds in contraceptive at first sex, perhaps because reproductive health was not often discussed directly with grandparents and thus the financial resources, supervisio n, love and support outweighed any traditional anti contraception values grandparents may or may not have had. Grandparents may also put the love they have for their

PAGE 307

291 grandchildren and their desires to protect their grandchildren ahead of cultural tradition s who told her not to get pregnant, the young women used contraception to prevent a pregnancy that would disappoint their grandparents. My results demonstrate the nee d to include grandparents in segmented assimilation theory for two reasons. First, they represent an important and protective link to cultural and country of origin, as the theory asserts more globally for families who reinforce traditional values. Second, similar to the positive relationship between family SES and reproductive health outcomes that demonstrated an more complex relationship than segmented assimilation theory can explain as discussed in Chapter 3, the traditional values that grandparents prov ide are woven together with supervision, financial resources, and emotional support. Future research could specifically focus on different dimensions of grandparents in Mexican origin families. Interventions could include outreach to grandparents, particul arly in single parent families. Health care providers could inquire about grandparents and actively involve grandparents in patient education. Older siblings Although most research has shown that the presence of an older sibling is associated with risky sexual behaviors in younger siblings, this relationship is moderated by genetics, gender, sibling relationships and other family factors. 110,111,113,200 In fact, prior research found that younger sisters who reported spending more time with their older sisters who were adolescent mothers, were more likely to experience an adolescent birth themselves than adolescents who reported more conflict with their pare nting older sisters. Companionship with an older sister increased the odds of an adolescent birth

PAGE 308

292 regardless of whether the sister was an adolescent mother or not suggesting older sisters are not protective against early childbearing. 38 Similarly, high levels of modeling, when younger siblings reported they wanted to be like their older siblings, and a higher sibling relationship quality was associated with a higher correlation in younger and old er siblings deviant behavior. 110,200 During qualitative data collection, I noticed that women described close relationships with their older siblings and often reported they learned about reproductive health through discussions with their older siblings or watching their experiences, thus, I decided to add models on the presence of an older sibling in the home into the quantitative models. Prevalence of an older sibling in the home. Table 7.1 shows that over 60% of first and second generation Add Health participants had an older sibling in the home at Wave I and 100% of first generation qualitative participants and 80% of second generation qualitative participants had an older sibling. The two qualitative participants who did not have an older sibling, were older siblings themselves, none we re only children. Older siblings and reproductive health outcomes. Based on my qualitative results, I hypothesized that older siblings would be protective for all three reproductive health behaviors. Table 7.9 shows the influence of having an older sibli ng in the home on reproductive health behaviors. Consistent with my hypothesis, Model 1 shows that having an older sibling in the home reduced the hazard of sexual initiation by 15% (1 0.85) (p<0.10). Model 2 shows there was no significant interaction betw een having an older sibling and immigrant generation. Model 3 shows the relationship between having an older sibling in the home and the odds of contraceptive use were not significantly

PAGE 309

293 different than one. However, also consistent with my hypothesis, Model 4 shows the odds of contraceptive use in second generation adolescents with an older sibling in the home was 1.6 times higher than third generation adolescents with an older sibling in the home (2.27*0.71) (p<0.10). Models 5 and 6 show there was no signif icant main effect nor interaction for the presence of an older sibling in the home. However, the hazard ratio of an adolescent birth for second generation adolescents with an older sibling in the home is about 31% less (0.63*1.09) than third generation ado lescents with an older sibling at home and approached significance p<0.195. Table 7 9 : The Presence of an Older S ibling in the Home and R eproductive Health O utcomes Model 1 Model 2 Model 3 Model 4 Model 5 M odel 6 HR for age at first sex HR for age at first sex OR of contraceptive use OR of contraceptive use HR for an adolescent birth HR for an adolescent birth First generation 0.69** 0.78 0.43* 0.59 0.71 (0.54 0.88) (0.55 1.13) (0.20 0.93 ) (0.16 1.02) (0.32 1.09) (0.29 1.74) Second generation 0.71*** 0.62*** 0.75 0.46* 0.51** 0.67 (0.60 0.85) (0.48 0.79) (0.49 1.16) (0.21 1.00) (0.34 0.76) (0.40 1.12) Third generation Ref. Ref. Ref. Ref. Ref. Ref. Older sibli ng 0.82 0.98 0.71 0.91 1.09 (0.73 1.02) (0.61 1.09) (0.68 1.39) (0.44 1.15) (0.69 1.20) (0.71 1.68) First gen X older sibling 0.85 1.16 0.72 (0.54 1.33) (0.49 2.78) (0.32 1.62) Second gen X older sibling 1.28 0. 63 (0.88 1.86) (0.94 5.46) (0.31 1.27) Third gen X older sibling Ref. Ref. Ref. Constant 0.00*** 0.00*** 1.81 2.03 8.11*** 8.09*** (0.00 0.00) (0.00 0.00) (0.19 17.31) (0.21 20.16) (4.70 13.99) (4.68 13.99) Observatio ns 10,562 10,562 1,425 1,425 144,819 144,819 *** p<0.001, ** p<0.01, p<0.05, p<0.10

PAGE 310

294 These findings are particularly interesting because they challenge prior research has found that having an older sibling generally increases risky behaviors, includi ng sexual initiation and substance use. 38,111,112,257 I argue that the reason my results are different than those in multiracial studies are because the relationship between siblings works differently in Mexican origin families, particularl y in migrant families. Through my qualitative data, I found three ways in which older siblings helped protect their contraceptive use, and adolescent childbearing; (1) older siblings, especially sisters, were expected to be responsible for their younger siblings; (2) siblings relationships were often described as close and supportive; (3) some older siblings provided direct advice on reproductive health and academic guidance; and (4) younger siblings learned from watch Older sibling responsibility. Consistent with prior research 108,144,145,220 and as discussed in detail in Chapter 5, I found that older siblings, particularly older sisters, had more domestic responsibility in the home than younger siblings and brothers, which oft en included the responsibility to care for younger siblings. Although only two young women were actually the oldest, those who were the oldest girl, or even the second oldest girl, described how they were expected to help with their younger siblings. First generation Mercedes helped her younger siblings with their homework and combing their hair, Eva sisters, and Catalina helped her older sister and her single mother wit h her younger brothers. Second generation Jocelyn and Alicia both cared for their younger siblings from changing diapers to walking them to school and caring for them when their mothers were

PAGE 311

295 at work. Eva is currently unsure if she wants to have children, d espite the familial and social pressure she feels, because she feels like she already raised her siblings, them to their first day or school and I was there during their recitals and this and t three night if I want to. At 30 now! (laughs) Women who were among the younger in their fa milies, described how their older sisters helped raise them. First generation Juana believed she was mostly raised by her older member who helped her with her homework. Car la was eight years younger than her older brothers who were told by her parents they had a responsibility to look after her, such as picking her up from school and even disciplining her. While there has been a lack of research on how sibling relationships specifically sibling responsibility, influences reproductive health behaviors, East and Khoo 113 found that non report on the frequency her older sister tells her what to do, was related to an increase in risky sexual behaviors. However, older sister power was sign ificantly and positively correlated with older sister warmth which was associated with a decrease in risky sexual behaviors and older sister power was on the pathway between families receipt of welfare and risky sexual behaviors suggesting older sisters ha These findings do not contradict my findings, because I am not convinced that the qualitative participants who were partially raised by their siblings would have felt their

PAGE 312

296 extremely gendered household, described in detail in Chapter 5, where her older brothers could tell her what to do and punish her. Further, quality of sibling relationships perhaps moderated the effects of sibling responsibil ity; younger siblings who perceive their older siblings responsibility of them through a caring supportive relationship will likely experience different outcomes than those who felt their older siblings were controlling In sum, exp eriences from older and younger siblings support the argument for older siblings, in particular girls, having responsibility and feeling a responsibility to care for and raise their younger siblings. This means that younger siblings have an additional care giver to their parents or, as in the case of grandparents, could provide care and support when their parents were absent, physically or emotionally. This additional support may be a reason that my quantitative analyses found older siblings to be protective against early sex, non contraceptive use and early birth. Quality of sibling relationships. Qualitative participants described varying levels of detail about relationships with their siblings during their childhood and adolescence. Of the 18 women who de scribed enough detail for me to gauge whether or not they had a positive or close relationship with siblings or a negative relationship with siblings, 11 participants specifically described supportive or close relationships with at least one sibling and on ly five participants described generally unsupportive or negative relationships with at least one sibling. Relatively equal numbers of first and second generation participants described positive and supportive relationships (6/10 and 5/8, respectively). Of int erest, only two participants out of nine who experienced an adolescent birth described supportive relationships as compared to nine of the 12

PAGE 313

297 participants who delayed birth. This suggests that close sibling relationships may protect against early child bearing. Close relationships were often described as playing together as young children and spending time together and having someone to talk to as they got older; some young women reported they talked more with their siblings than parents. For example, f irst generation and college graduate Isabel, recalled that her siblings rode bikes, played chase, I do. I want four and the reason I want four is because of my childhood. Because I loved growing up with my siblings and I love now growing up and having I can call my brother. I [boyfriend] only wants one. He wants one, I think our compromise if anything As Isabel described, the majority of participants recalled that during their young childhoods, played and fought with their siblings, but became closer as they became older, particularly with sisters. In support of some prior research that found that sibling relationships may be closer in single parent families, 113 I found that siblings were sometimes close in single parent families because older siblings had responsibility for their younger siblings, but also because siblings, especially older siblings, could buffer the turmoil and lack of attention often fo und in single parent families. For example, first generation Eva, described her relationship with her older brother. We experienced similar things and he was there for me and I was there for him, hap No te

PAGE 314

298 preocupes negritas That was always my big que stion because he was always there for me and my ready for her and set the table and she would go take a sho wer and lay in bed. would have to get up early the next morning. And then my brother would do my hair. Those are the best. To this day I still see pictures of me, school pictures, a nd it was just horrible. Those were not very pleasant pictures. (laughs) Eva was the oldest girl so she had primary responsibility for raising her three younger sisters, howev surviving the difficulties of being a new immigrant and having an overworked single mother. The negatively described relationships with siblings were usually explained by emotional physical, or age distance. First generation Inez grew up with her grandmother while her siblings were raised by her mother and although they were about three miles st refuse take her places although now as an adult, Beatriz described a close relationship with her sister. A few women who were treated differently than their brother s or older siblings, described more frustrating relationships, such as Juana. I asked what it was like to be the youngest of eight. ur older ones? I mean I would have oldest and you respect them. So my older sister was more like my mother. She hard because I would have respect that decision because they are the older sister or brother.

PAGE 315

299 Ju ana also described famlismo and respeto which in this case, did not translate into closer sibling relationships as it did in some families. Advice from older siblings. In addition to the quality of the sibling relationship, some women described learning about sex and receiving advice from their older siblings that they believed protected them. Four women recalled they learned about sex and contraception from their older sisters and two other women were warned against having sex and the dangers of predato ry men from their older brothers. When asked if she used contraception at her first sex, second generation Arely, who planned her first pregnancy started talking to m few women discussed that they talked to their younger siblings, or younger cousins, sisters in the re ligious sex education class that Eva had to sit through because they had contraception and delaying sex. Some participants also received support and guidance from their old er siblings regarding their education. In addition to older siblings helping with homework, some older siblings actively pushed their younger siblings to succeed academically. First generation Guadalupe, whose mother passed away prior to her high school gr aduation, credits her older siblings encouragement for her academic success. Second generation

PAGE 316

300 my results have shown, high goals for education and receiving parental, and possibly sibling, involvement, may decrease risk for early sexual initiati on, non contraceptive use and delay childbearing in order to achieve academic goals. Birth order may be of particular importance in migrant families. Five second generation women, four of whom delayed their first sex and birth, recalled how their older s iblings brought English home from school, making it easier for them once they started school. Isabel and Ana believed their older sisters had more difficult experiences in school than they did because they were the first in the family to go to school. Isab el explained, I think my oldest sister had a hard time. I think my experience is different than hers, because she was the first one to go to elementary school. She was the first one to really learn English and she brought English home and I was better prep ared when I went to school. I think she was the one that dealt with growing up in a really Mexican household and going to her elementary classes where no one really spoke Spanish. Even though my sister had blonde hair and she might have looked white, she w than we did. The older siblings had a language barrier and experienced more discrimination from the other children than the younger siblings did. By the second child, parents also have a better understanding of the U.S. school system and the younger siblings had a better grasp on English and what school would be like for them. It is plausible that younger siblings have an easier transition to school that may ultimately lead to better academic success and less risky behaviors because their older siblings paved the way for them. Learning by exp erience. Some young women, even those who did not report particularly supportive sibling relationships and did not receive direct advice from their

PAGE 317

301 reproductive health decisio ns based on what they learned. Second generation Ana and Jessica, both college graduates, reported they learned to be sneakier after watching their older siblings get caught dating or with contraception. While this type of learning may not necessarily be a dvantageous, Jessica learned to hide her birth control pills in her dating parents. Four women reported they had an older sibling with an unplanned pregnancy or childbirth. Two went on to be adolescent mothers themselves, and the two who did not, cred ited their successful delay of pregnancy with learning from their siblings negative experiences with an adolescent pregnancy. The women I interviewed generally adapted their behavior to avoid an early pregnancy after watching their older siblings experienc es, however, as prior research has shown, modeling can also be an influence that increases risky behaviors if the older sibling is engaging in risky behaviors. Despite prior research suggesting the presence of an older sibling increases risky behaviors, t he difficulty of consistently measuring sibling relationship characteristics has led to somewhat mixed results on sibling relationships and behaviors and few studies have qualitatively studied sibling relationships and behaviors in Mexican origin immigrant families. In addition to helping their parents care for and raise younger siblings, the women in my study often described supportive and loving relationships with their siblings, received advice and guidance from older siblings and learned how to adapt th eir reproductive health behaviors after watching their older siblings experiences. In support of my findings, Updegraff and colleagues 107 found little evidence of negative an d

PAGE 318

302 conflict oriented sibling relationships in their study of Mexican American siblings, East and Koo found higher sibling warmth was associated with a decrease in risky behaviors in non Latino black and Latino siblings, 113 and Killoren and Roach found that when sisters discussed sexual and reproductive health they were sources of support, co nfidants, and older sisters mentored younger sisters. 258 Further, Siblings in Mexican origin families spend more time together than non Latino white siblings 107,200 which could symbolize closer relationships between siblings which has been fou nd to have lower sexual risk behavior. 113 Older siblings may spend more time with, be closer to, and feel more responsible for their younger siblings, resulting in more guidance, sexual education, and academic support for younger siblings. My quantitative results showed that having an older sibling in the home increased the odds of contrace ptive use at first sex for the second generation as compared to the first and third generation. Perhaps third generation adolescents are more similar to non Latino white and black adolescents and with less sibling responsibility, time spent together, or cl oseness, older siblings may provide more of a risk, as prior literature has found. 38,111,112,257 The reason behind the difference between the first and second generation is less clear. I did not find any substantial qualitative differences in sibling relationships between first and second generation adolescents. However, because first generation adolescents are less likely to use contraception than second and third generation adolescents, perhaps they are less likely to provide direct advice regarding contraception to younger siblings than second generation adolescents. Future research on sibling relationships that includes differences by immigrant generation is needed.

PAGE 319

303 Limitations. Using a qualitative local sample to understand a national qu antitative sample means that differences between samples may account for differences in findings. However, I ensured that the qualitative participants could have theoretical participated in the Add Health survey in order to provide the ability to draw conc lusions between the samples. Although I asked all qualitative participants to describe their relationships with their siblings, I did not probe further if they did not give detail regarding siblings because sibling relationships were not the focus of my st udy. Prior research 113,200,257 and my results sh ow that sibling gender is an important factor in sibling relationships but I did not examine the difference between having an older brother or an older sister in the home in the quantitative models because it was again, not the focus of my study. Quantitat ively, having an older sibling was measured by having an older biological sibling. Considering stepsiblings or half siblings is an important next step in sibling research. Conclusion. My quantitative and qualitative results suggest that older siblings, pa rticularly sisters, in Mexican origin families may reduce the risk of early sexual initiation, non contraceptive use, and adolescent childbearing. Future research is also needed on the context of how older siblings influence reproductive health decisions, including more detailed information on relationship quality, direct advice, and how implications include, involving older siblings in sexual and reproductive health educati on, specifically providing older siblings with accurate sexual health information that they can then share with younger siblings. Family planning clinics may also seek to provide dyad care to siblings or specifically invite siblings to particular visits.

PAGE 320

304 CHAPTER VIII CONCLUSION esearch in the social, behavioral, and reproductive health sciences continues to interpret Latino families' beliefs and practices through culturally defined theoretical concepts such as machismo, marianismo hembrismo familismo personalismo and the madonna/whore dichotomy, among others, including a socalled "Latino culture." Besides the need to recognize not one but many "Latino cultures," an overemphasis on these categories of analys is may promote inaccurate images of Latinas and Latinos who live in the United States. Without compromising the importance of cultural forces, we need to explore alternative and comprehensive theoretical frameworks aimed at examining how and why socioecono mic structures shape parenting styles and fatherhood experiences, gender relations, and the sex education of a new generation of Latina and Latino children. Gloria Gonzalez Lopez 195(p1128) In this dissertation, as Garcia 145 nt cultural frameworks by examining the complex associations between migration patterns, family structure, family socioeconomic status, gendered roles and discourses, academic discourses, and reproductive health behaviors in Mexican origin young women. To How do cultural constructions of gender and sexuality help explain the relationship between immigrant generation, family SES, family structure and reproductive health behavior of Mexican origin adolescents/young ad I found that gendered discourses and roles do help explain the relationship between migration, family structure, family SES and reproductive health outcomes. However, I found that academic discourses and active parental academic involvement appear to be a stronger mediator of the relationship between immigrant generation, family structure and SES and reproductive health outcomes in Mexican origin adolescents. More importantly, I have begun to unravel how socioeconomic factors are interwoven with th e migration process, family structure, gendered discourses, parenting and ultimately reproductive health decision making in Mexican origin young women.

PAGE 321

305 Figure 8.1 shows a conceptual framework of my results. I conclude this dissertation by summarizing my co nclusions for each of the factors contributing to reproductive health behaviors and describing implications for future work and current public health and clinical practice. I close with discussing the utility of segmented assimilation theory to understand reproductive health behaviors, the measurement of acculturation, and I offer Figure 8 1 : Conceptual Model 17 M igr ation and Family Migration Advancing prior literature, I found that independent migration, when a single adult migrates to the U.S. alone or with children, was often associated with long term family instability and single parenting, especially for female migrants. The socioeconomic status of families in Mexico influenced migration patterns; the majority of participants with a parent who migrated independently believed it was to support their poor parents and siblings in Mexico. The relationship between ind ependent migration and 17 I acknowledge that the gende red and economic antecedents of migration, including the region of origin in Mexico (urban vs rural), directly influence gendered discourses and roles, I did not measure them in this study and thus I did not draw a line connecting migration and gendered di scourses and roles in the figure. Migration Family structure & SES Academic discourses and parent action Reprod uctive health outcomes Gendered discourses and GDL, GDP

PAGE 322

306 subsequent unhealthy common law partnerships occurred more often in women, who are often more economically dependent upon men, especially if they have children. Consistent with prior research, 221,244 I found that migration caused family disruption that almost always separated children from at least one parent. Second generation participants who were born after their parents reunified in the U.S. were more likely to delay sex and their first birth than those who were separated from one or both parents. Family instability in migrant families should not be seen as a loss of the value of family or adaptation to a society with a high divorce rate but should rat her be seen as consequences of poverty and migration. Segmented assimilation theory considers the influence of communities where migrants settle and sociopolitical climate of the U.S. but it does not recognize the how the family migration process is relat ed to family structure and family socioeconomic status, key predictors of adolescent outcomes in the model. I found that the migration patterns of parents of the second generation were associated with family structure and family SES and thus migration patt erns are missing elements in segmented assimilation theory. Family SES Socioeconomics play an important role in both the migration and settlement processes. All of the qualitative participants grew up with undereducated parents (mean parental education <1 2 years) in poor or working class families. Participants who low wage unskilled jobs in order to feed their children and spent less time responding to their emotional needs and providing supervision, did not feel supported by their parents

PAGE 323

307 and more often experienced early sexual initiation, non contraceptive use, and an adolescent birth. Quantitatively, I advanced prior literature by finding that parental education an d parental employment were not protective for all reproductive health outcomes in all sexual initiation increased in the first and second generation. I speculated th at higher educated migrant parents would be more liberal in their sexual views and may not emphasize virginity or have as strict dating rules as lower educated parents. I did not have enough variation in parental education in my qualitative sample to addre ss this finding and I did not interview third generation women. My quantitative analyses found that parental expectations about sex and contraception, parental communication about sex, gendered expectations and religiosity did not mediate the interaction between parental education and the hazard of sexual initiation in the first and second generation This does not mean that higher educated parents do not have more liberal views of sex but rather the Add Health variables may not have captured the concepts well. In fact, Gonzalez Lopez 195 has argued that migrants from urban Mexico are more highly educated than rural migrants and hold more liberal sexual in itiation. Add Health is a nationally representative data set but does not address area of origin in Mexico for Mexican participants and thus regional differences in Mexico cannot be controlled for quantitatively. Further, the quantitative analyses found th e largest differences in hazard and odds ratios between the first and third or second and

PAGE 324

308 third generations. Unfortunately, I did not interview third generation women and cannot address why parental education and employment worked differently in the third generation than the first and second using qualitative data. Although the interaction between parental education and immigrant generation most consistent predictor in al l three reproductive health outcome analyses. In the first generation, having a mother not working reduced the hazard of sexual initiation, increased the odds of contraceptive use and reduced the hazard of an adolescent birth compared to participants with professional mothers. Having a mother not working increased the odds of contraceptive use among all three generations. Mothers who are not working outside the home spend more time on childcare than mothers working outside the home. Married mothers are more likely to stay at home because they want to care for their families and migrant stay at home mothers are more likely to be married. 141 Migrant mothers may be more likely to choose to stay at home while native born mothers may be at home due to disability or difficulty finding a job thu s a not working mother may not be as protective in the third generation because mothers are not choosing to stay at home and may not provide the support that migrant mothers provide. My qualitative results supported the quantitative relationship because m ore women with stay at home mothers than working mothers delayed sex, used contraception and delayed their first birth. Stay at home mothers may provide more supervision and support than working mothers but the underlying family structure and socioeconomic s that allows mothers to choose to stay home should not be underestimated because qualitative data showed that participants with stay at home mothers had two biological

PAGE 325

309 parents in the home and a working father. I did not interview third generation women so I cannot explain why there is a quantitative difference in mothers not working between the t working in the first generation and the hazard of sexual initiation or adolescent birth. Again, this does not mean that stay at home mothers do not provide more support or supervision, but rather the variables may not have captured what I intended for th em to capture or there are other reasons, such as more conservative sexual views, that I did not test. Family structure Parents. My quantitative models and qualitative results found that the importance of two biological parents in the home cannot be unde restimated. Having two biological parents in the home reduced the hazard of sexual initiation, increased contraceptive use and reduced the hazard of an adolescent birth among all generations in the quantitative models and among the first and second generat ion qualitative participants. Having two biological parents also accounted for some of the difference between the second and third generation in all three outcomes, mediating immigrant generation, sexual initiation and adolescent birth and suppressing the odds of contraception use. Two parents can provide more economic resources, supervision and support to children. Qualitative participants who described more supportive and positive relationships with their mothers and fathers also had better reproductive h quality and communication reduced the hazard for first sex and maternal communication reduced the hazard of an adolescent birth.

PAGE 326

310 Fathers were of particular importance in qualitative interviews and are often left out of reproductive health literature. Women who described loving and supportive relationships with their fathers more often delayed sex and first birth. A few of these high achieving women felt closer to their fathers and that their fathers pus hed them more to succeed than their mothers. After controlling for maternal characteristics and family income, having a father in the home significantly reduced the hazard of sexual initiation highlighting the importance of a father in the home regardless of family income and how supportive the mother is. This association was not significant for adolescent birth models, perhaps because the relationship with the father is more important than merely having a father in the home when it comes to childbearing. T hese findings are crucial to dismantling common stereotypes that fathers are possessive of their women and seek to Grandparents. To my knowledge, I am the first to explore the role of grandpar ents in the lives and reproductive health decisions of Mexican origin young women. Quantitatively, I found that having a grandparent in the home during adolescence reduced the hazard of sexual initiation and increased the odds of contraceptive use. Through qualitative interviews, I found that grandparents provided their granddaughters with a link to their culture roots that was associated with the importance of family and sacrifice; financial resources and supervision; and love and emotional support, all wh ich could delay sexual initiation, increase contraceptive use, and decrease risk of adolescent childbearing. Siblings. I have ch allenged prior research that found having an older sibling increases risky behavior in adolescents. 110,111,113,200 During initial qualitative data

PAGE 327

311 collection I noticed that relationships with older siblings were discussed positively and with importance during interviews. Examining t he role of older siblings in quantitative models, I found that having an older biological sibling was associated with a reduction in the hazard of sexual initiation among all immigrant generations and increased in the odds of contraceptive use in the secon d generation. Qualitative analyses found that older siblings especially sisters, were expected to be responsible for their younger siblings; sibling relationships were often described as close and supportive; some older siblings provided direct advice on r eproductive health and academic guidance; and younger siblings in Mexican origin families spend more time together than non Latino white siblings 107,200 which could symbolize closer relationshi ps between siblings which has been found to have lower sexual risk behavior. 113 Older siblings may spend more time with, be closer to, and feel more responsible for their younger siblings, resulting in more guidance, sexual education, and academic support for younger siblings. Older siblings may not be protective in third generation adolesc ents perhaps because they are more similar to non Latino white and black adolescents, who may have less sibling responsibility, time spent together, or more distant sibling relationships than Mexican origin siblings. This would help explain why prior resea rch in multiracial samples found older siblings increased risky behaviors in their younger siblings. 38,111,112,257 F amily, Gender and Academics Gendered discourses I hypothesized that gender and sexuality mediated the relationship between family structure and SES and reproductive health outcomes in Mexican origin young women.

PAGE 328

31 2 My results partially supported this hypothesis; they were directly related to family structure, family SES and reproductive health outcomes as hypothesized, but gende red roles and discourses were similar in the first and second generation. Based on my results, I advocate for a more nuanced view of the currently accepted cultural framework. First, gendered discourses were often at odds to what was actually occurring in the household. Mothers and fathers who encouraged a gendered division of labor and power in the household often did not follow their own advice. Second generation Teresa learned that women must be subservient to men and that marriage and motherhood were he r ultimate erence draw upon as they negotiated gendered expectations and made their decisions. Second, I found definitions of machismo and marianismo were also highly nuanced and were not useful terms to understand gendered identities of Mexican origin men and women particularly because they are removed from the socioeconomic context participants described h ow their parents did not want them moving out of the house until they were married although their brothers could move out alone, supporting the view that female independence is gained through marriage. However, this was not the case in all families and par ents did not appear to encourage early marriage nor discourage educational attainment even when they wanted their daughters to remain at home until married. Of the five women who reported they were expected to live at home until

PAGE 329

313 marriage, four who followed this advice completed college degrees. Concurring with other researchers, 137,195,224 I advocate moving away from using these labels that only contribute to stereotypes. When I asked qualitative participants to share the va lues they learned from their parents, three first generation and six second generation women mentioned the importance of family. Six first generation and two second generation women mentioned the importance of respecting others. W hile w omen were clear that machismo and marianismo were not a pplicable to their lives today, a bout half of them supported cultural values of familismo and resp e to Looking more closely at the data, all but one of the nine women who reported the importance of family came from a two parent household. This again, cautions us against taking even positive stereotypes at face value and reinforces the link between socioeconomics, culture, family and immigrant generation. Finally, contrary to common stereotypes, I found that the most extre mely gendered households, including a those with violent fathers or young women not being allowed out of the house without their brother or father, were related to earlier sexual initiation and unmarried and unplanned adolescent childbearing. Moderately ge ndered and egalitarian households, where fathers helped mothers in the home when they coul d and mothers shared decision making power were associated with delayed sexual initiation and delayed first birth. Households with a more extreme gendered division of labor, power and discourses were of lower socioeconomic status and experienced more family turmoil and instability. A negative association between egalitarian attitudes and poverty supports this important finding. 222 This information is critical to the discussion of how gender is constructed in families, how young women interpret it and how it

PAGE 330

314 influences their reproductive health decision making. These resul ts challenge current public discourse and prior research that uses culture and gender as the reason for the current Latina adolescent birth rate and other reproductive health disparities without considering the relationship between culture, gender, and soc ioeconomic status. These findings do not contradict quantitative findings of the positive relationship between parental education and the hazard of sexual initiation discussed in Chapter 3. Conservative sexual views are not the same as extremely gendered h ouseholds and qualitative data showed that women in extremely gendered households received less parental messages of virginity than moderately gendered households. This highlights the complexity of gender and sexuality. How parents educate their daughters Sex education. Contrary to public discourse and some prior research, Mexican varying levels of avoidance and parental discomfort regarding sexual topics, this is not l imited to Mexican origin parents, but parents of all race/ethnic groups. 148,226 Messages cuidate cuidate and a delay in childbearing. It is poss ible that it is improved parent child communication and parent al support associated with these messages of virginity, rather than the messages themselves, that are associated with improved reproductive health outcomes. I found that although most sexual messages and education came from mothers, fathers also encourage d delaying sexual intercourse, although contrary to stereotypes it

PAGE 331

315 was not necessarily to preserve virginity and family reputation. The women believed their fathers cared about their well beings and wanted them to avoid consequences such as adolescent preg nancy and protect them from harm. 195 Consistent with Garcia, 145 parents want their daughters to be able to protect themselves from an early pregnancy and sexually transmitted infections even if they are not able to articulate how to do this. In Chapter 3, I discussed how higher parental educat ion was associated with a higher hazard of sexual initiation in first and second generation adolescents and having a not working mother was associated with a reduced hazard of sexual initiation, adolescent birth and increased the odds of contraceptive use. Although the quantitative models in Chapter 7 found that parental expectations about sex did not mediate the relationship between parental education and sexual initiation, it is still plausible that conservative sexual values in undereducated, probably ru ral, migrant parents account for this finding. Further, in my qualitative sample second generation women were more likely to complain about and negotiate gendered expectations about dating than first generation women which could also account for some of th e difference in sexual initiation between the first and second generation. I did not interview third generation women which is unfortunate because they have native born parents and have the largest quantitative differences in reproductive health behaviors. Sexual experience. Parental messages conveying the vulnerability of women and pregnant without providing any sex education was not necessarily helpful and was relat ed to earlier sexual initiation, earlier adolescent births and more negative sexual experiences. The majority of women described uncomfortable and awkward sexual experiences,

PAGE 332

316 particularly if they had not received sexual education. Some women made conscious decisions to learn about sex and contraception and prepare for future sexual intercourse research, 138,145,259 women are still relucta nt to view themselves as sexual agents who may desire and enjoy sex. Further, women generally felt unprepared and rarely discussed having sex or preventing pregnancy or STIs prior to the event which increases risk for pregnancy and STIs. Whether or not we believe that adolescents should be having sex, they are, and the lack of preparation and negative first sexual experiences is not acceptable and needs to be addressed. Academic expectations and parental action. Although Mexican origin parents value educat ion and want their children to become educated and economically independent adults, there is room for improvement in Mexican involvement in the education of their children. Mexican origin migrants face a number of barriers that make it difficult to convey their expectations about school to their children their parents faced language barriers, struggled with the differences between the education syst em between Mexico and the U.S., had very low levels of education parenting and trying to put food on the table. Despite these barriers, the women all knew their pare nts wanted them to do well and complete school. Some parents, however, were able to promote educational success in their children through their messages of sacrifice and through active involvement in education.

PAGE 333

317 Parental messages of sacrifice to their chil dren has been documented in migrant parents whose children remain in their home country. 220,244 I am the first to document these messages in children in the U.S. and the first to document their relationship to both academic success and reproductive health outcomes. In fact, messages of sacrifice and parental action appear to be more powerful mediators of the relationship between family structure and SES and reproductive h ealth outcomes than gendered discourses were. While gendered discourses were actually similar in the first and second generation, second generation women more often reported messages of sacrifice and active parental involvement. Further, these messages wer e able to buffer family turmoil and extremely gendered discourses in some of the women. All five women whose parents told them that they migrated to the U.S. so their children could succeed, delayed their first sex, first birth and four completed college. Further, active parental involvement in education, such as checking grades and attendance or placing the importance of homework above chores, was also associated with delayed sexual initiation, first birth and higher educational attainment. Contrary to so me prior research and public discourse, but consistent with a previous study, 109 the women in my study believ ed their parents wanted them to finish high school, attend college or start a career, and then marry and have children. Many parents specifically told their adolescent daughters not to get pregnant and as discussed in isappointed she wanted to marry at 21. The majority of the women in my sample believed their parents migrated to give them a better life and knew their success in life was important to their parents. Some of the college graduates I interviewed felt that af ter they completed their education their parents placed

PAGE 334

318 a higher value on motherhood and childbearing. However, motherhood and education were not opposing goals; when they were growing up and actually attaining their education, their parents pushed them to excel. Further, desiring grandchildren or giving grandchildren more attention than a college degree is not limited to Mexican origin families. I have never seen an older adult wearing a t shirt listing the degrees of their children as they wear t shirts w ith the number of grandchildren they have. The paradox between parents migrating to give their children a better chance at economic success and their difficulty promoting higher education in their children is an Swidler, 237 educational success is not because they cling to cultural values but because they are reluctant to abandon familiar strategies of action for which they have the cult 237(p287) Migrant children to succeed in school but the value is an end, not the strategy of action. Many migrant parents may be unable to act on education system, knowledge of early home learning, and because they are unsure how higher education is achieved. This may be why families who were able to actively married or having children because that did not fit into their world view. Supporting this argument, I told my research assistant, a second generation Guatemalan who is a recent college graduate, how the women I interviewed believed

PAGE 335

319 232 U.S. white women who delay marriage and childbearing and obtain college degrees and professional employment have a very different life course than women who marry in their teens in rural Mexico and become stay at home mothers. It is possible that Mexican migrant parents have no idea what ch different than any life they have ever seen. Immigrant generation Table 8.1 shows the quantitative and qualitative findings for the three main reproductive health outcomes and other studied factors. My qualitative results generally supported my quantit ative findings, at least related to the first and second generations, and either advanced or extended prior literature. Table 8 1: Quantitative and Qualitative Findings and Relationship to Prior Literature Quantitative Qualitative Pri or literature Sexual initiation a significantly lower hazard of sexual initiation than the second. difference was small (HR 0.68 vs 0.71) was actual earlier for the secon d generation. When an outlier was removed (first sex at 22) from the second generation the means were identical. generation women in my sample grew up with two biological parents than was found in Add Health. Co ntraceptive use at sexual initiation between first and second generation. odds of contraceptive use after controlling for two biological parents. generation women us ed contraception at their first sex than the first. Second generation women were more likely to have two biological parents. research showed greater acculturation or time in the U.S. was associated with greater odds of contraceptive use, r egardless of family structure

PAGE 336

320 Table 8 1: Quantitative and Qualitative Findings and Relationship to Prior Literature Adolescent birth a lower hazard of an adolescent birth than the first generation. generation women experienced an adolescent birth than the second generation. ends. Most prior research looked at nativity alone or acculturation. found the lowest hazard of adolescent birth in the second generation. Migration shows that the hazard for sexual initiation increases as age at migration decreases (or time in the U.S. increases). increased as age of migration decreased. birth, age at migration is not linear. Generation 1.25 (migrated age 13 17) had the lowest haza rd, followed by the second generation and then generation 1.5 (migrated 6 13). who migrated prior to age 10 delayed their first birth and completed college than those who migrated after age 10. i n both first and second generation families was associated with more family turmoil and instability and more extreme gendered roles and discourses. participants experienced separation from a parent than the second. Extends by findi ng an association between parental migration patterns, family structure, gender and outcomes. finding reproductive health differences in age at migration. Family structure were protective for all three outcomes amon g all three generations. protective for all three generations for sexual initiation and contraceptive use. protective against sexual initiation in all three generations and non contraceptive use in the second ge neration and adolescent birth in the second generation after controlling for age at sexual initiation. employment, having a resident father still reduced the hazard of first sex. women with two biological parents more often delayed first sex, used contraception at first sex, and delayed their first birth. grandparent during adolescents, grandparents were described as s ources of support. support from older siblings, including sex education, academic and emotional support. Older sisters often had responsibility for their younger siblings. important to delaying fi rst birth and academic success in qualitative participants. research on grandparents and challenges previous research showing older siblings increase risky behaviors. examining the role of fathers.

PAGE 337

321 Table 8 1: Quantitative and Qualitative Findings and Relationship to Prior Literature Family SES education increased the hazard of sexual initiation in the first and second generation. mother was protective for all three outcomes in the first generation and contraceptive use in all three generations. m and communication did not mediate the relationship employment and immigrant generation quality reduced the hazard of sex and maternal communication and supervision reduced the hazard of adolescent birth. diversity in parental education, more first and second generation women with stay at home mothers delayed their first sex and their first birth than those with working mothers. "surviva single or step parent homes, had a parent who migrated independently and extremely gendered roles and discourses in the house. showing parental SES works differently in the first and second generation than third. ends by showing benefits of stay at home mothers in first and second generation women. Gendered discourses different parental expectations for girls than boys was associated with a decrease in the hazard of sexual initiation an d an adolescent birth. between gendered discourses in families and what was actually happened and a lot of variation in gendered roles and discourses within families. reported differences in expectat ion and treatment of sons and daughters. gendered roles and discourses between first and second generation. gendered roles were more likely to have instability, have experienced independent migration often had early sex and an adolescent pregnancy. Marianismo and machismo do not appear to be useful constructs because of the nuance in male and female expectations literature on linking gendered disc ourses and roles in families with migration patterns, family structure, or reproductive health outcomes in adolescents.

PAGE 338

322 Table 8 1: Quantitative and Qualitative Findings and Relationship to Prior Literature Sex education expectations, parental expectations about sex and contraception and religiosity did not mediate pare ntal education and immigrant generation in sexual initiation models. It is still plausible that higher educated migrant parents hold more liberal sexual views. sexual knowledge from parents. educat ion often conveyed that females were both vulnerable and responsible for the actions of men. women complained about differential treatment of brothers when it came to dating and privileges. messages of virginity from their parents and these women delayed sex and delayed first birth as compared to women who did not receive these messages. described in women with early sex and first birth. y women believed their fathers wanted them to wait to have sex until married to protect them, not necessarily due to preventing shame. are consistent with prior research. 109,142,145 of link between sexual messages and reproductive health outcomes. Sexual experience gener ation believed sex was to them and many reported negative first sexual experiences. education from their parents or sought it out themselves did not report negative first sexual experiences. irst generation women felt unprepared for sex than second generation women. lack of sexual subjectivity 145 and commonality of negative first sexual experiences for Mexican origin women. 217 finding first generation felt more unprepared for sex.

PAGE 339

323 Table 8 1: Quantitative and Qualitative Findings and Relationship to Prior Literature Academic education their parents wanted them to succeed in school, those who received parental messages of sacrifice or active parental involvement in their education more often delayed their first birth and achieved higher education were more likely to experience both messages of sacrifice and parental active involvement. messages of sacrifice reported in children left behind in migrant families. linking messages of sacrific e and parental educational actions to reproductive health outcomes. Immigrant generation hazard of sexual initiation than second and third. odds of contraceptive use than second and third. ond generation had the lowest hazard of an adolescent birth, followed the first and finally third generation. employment and family income increased with each generation. and second generation adolesc ents had mothers not working. more likely to live with two biological parents than first and third. parents accounted for some of the differences between the second and third generation in all three outc omes. women had similar ages of sexual initiation had adolescent births women had college degrees and a higher mean years of education women lived in two parent households women had stay at home mothers. reported feeling unprepared at their first sexual experience women experienced messages of sacrifice and parental invo lvement in education prior sexual initiation papers. examining interaction between generation and family factors. exploring gendered discourses, roles and academic discourses between immigrant generations. My primary research question focused on understanding the reasons behind reproductive health differences by immigrant generations. Because I did not interview third generation women, I can only provide evidence for the differences between the first and the se cond generation. It is likely that the biggest reason the third generation is

PAGE 340

324 quantitatively different than the first and second is because the third generation was born to native born (previously second generation) parents and will generally have less tie s to Table 8.1 lists the differences I found in reproductive health outcomes between first and second generation young women. Although there are a number of reasons first and second gener ation adolescents have different outcomes despite having migrant parents, it appears to be the timing of migration that matters most. Migration timing influences family stability, documentation status, language acquisition, and exposure to U.S. peers in th e neighborhood and the school system. The second generation may be more likely to succeed academically and potentially economically and delay their first birth for multiple reasons: they are more likely to live with two biological parents; they learned Eng lish early in school; they did have to not attend school in two different countries; they received messages of sacrifice from their parents; they are U.S. citizens and have access to financial aid and legal employment; and exposure to U.S. peers allows the m to see different traditions and ways of life. one of the biggest differences between the first and the second generation is that first generation women who migrate during a common 237(p281) 18 Second generation adolescents 18 However, one may argue that children in Mexico who have a migrant father in the U.S. may not have a settled life.

PAGE 341

325 multiple cultural world views, st 237(p287) For example, first generation women in my sample who migrated during adolescence were less likely to complain about sibling differential treatment than second generation women and those who migrated during childhood ; perhaps second generat ion women were exposed to different ways siblings are treated in the U.S. whereas differential sibling treatment may have been commonplace in the first generation. Adolescents who have greater exposure to a range of cultures in the U.S., will be more likel y to rebuild and establish new strategies of action in order to solve problems and achieve goals than adolescents who spent most of their childhoods and adolescence in Mexico. Learning new strategies of action will be more costly for later migrants. Altho ugh not necessarily a contradiction, I found that extremely gendered households were associated with family instability and turmoil, however, first and second generation women had similar numbers of extremely gendered households, despite the first generati on having more family instability and turmoil. I went back through my qualitative analysis and even reclassified extreme households and the results are still the same. I finally looked at differences between the eight extremely gendered households and I fo und three possible reasons for this finding. First, place matters. In first generation families, the gendered division of labor and power first occurred in parental relationships while the families were still in Mexico. In second generation families the ge ndered division of labor and power in the relationship occurred in relationships that were formed in the U.S. Thus, the location, including meaning of gender, differences in communities, access to resources and SES between the U.S. and Mexico, are essentia lly confounders or

PAGE 342

326 perhaps mediators in the relationship between family instability, extremely gendered households, and immigrant generations. Second, timing of family instability matters. In all three first generation households with extreme divisions of labor and power and family instability, the extreme gendered division of labor and power occurred before the instability. For example, two of these women knew their mothers migrated to the U.S. after ending unhealthy relati onships with their fathers and th e instability and the migration occurred after, and possibly as a result of the extreme division of labor and power in the household. Thus the instability and extreme gendered division of labor and power actually oc curred prior to migration in first gener ation families. In the second generation the instability and migration appeared to predate the gendered division of labor and power or at least it was more difficult to untangle because two mothers became pregnant out of wedlock and thus began unstable par tnerships that had extreme divisions of labor and power in their first and successive relationships. For example, one mother migrated independently and entered into a pattern of short term unhealthy extremely gendered relationships and one mother entered i nto an unhealthy extremely gendered relationship after her husband was deported. Third, socioeconomics are intertwined with family stability and gendered divisions of labor and power, thus it is difficult to untangle whether the association I found betwee n family instability and extreme divisions of labor and power is really due to underlying socioeconomics. If poverty is the main driver of both family instability and extreme divisions of gender and power in the household that would explain why the first g eneration did not have more households with extreme divisions of labor and power

PAGE 343

327 despite having more family instability. In other words, low SES is associated with extremely gendered households and because first and second generation families in my sample had similar socioeconomics, they had similarly gendered households. Furthermore, three of the four first generation families with extreme divisions of labor and power appeared to be relatively wealthy or at least the women did not describe growing up in p had a skilled job related to airplanes. In the second generation, three of the four women described parents had to work and they moved from state to state to find employment. Bringing culture back in, extremely gendered divisions of labor and power in the household in the first generation may reflect more social and cultural acceptability of male authority whereas in the second generation extremely gendered divisions of labor and power may refl ect poverty and lower SES. The difference in location, timing, and socioeconomics between the first and second generation participants my sample, may be why I found similarly gendered divisions of labor and power in the households, despite the first genera tion having more family instability. While I know the town or state of origin in Mexico of the qualitative participants, I do not have enough detail to understand the influence of regional patriarchies which could also be confounding the relationship. The differences between the first and second generation could be related to sampling bias or could reflect differences in the greater Mexican origin community in the U.S. These does

PAGE 344

328 not undermine my argument and in a sense it reinforces the importance of situa ting culture within a structural context, including place and time. Limitations I discussed limitations in each dissertation chapter. Although I interviewed women who are currently the same ages as the Add Health participants, the largest limitation to th e dissertation is the use of a national quantitative data set and a local qualitative data set, limiting my ability to draw inferences between the two. The most substantial quantitative differences were between the first and third generations and second an d third generations. However, I was not able to provide context for these relationships because I did not interview third generation women in order to keep the dissertation feasible. To me, this is the largest limitation in mixing my results and generating mixed methods conclusions and papers. There were also differences in the two samples, as discussed in Chapter 7, making it difficult to know if results differed between the two because of sampling differences. Area of origin in Mexico likely plays a role in the migration process, family factors, and gendered roles, discourses and sex education, and I did not have that information in Add Health and there was variability in the qualitative sample. I am aware that 21 women is a relatively small qualitative sa mple and that I cannot generalize conclusions to Mexican origin women as a whole. Throughout this dissertation, I have not intended to infer that my results are indisputable or causational or generalizable. However, my results have made an important first step in unravelling the complex migration, family, and gendered processes that influence how Mexican origin young women make reproductive health decisions. Many of my results can stand on their own in separate quantitative and qualitative papers

PAGE 345

329 and a few fit together very well, such as the influence of grandparents and siblings discussed in Chapter 7. Despite these limitations, I answered my research questions and have drawn conclusions and created new questions that I can work on with future research. I was also able to test segmented assimilation theory and conclude with how my results have informed the measurement of acculturation and framing of Latina adolescent childbearing. Implications for future research My results and my limitations have multiple implications for future research. First, addressing one of the largest limitations, I can interview third generation young women in order to provide context and test quant itative associations, like if third generation families hold less sexually conservat ive views and are less likely to encourage virginity than first and second generation families. Second, I can also conduct quantitative analyses with large migration data sets or conduct more qualitative research to corroborate the association between migr ation patterns, specifically independent migration, family instability, and subsequent adolescent outcomes. Third, I can use structural equation modeling (SEM) or further qualitative interviews in a sample with greater variation in parental education to he lp elucidate why higher educated migrant parents increase risk for early sexual initiation. SEM models may also uncover reasons why mothers not working are more protective in the first and second generation than the third. Fourth, I plan to focus future re search on the role of fathers, such as qualitative relationship quality variables to help uncover how fathers play a role in the reproductive health decisions. Fifth, I c an continue qualitative research to understand how

PAGE 346

330 grandparents and older siblings may protect Mexican origin adolescents from risky sexual behaviors. Sixth, community based participatory research (CBPR) may provide a better understanding of how parents ta lk to their children about sex, barriers to sexual communication and educating parents on sex education. Finally, I could use CBPR research and qualitative research to learn more about parental messages of sacrifice and how migrant parents are involved in Implications for practice and policy Policy implications include migration reform that keeps families together and providing support and resources to families who are currently separated due to migration. The recent Executive Ac tion on Immigration, announced on November 20 th 2014, makes great strides in keeping families together by allowing parents, spouses and siblings of U.S. citizens eligible for deferred action and legal employment. 260 However, congress must pass a similar immigration bill in order for elements of this action to be sustainable. Decreasing ex ploitation of migrant workers, providing a path to legal employment and a living wage may reduce the need for women to enter into potentially unstable common law marriages for economic purposes and will also allow parents to spend more time with their chil Public health and clinical practice implications include identifying families in migrant communities who are experiencing separation and providing community support such as helping single families with childcare and educating women and men on abusive relationships. Identifying children in schools and clinical practice who are recent migrants or children of migrants in order to provide emotional support may help bu ffer family disruption. It is also important to change the anti immigrant sentiment in public discourse to include a better

PAGE 347

331 understanding of the economic reasons underlying the migration and settlement process may lead to less discrimination and racism. Sex education from parents and in schools needs to move away from women as vulnerable and men as predatory. Teaching adolescents that they are sexual subjects who are able to sexual pleasure and have control over their bodies and decisions, is crucial to providing comprehensive and empowering sex education. 145,261 CBPR and working with commu nity organizations may help outreach to parents or provide classes in helping parents talk to their children about sex. My qualitative results and the opinions of the women themselves, suggested that teaching parents how to promote academic success in thei r children while considering the language, education, and time barriers that they have may promote education attainment in adolescents and may reduce adolescent childbirth. My qualitative data also suggested that conflict also occurred when young women mee t women may experience internal conflict as well as parent child conflict because they have chosen higher education, and that should not be overlooked in the school sy stem because of their high achieving status. M easuring assimilation and acculturation It is important to remember that while I tested the utility of segmented assimilation theory to understand reproductive health outcomes in Mexican origin adolescents of varying immigrant generations, segmented assimilation theory was designed to predict economic assimilation in the second generation. Further, the influence of neighborhoods and ethnic communities on economic assimilation, a key feature of segmented assimi lation theory, was beyond the scope of my dissertation. That being said,

PAGE 348

332 adolescent childbearing has been associated with downward or at least more difficult upward assimilation and Portes has tested the theory with adolescent childbearing as an outcome. 262 Based on m y quantitative and qualitative results, I now draw conclusions on the strengths and weaknesses of using segmented assimilation theory to understand reproductive health behaviors. Strengths of SAT My results have supported a number of aspects of segmented assimilation theory. First, both qualitative and quantitative results support the association between an intact family structure and delayed sexual initiation and childbearing and an increase in contraceptive use at first sex. I extended the theory to hyp othesize that family structure would be more protective in the first and second generation but this was not supported by my results; having two biological parents in the home is protective in all generations. Extension of SAT to include grandparents and ol der siblings as protective was supported in my qualitative and quantitative analyses further demonstrating that close, supportive families can improve reproductive health outcomes in adolescents. ulturation, its association with upward economic assimilation, and its relationship to an intact family structure and generally higher SES. Qualitative participants, such as Isabel, Cristina, and Arely, discussed in detail in Chapter 4, had parents who wer e able to provide cultural teachings such as the importance of the Spanish language, Mexican food and traditions, future upward mobility through parental messages of sac rifice and active participation in

PAGE 349

333 from stable two parent families, although there was some variation in socioeconomic status. Women who experienced selective acculturation Although peer groups were not the focus of this study, I did find some qualitative evidence to support that two biological parent families an d selective acculturation could buffer deviant peer influences as suggested by SAT. Finally, consistent with SAT, there are different economic life trajectories of the second generation and not all children of migrants assimilate into a middle class as on ce thought. This trajectory was associated with family SES and family structure as expected, although not necessarily in the way SAT predicts. Weaknesses of SAT There are numerous weakness of segmented assimilation theory in the use of predicting and und erstanding reproductive health outcomes. First, I used the theory of gender and power as an organizational framework to explore the role of gender in migration, family, and reproductive health decision making of Mexican origin young women because SAT does not even acknowledge the role of gender. My qualitative results show that gendered discourses and roles and parental expectations of sex and dating are crucial to reproductive health decision making in women and are linked to family structure and SES. For example, almost all of my participants described how their parents treated them differently than their brothers including responsibilities and expectations. This gender difference is crucial to understanding adolescent decisions and outcomes. Second, SAT a ssumes that parental cultural teachings derived from the country of origin is always protective. I found that cultural teachings are not always

PAGE 350

334 protective and in extremely gendered households reinforcement of a gendered division of labor and power and a di scourse on the vulnerability of women and their responsibility for the behavior of men increased risk for early sexual initiation, early childbearing and non contraceptive use. Third, SAT asserts that higher educated parents and skilled or professionally employed parents are associated with improved economic assimilation. My quantitative the first, and sometimes second generation, and are not necessarily protective for sexu al initiation, contraceptive use or risk for an adolescent birth. In fact, in the first generation, having a mother not working was more protective for the three reproductive health outcomes than having a professionally employed mother. It is possible that negative selection, not necessarily segmented assimilation, is the cause of different adolescent trajectories. A very small proportion of Mexican migrants have achieved higher education or skilled and professional employment so the greater variation in ec onomic outcomes of the second and third generation could be related to greater parental economic variation. Fourth, while some participants, such as Isabel, Cristina, and Arely, exemplified les of consonant or dissonant acculturation. This does not mean they do not exist but it means there are other patterns of acculturation that SAT does not address. Although there is general consensus that there is not just one pattern of upward assimilatio n as classic assimilation theorists thought, 86 my results and other researchers suggest that there are more than the three 215,263

PAGE 351

335 Fifth, although not necessarily related to my research question, my results suggest that there is a spectrum of upward and downward assimilation that is f lexible with time. While the women in my study described that having a child during adolescence made it more difficult for them to continue their education, four of the nine women who experienced an adolescent birth, returned to school and finished their G ED or associates degrees and two are enrolled in Bachelor degrees programs in their late 20s. Some married or se juntaron with men who have skilled employment, one previous adolescent mother has her children enrolled in private school and some are home own ers. I do not believe that adolescent motherhood necessarily caused downward assimilation into an Finally, I found that family migration patterns were associated with family structure and SES which then impacted gender and subsequent reproductive health decisions. SAT considers U.S. labor markets, the sociopolitical climate of the U.S., and more recent papers have examined the relationship between migrant flows, social networks and remittances with segmented assimilation, 262 it does not account fo r family patterns of migration. Table 8.2 summarizes the strengths and weaknesses of SAT. Table 8 1 : Strengths and Weaknesses of SAT to Understand Reproductive Health B ehaviors Strengths Weaknesses Recogniti on of the important and protective role of a stable, two parent family in improving reproductive health and educational outcomes of adolescents in migrant families No acknowledgement of the critical role of gender in the assimilation process or reproductiv e health outcomes of adolescents in migrant families. Selective acculturation, parental cultural teachings educational success is associated with improved reproductive health and educational outcomes of adole scents in migrant families Parental cultural teachings and values are not always protective for reproductive health outcomes. Acknowledgement of multiple life trajectories and varying types of economic assimilation rather than on into a U.S. middle class. Higher parental education and skilled or professional employment is not necessarily protective for reproductive health outcomes, particularly in the first and second generation.

PAGE 352

336 Table 8 1 : Strengths and Weaknesses of SAT to Understand Reproductive Health B ehaviors There is more than the selective, consonant, a nd dissonant acculturation. There are more than three segments of assimilation and mobility has a wide range spectrum that changes with time. Segmented assimilation theory has a number of strengths and prior research has found full or partial support for its use in examining academic and economic outcomes of second generation youth as well as risky behaviors such as substance abuse. 1 22 124,204 After examining my own research and current literature, I do not recommend segmented assimilation theory as the best theoretical framework for understanding reproductive health outcomes in Mexican origin adolescents because migration patterns a nd gender are missing from the theory and because adaptation, whether cultural or economic, has more than three segments. Although no theory can address all influences of human behavior, combining the theory and gender and power significantly strengthens t he use of the theory as well as acknowledging the migration process. In my future work, I will consider the useful aspects of segmented assimilation theory such as the importance of family structure, and the potential benefits of maintenance of some cultur al ties. I also believe the influence of neighborhoods and communities on the reproductive health outcomes of Mexican origin adolescents is worth further examination. I will take these strengths of SAT, the organization of the theory of gender and power, a the processes of adaptation, change, and economic assimilation, in Mexican origin kinds of adaptations to changing circumstances and with the line having no predictable 263(p44) Instead of only three patterns of economic assimilation, I will allow for a spectrum of trajectories and outcomes that are influenced by many factors, including,

PAGE 353

337 family migration patterns, family structure, family SES, gendered discourses, academics, employment, neighborhoods, peers, intermarriage, discrimination, personal agency and outcomes of each deci sion. While I cannot examine every life experience that is related the life course, adaptation, and economic incorporation of migrants and children of migrants, it is important to consider there are more than only three paths and multiple factors that infl uence human behavior and decision making. Acculturation In Chapter 1, I discussed how acculturation has been inconsistently defined and measured in social science and public health literature. I hoped that my dissertation would add to a more comprehensiv e understanding of acculturation but instead, I now agree with Hunt, Schneider, and Comer 84 that acculturation should not be used as a variable in health research, at l east in the way it has been measured in the past. For one reason, one dimensional variables, such as language, have been used to measure the complex process of adaptation in Latino adolescents. Studies assume that adolescents who speak Spanish at home are less acculturated tha n adolescents who speak English at home which may account for sexual and reproductive health differences. However, all of my qualitative participants grew up speaking Spanish in the home and there is much variation in their reproductiv e health outcomes. My qualitative results have shown that there was not one set of traditional values but examined through gendered discourses and roles, academic expectations, educational actions, cultural practices and teachings of Mexican family membe rs, were associated with socioeconomic and structural factors. Socioeconomics contributed to the migration process, the formation and stability of families, the education and employment of parents

PAGE 354

338 and their children, and access to resources. My quantitativ e results also supported the relationship between socioeconomics and culture because I found that higher parental education and employment is not necessarily beneficial for first and second generation adolescents perhaps because of differences in sexual vi ews is connected to region in Mexico (urban and rural) and access to higher education and employment in those regions. 195 I also found that the presence of a grandparent, who would be the least contraceptive use at first sex. Clearly, cultu re is not the only explanation for health behavior differences among adolescents. reproductive health literature has a tendency cture and the dynamic social processes in which behavior and beliefs are generated, and to relegate consideration of the socio economic challenges associated with immigration, poor English language skills, and poverty, to their effects as separate or confo unding variables. 84(p981) For example, Chapter 1, Table 1.2, listed 13 quantitative studies examining reproductive h ealth outcomes by nativity or acculturation in Latino adolescents. All 13 used SES measures as controls. Even in papers using SEM models, SES was not part of the pathway connecting family, expectations, acculturation, nativity and outcomes. 25,190 I firmly believe that cultural change and adaptation through migrati on and settlement need to be studied but there are more valid and comprehensive ways in which to do this, without separating cultural and structure and without reducing cultural processes to simple stereotypes. Hunt, Schneider, and Comer suggest, There is great potential for interdisciplinary research to generate more realistic and useful models of the impact of culture on health, incorporating consideration

PAGE 355

339 of the range of cultural, social, economic, and political conditions pertinent to the groups in ques the specific cultural elements in question within their actual cross cultural and historical context to replace sweeping assertions about Hispanic culture, which could lead to better understa nding of key issues that impact both cultural change and health, such as the practical realities of immigrant life and the harsh influences of discrimination. 84(pp981 2) My dissertation provides an important step in working toward this suggested interdisciplinary research. I have explored specific aspects of culture, such as gendered discourses and roles, within a socioeconomic context that considers time, place, and country of origin. The association between migration patterns, family socioeconomics and structure, gendered discourses and roles, expectations of sex and dating, parental academic expectations and ac tions, and reproductive health behaviors of adolescents is absolutely crucial to providing a more complicated and accurate study and framing of reproductive health in Mexican origin adolescents. In my future research, I will not study or measure acculturat ion but rather explore these complex processes with qualitative research or include SES variables in quantitative models. R eframing Latina Adolescent Childbearing In addition to reframing and complicating cultural constru ctions to better understand how migrants and their families change through migration, settlement and intergenerational processes, I now recommend ways to reframe Latina adolescent childbearing. I advocate for a more comprehensive understanding of adolescen t childbearing that moves away from blaming the young woman, her family, and her culture. In Chapter 1, I discussed two ways that adolescent childbearing has been inappropriately framed. First, literature and public discourse have focused on adolescent chi ldbearing, particularly Latina adolescent childbearing, as a cause of poverty and other

PAGE 356

340 social ills. Second, literature and public discourse have oversimplified poverty and culture as mutually exclusive causes of adolescent pregnancy and childbearing. Thro ugh the results and conclusions I have drawn from my dissertation, I recommend that we concentrate on the sociocultural and socioeconomic factors and the connection between them that predisposes young women to adolescent births and other risky reproductive health behaviors. Latinas and their parents are often blamed for high rates of adolescent motherhood above education. 71 A 2004 articl pregnancy prevention goals without interviewing any La tino families. This paragraph was part of their background section, Research on Hispanic cultural values, attitudes and behaviors that influence sexual and contraceptive behavior has shown that sexuality is often a taboo subject and that parental communic ation regarding sexuality is often lacking in Hispanic homes. At the same time, Hispanic culture supports early and high fertility, as well as the belief that early motherhood and continued education are incompatible. These cultural values may explain why Hispanic women desire marriage and children at a younger age than do blacks, Southeast Asians and whites. In addition, Mexican Americans are more likely than whites to believe 71(p142) Although awareness of cultural differences is importance, researchers must take great care in how they present findings related to culture in race/ethnic populations. This quote feeds into stereotypes that paint a pictu re of Mexican origin families who do not provide any sexual education to their adolescents, encourage early childbearing and marriage, and do not value education. In addition to increasing racism and discrimination,

PAGE 357

341 being at risk of confirming, as self 264(p797) The stereotype threat has been associated with poor academic performance in minority students. 264,265 If the media and research continues to foster these stereotypes, are we not contributing to Latino families and their children confirming thes e stereotypes instead of proving them wrong? Prior research 109,145 and my own have chal lenged these stereotypes about Latino families. Although the Latina adolescent birth rate is higher than other race/ethnic groups, in the last decade, they experienced the most rapid decline. In 2010 the Latina adolescent birth rate was 0.8 per 1,000 10 14 year olds and 55.7 per 1,000 15 19 year olds. Although this rate is higher than other race/ethnic groups, less than half of Latina adolescents are giving birth. This should caution us from blaming Latino families for promot ing early luing motherhood above education adolescents are not becoming adolescent mothers. and value education, is the fact that a higher proportion of second generation Latino and Asian immigrants have college degrees than the general U.S. population. More second generation Latinos hold college degrees than the general Latino population as a whole. 233 My results showed that second generation Mexican origin adolescents had a lower hazard of first birth than firs t or third generation adolescents. Clearly, the upward mobility of the second generation means that Latino migrant parents are not failing as parents nor do they promote motherhood above education. In fact, looking at differences between the second and thi rd generations, one might argue that migrant parents of U.S. born children

PAGE 358

342 are parenting better than native born Latino parents, thus we cannot continue to blame a I am not implying that we do not need to study and work to reduce the Latina adolescent birth rate. I have provided this data because I argue that breaking down reducing health disparities in the Latino and Mexican origin population. Instead of 237 argument, also supported and used by Garcia, 145 My results have demonstrated that cultural values are highly nuanced and depend upon time period, place, and socioeconomics; there is not a specific set of traditional cultural values. Further, i t is their physical health yet we have a global obesity epidemic. Swidler suggests we study culture thro provides an example about the disconnect between values and outcomes in poor minority youth. Prior research has found that poor minority youth, who are often unlikely to atten d college or even finish high school, tell researchers that they value education and want to style, and informal know how are unfamiliar. One does better to look for a line of action 237(p275) This vi ew allows us to having the cultural strategies of action or tools needed to promote academic success in their children and allows us to understand why young poor Mexicanas may choose a culturally acceptable motherhood, as their mothers and grandmothers chose in Mexico,

PAGE 359

343 y and strength to one of my central arguments, that Mexican origin adolescent girls and their families value education and a delay in childbearing but their structural circumstances make it difficult have the cultural 237(p281) Removing blame and stereotypes from Latin o families does not mean that Latino families play no role in adolescent reproductive health disparities but places some that they are vulnerable yet responsible for the a ctions of men or not taking an active role of shaming parents who cannot attend parent teacher conferences or help with homework due to language or time barriers, we could continue to fight for policy reform that constraints that they have. Thro ughout this dissertation, I did not mean to ignore other ways to reduce reproductive health disparities in adolescents, including increasing availability and access to comprehensive sex education and long acting reversible contraception which have been sho wn to reduce adolescent birth rates in low income populations. I do not suggest that adolescents have no agency and are at the mercy of their family and other factors I have studied. Rather, given my focus on migration, family, and gender and how they impa ct reproductive health decision making in adolescents, I am not able to infer more

PAGE 360

344 about the processes through which the young women negotiates, rejects or reproduces these influences. Adolescents of all race/ethnic groups need to learn personal responsibi lity, as well view themselves as sexual subjects who have a right to sexual pleasure, agency over what they do with their bodies, and the ability to make decisions such as when to have sex, when to use contraception, and if they want to terminate, adopt, o agency is limited by forces larger than themselves such as the process of migration, family structure and SES, parental teachings, institutionalized racism, anti immigrant sentiment After decades of research confirming that poverty and its surrounding web of peril (family instability, substance abuse, child abuse and neglect, intimate partner violence, rape) increases risk for adolescent pregnancy and childbearing, it is time for pu blic discourse, the academy, and politicians to make structural forces a priority.

PAGE 361

345 REFERENCES 1. Palacio M, Reynolds R, Drisko J, Lucero C, Hunt C, Phi K. Racial and ethnic health disparities in Colorado 2009. Office of Health Disparities, Colorado Department of Public Health and Environment 2009. 2. Centers for Disease Control and Prevention. Sexually transmitted disease surveillance 2009. Atlanta: US Department of Health and Hu man Services 2010. 3. Centers for Disease Control and Prevention. HIV surveillance report, 2009; vol. 21. 2011. http://www.cdc.gov/hiv/topics/surveillance/resources/reports/. Accessed October 30, 2011. 4. Martin JA, Hamilton BE, Osterman MJK, Curtin SC, Mathews TJ. Births: Final Data for 2012 U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, National Vital Statistics System; 2013. http://www.cdc.gov/nchs/data/nvsr/nvsr62/nvsr 62_09.pdf. Accessed May 27, 2014. 5. Kearney MS, Levine PB. Why Is the Teen Birth Rate in the United States so High and Why Does It Matter? National Bureau of Economic Research; 2012. http://www.nber.org/papers/w17965. Accessed June 30, 2013. 6. US Cens us Bureau. Projected Population by Single Year of Age, Sex, Race, and Hispanic Origin for the United States: 2012 to 2060 U.S. Census Bureau Population Division; 2012. 7. Ennis SR, Rios Vargas M, Albert NG. The Hispanic Population: 2010. United States Census Bureau 2011. 8. Passel J, Cohn D. U.S. Population Projections: 2005 2050. 2008. http://www.pewhispanic.org/2008/02/11/us population projections 2005 2050/. 9. Abraido Lanza AF, Armbrister AN, Florez KR, Aguirre AN. Toward a theory driven model of acculturation in public health research. Am J Public Health 2006;96(8):1342 1346. doi:10.2105/AJPH.2005.064980. 10. Brindis C, Wolfe AL, McCarter V, Ball S, Starbuck Morales S. The associations between immigrant status and risk behavior patterns in Lati no adolescents. Journal of Adolescent Health 1995;17(2):99 105. doi:doi: DOI: 10.1016/1054 139X(94)00101 J. 11. Afable Munsuz A, Brindis CD. Acculturation and the sexual and reproductive health of Latino youth in the United States: a literature review. P erspectives on Sexual and Reproductive Health 2006;38(4):208 219.

PAGE 362

346 12. Pew Hispanic Center. Demographic Profile of Hispanics in Colorado, 2009. 2011. http://pewhispanic.org/. Accessed October 11, 2011. 13. Lopez MH, Gonzalez Barrera A, Cuddington D. Appe ndix: Rankings of the 14 Largest Hispanic Origin Groups. 2013. http://www.pewhispanic.org/2013/06/19/appendix rankings of the 14 largest hispanic origin groups/. Accessed April 5, 2015. 14. Upchurch DM, Mason WM, Kusunoki Y, Kriechbaum MJ. Social and Behavioral Determinants of Self Reported STD Among Adolescents. Perspectives on Sexual and Reproductive Health 2004;36(6):276 287. 15. Boonstra HD. What is behind the declines in teen pregnancy rates? Guttmacher Policy Review 2014;17(3). https://www.guttmacher.org/pubs/gpr/17/3/gpr170315.html. Accessed April 5, 2015. 16. Hamilton BE, Martin JA, Ventura SJ. Births: Preliminary Data for 2012 National Vital Statistics System, Centers for Disease Control and Preve ntion; 2013:20. http://www.cdc.gov/nchs/data/nvsr/nvsr62/nvsr62_03.pdf. Accessed March 23, 2014. 17. Hamilton BE, Ventura SJ. Birth rates for U.S. teenagers reach historic lows for all age and ethnic groups. NCH data brief, no. 89. 2012. 18. Ventura SJ, Hamilton BE, Mathews TJ. Pregnancy and Childbirth Among Females Aged 10 19 Years United States, 2007 2010. Morbity and Mortality Weekly 2013;62(3):71 76. 19. Ford K, Norris AE. Urban Hispanic adolescents and young adults: Relationship of acculturation to sexual behavior. Journal of Sex Research 1993;30(4):316 323. 20. Gilliam ML, Berlin A, Kozloski M, Hernandez M, Grundy M. Interpersonal and personal factors influencing sexual debut among Mexican American young women in the United States. J Adolesc He alth 2007;41(5):495 503. 21. Guarini TE, Marks AK, Patton F, Coll CG. The immigrant paradox in sexual risk behavior among Latino adolescents: impact of immigrant generation and gender. Applied Developmental Science 2011;15(4):201 209. doi:10.1080/108886 91.2011.618100. 22. Guilamo Ramos V, Jaccard J, Pena J, Goldberg V. Acculturation Related Variables, Sexual Initiation, and Subsequent Sexual Behavior Among Puerto Rican, Mexican, and Cuban Youth. Health Psychology 2005;24:88 95. doi:10.1037/0278 6133.24 .1.88.

PAGE 363

347 23. Jimenez J, Potts MK, Jimenez D. Reproductive attitudes and behavior among Latina adolescents. Journal of Ethnic & Cultural Diversity in Social Work 2002;11(3/4):221 249. doi:10.1300/J051v11n03_04. 24. Kaplan CP, Erickson PI, Juarez Reyes M. A cculturation, gender role orientation, and reproductive risk taking behavior among Latina adolescent family planning clients. Journal of Adolescent Research 2002;17(2):103 121. doi:10.1177/0743558402172001. 25. Killoren SE, Deutsch AR. A Longitudinal Exa mination of Parenting Processes and J Youth Adolescence 2013:1 12. doi:10.1007/s10964 013 0053 z. 26. Landale NS, Schoen R, Daniels K. Early family formation among white, black, and Mexican American women. J Fam Iss ues 2009;31:445 474. doi:10.1177/0192513X09342847. 27. Lee J, Hahm HC. Acculturation and sexual risk behaviors among Latina adolescents transitioning to young adulthood. J Youth Adolescence 2009;39(4):414 427. doi:10.1007/s10964 009 9495 8. 28. McDonal d JA, Manlove J, Ikramullah EN. Immigration measures and reproductive health among Hispanic youth: Findings from the national longitudinal survey of youth, 1997 2003. Journal of Adolescent Health 2009;44(1):14 24. 29. Minnis AM, Padian NS. Reproductive h ealth differences among Latin American and US born young women. Journal of Urban Health: Bulletin of the New York Academy of Medicine 2001;78(4):627 637. doi:10.1093/jurban/78.4.627. 30. Collins PH. Toward a new vision: Race, class, and gender as catego ries of analysis and connection. In: Race, Sex & Class .Vol 1.; 1993:25 45. 31. Driscoll AK, Biggs MA, Brindis CD, Yankah E. Adolescent Latino reproductive health: a review of the literature. Hispanic J Behav Sci 2001;23(3):255. 32. Chen X K, Wen SW, Fle ming N, Demissie K, Rhoads GG, Walker M. Teenage Pregnancy and Adverse Birth Outcomes: A Large Population Based Retrospective Cohort Study. Int J Epidemiol 2007;36(2):368 373. doi:10.1093/ije/dyl284. 33. Hoffman S. By the numbers: The public costs of tee n childbearing. Washington, DC: National Campaign to Prevent Teen Pregnancy 2006. 34. Manlove JS, Terry Humen E, Mincieli LA, Moore KA. Outcomes for children of teen mothers from kindergarten through adolescence. In: Hoffman SD, Maynard RA, eds. Kids Hav Pregnancy Second. Washington, D.C.: Urban Institute Press; 2008:161 196. 35. Sainer S. The state of adolescent sexual health in Colorado. 2011.

PAGE 364

348 36. Goerge RM, Harden A, Joo Lee B. Consequences of t een childbearing for child abuse, neglect, and foster care placement. In: Hoffman SD, Maynard RA, eds. Kids Second. Washington, D.C.: Urban Institute Press; 2008:257 288. 37. Hardy JB, Shapiro S, Astone NM, Miller TL, Brooks Gunn J, Hilton SC. Adolescent Childbearing Revisited: The Age of Inner City Mothers at Delivery Is Sufficiency at Age 27 to 33. Pediatrics 1997;100(5):802 809. doi:10.1542/peds .100.5.802. 38. East P, Reyes BT, Horn EJ. Association between adolescent pregnancy and a family history of teenage births. Perspectives on Sexual and Reproductive Health 2007;39(2):108 115. doi:10.1363/3910807. 39. East P, Slonim A, Horn EJ, Trinh C, R Perspect Sex Reprod Health 2009;41(4):210 217. 40. Grogger J. Consequences for incarceration among adult children: Approach and e stimates through 1991. In: Hoffman SD, Maynard RA, eds. Economic Costs & Social Consequences of Teen Pregnancy Second. Washington, D.C.: Urban Institute Press; 2008:290 311. 41. Scher LS, Hoffman SD. Consequences for incarceration among adult children: Updated estimates through 2002. In: H offman SD, Maynard RA, eds. Kids Having Second. Washington, D.C.: Urban Institute Press; 2008:311 321. 42. Hotz VJ, Williams Elroy S, Sanders SG. Consequences of teen childbearing for mothers through 1993. In: Hoffman SD, Maynard RA, eds. Economic Costs & Social Consequences of Teen Pregnancy Second. Washington, D.C.: Urban Institute Press; 2008:52 118. 43. The National Campaign to Prevent Teen and Unplanned Pregnancy. Coun ting It Up. The Public Costs of Teen Childbearing: Key Data .; 2013. https://thenationalcampaign.org/resource/counting it key data 2013. Accessed September 27, 2014. 44. Reid D, Roberts D, Gonzalez Rojas J, Gilliam M. The continuing debate on the Medicaid 2014. 45. Furstenberg FF. Teenage childbearing as a public issue and private concern. Annu Rev Sociol 2003;29(1):23 39. doi:10.1146/annurev.soc.29.010202.100205. 46. Making the Case: Fo r Wanted and Welcomed Pregnancy. The National Campaign to Prevent Teen and Unplanned Pregnancy 2014. http://thenationalcampaign.org/why it matters. Accessed October 16, 2014.

PAGE 365

349 47. Furstenberg FF. As the pendulum swings: Teenage childbearing and social con cern. Family Relations 1991;40(2):127 138. 48. Singh S, Darroch JE, Frost JJ. Socioeconomic Disadvantage and Adolescent Countries. Family Planning Perspectives 2001;33(6):251 289. doi: 10.2307/3030192. 49. Mollborn S, Morningstar E. Investigating the relationship between teenage childbearing and psychological distress using longitudinal evidence. Journal of Health and Social Behavior 2009;50(3):310 326. doi:10.1177/002214650905000305. 50. Geronimus AT, Korenman S. The Socioeconomic Consequences of Teen Childbearing Reconsidered. The Quarterly Journal of Economics 1992;107(4):1187 1214. doi:10.2307/2118385. 51. Manlove J, Steward Streng N, Peterson K, Scott M, Wildsmith E. Racial and Ethnic Differences in the Transition To a Teenage Birth in the United States. Perspectives on Sexual and Reproductive Health 2013;45(2):89 100. doi:10.1363/4508913. 52. National Center for Health Statistics. Health, United States, 2011: with special feat ure on socioeconomic status and health. 2012. http://www.cdc.gov/nchs/data/hus/hus11.pdf. Accessed December 10, 2012. 53. Lopez MH, Velasco G. Childhood poverty among Hispanics sets record, leads nation. Pews Hispanic Center 2011. 54. Brown A, Patten E. Statistical Portrait of the Foreign Born Population in the United States, 2012 Washington, DC: Pew Research Center; 2014. http://www.pewhispanic.org/2014/04/29/statistical portrait of the foreign born population in the united states 2012/. Accessed June 13, 2014. 55. Brown A, Patten E. Statistical Portrait of Hispanics in the United States, 2012 Washington, DC: Pew Research Center; 2014. http://www.pewhispanic.org/2014/04/29/statistical portrait of hispanics in the united states 2012/. Accessed Septembe r 11, 2014. 56. Fernandez Kelly MP, Schauffler R. Divided fates: Immigrant children and the new assimilation. In: Portes A, ed. The New Second Generation New York: Russell Sage Foundation; 1996:30 53. 57. Portes A, Fernndez Kelly P, Haller W. The Adap tation of the Immigrant Second Generation in America: A Theoretical Overview and Recent Evidence. Journal of Ethnic and Migration Studies 2009;35(7):1077 1104. doi:10.1080/13691830903006127.

PAGE 366

350 58. Rumbaut RG. The crudible within: Ethnic identity, self est eem, and segmented assimilation among children of immigrants. In: Portes A, ed. The New Second Generation New York: Russell Sage Foundation; 1996:119 170. 59. Ngai M. Princeton N.J.: Princeton University Press; 2004. 60. Calavita K. New York u.a.: Routledge; 1992. 61. The Columbia History of Latinos in the United States since 1960 New York N.Y.: Columbi a University Press; 2004. 62. State of Colorado. House Bill 06S 1023 frequenty asked questions. 2006. http://www.colorado.gov/cs/Satellite?blobcol=urldata&blobheadername1=Content Disposition&blobheadername2=Content Type&blobheadervalue1=inline%3B+filename %3D%22HB06S+1023+Frequently +Asked+Questions.pdf%22&blobheadervalue2=application%2Fpdf&blobkey=id& blobtable=MungoBlobs&blobwhere=1251810020076&ssbinary=true. Accessed September 2, 2012. 63. State of Arizona. State of Arizona House of Representatives Forty ninth Legislature Second Regular Session 2010 HOUSE BILL 2281. 2011. http://www.azleg.gov/legtext/49leg/2r/bills/hb2281s.pdf. 64. State of Colorado. Colorado ASSET .; 2013. http://www.coloradoasset.com/. Accessed September 21, 2014. 65. Castles S, Miller MJ. The Age of Migration Guilford Press; 2003. 66. Mahler S. Princeton N.J.: Princeton University Press; 1995. 67. Massey DS, Riosmena F. Undocumented Migration from Latin America in an Era of Rising U.S. Enforcement. The ANNALS of the American Academy of Politica l and Social Science 2010;630(1):294 321. doi:10.1177/0002716210368114. 68. Passel J, Cohn D. Unauthorized immigrant population: National and state trends, 2010. 2011. 69. Frank R, Heuveline P. A crossover in Mexican and Mexican American fertility rates : Evidence and explanations for an emerging paradox. Demographic Research 2005;12(4):77 104. 70. Chavez LR. A Glass Half Empty: Latina Reproduction and Public Discourse. Human Organization 2004;63(2):173 188.

PAGE 367

351 71. pectives on effective practices for Hispanic teenage pregnancy prevention. Perspectives on Sexual and Reproductive Health 2004;36(4):142 149. 72. Chavez L. The Latino Threat: Constructing Immigrants, Citizens, and the Nation Stanford University Press; 2 013. 73. Farmer P, Conners M, Simmons J. Women, Poverty and AIDS First. Monroe, ME: Common Courage Press; 1997. 74. Adolph C, Ramos DE, Linton KL., Grimes DA. Pregnancy among Hispanic teenagers: Is good parental communication a deterrent? Contraception 1995;51(5):303 306. 75. Talashek ML, Peragallo N (Nena), Norr K, Dancy BL. The context of risky behaviors for Latino youth. J Transcult Nurs 2004;15(2):131 138. doi:10.1177/1043659603262489. 76. Rocca CH, Hubbard AE, Johnson Hanks J, Padian NS, Minnis AM. Predictive Latino Community. Studies in Family Planning 2010;41(3):179 192. 77. Biggs MA, Ralph L, Minnis AM, et al. Factors associated with delayed childbearing: From the voices of expectant Latina adults and teens in California. Hispanic Journal of Behavioral Sciences 2010;32(1):77 103. doi:10.1177/0739986309353000. 78. Gilliam ML, Warden MM, Tapia B. Young Latinas recall contraceptive use before and after pregnancy: a f ocus group study. J Pediatr Adolesc Gynecol 2004;17(4):279 287. 79. Rivera CP, Mndez CB, Gueye N A, Bachmann GA. Family planning attitudes of medically underserved Latinas. 2007;16(6):879 882. doi:10.1089/jwh.2007.0366. 80. D uRant RH, Pendergrast R, Seymore C. Sexual behavior among Hispanic female adolescents in the United States. Pediatrics 1990;85(6):1051. doi:Article. 81. Heavey EJ, Moysich KB, Hyland A, Druschel CM, Sill MW. Differences in pregnancy desire among pregnant female adolescents at a state funded family planning clinic. J Midwifery Womens Health 2008;53(2):130 137. 82. Dogan Ates A, Carrion Basham CY. Teenage pregnancy among Latinas: examining risk and protective factors. Hispanic Journal of Behavioral Scienc es 2007;29(4):554 569. doi:10.1177/0739986307308163. 83. Lopez Class M, Castro FG, Ramirez AG. Conceptions of acculturation: A review and statement of critical issues. Soc Sci Med 2011;72:155 162.

PAGE 368

352 84. research? A critical review of research on US Hispanics. Social Science & Medicine 2004;59(5):973 986. doi:10.1016/j.socscimed.2003.12.009. 85. Riosmena F, Everett BG, Rogers RG, Dennis JA. Negative Acculturation and Nothing Mo re? Cumulative Disadvantage and Mortality during the Immigrant Adaptation Process among Latinos in the U.S. Int Migr Rev 2014:n/a n/a. doi:10.1111/imre.12102. 86. Zhou M. Segmented assimilation: Issues, controversies, and recent research on the new sec ond generation. International Migration Review 1997;31(4):975 1008. 87. Portes A, Zhou M. The new second generation: Segmented assimilation and its variants. The ANNALS of the American Academy of Political and Social Science 1993;530(1):74 96. doi:10.11 77/0002716293530001006. 88. Portes A, Fernndez Kelly P, Haller W. Segmented assimilation on the ground: The new second generation in early adulthood. Ethnic and Racial Studies 2005;28(6):1000 1040. doi:10.1080/01419870500224117. 89. Clark RL, Glick JE, Bures RM. Immigrant Families Over the Life Course Research Directions and Needs. Journal of Family Issues 2009;30(6):852 872. doi:10.1177/0192513X09332162. 90. Glick JE, Bean FD, Hook JVWV. Immigration and Changing Patterns of Extended Family Household Structure in the United States: 1970 1990. Journal of Marriage and Family 1997;59(1):177 191. doi:10.2307/353671. 91. Van Hook J, Glick JE. Immigration and Living Arrangements: Moving Beyond Economic Need Versus Acculturation. Demography 2007;44(2):225 249. doi:10.1353/dem.2007.0019. 92. Landale NS, Thomas KJA, Van Hook J. The Living Arrangements of Children of Immigrants. The Future of Children 2011;21(1):43 70. 93. Aneshensel CS, Fielder EP, Becerra RM. Fertility and Fertility Related Behavior Among Mexican American and Non Hispanic White Female Adolescents. Journal of Health and Social Behavior 1989;30(1):56 76. 94. postpone onset of sexual intercourse: a survival anal ysis of virginity among youths aged 13 to 18 years. Journal of Adolescent Health 2000;26(1):42 48. doi:10.1016/S1054 139X(99)00041 5. 95. Merten MJ, Henry CS. Family Structure, Mother Daughter Relationship Quality, Race and Ethnicity, and Adolescent Girl Journal of Divorce & Remarriage 2011;52(3):164 186. doi:10.1080/10502556.2011.556966.

PAGE 369

353 96. Upchurch DM, Aneshensel CS, Mudgal J, McNeely CS. Sociocultural contexts of time to first sex among Hispanic adolescents. Journal of Marriage and F amily 2001;63(4):1158 1169. 97. Costa FM, Jessor R, Fortenberry JD, Donovan JE. Psychosocial conventionality, health orientation, and contraceptive use in adolescence. Journal of Adolescent Health 1996;18(6):404 416. doi:10.1016/1054 139X(95)00192 U. 98 Baumeister LM, Flores E, Marn BV. Sex information given to Latina adolescents by parents. Health Educ Res 1995;10(2):233 239. doi:10.1093/her/10.2.233. 99. Corcoran J, Franklin C, Bennett P. Ecological factors associated with adolescent pregnancy and parenting. Social Work Research 2000;24(1):29 39. doi:Article. 100. Hill MS, Yeung W JJ, Duncan GJ. Childhood family structure and young adult behaviors. Journal of Population Economics 2001;14(2):271 299. doi:10.1007/s001480000039. 101. Bryson K, Cas per L. Current population reports: Coresident grandparents and grandchildren. 1999. 102. Caputo RK. Grandparents and Coresident Grandchildren in a Youth Cohort. Journal of Family Issues 2001;22(5):541 556. doi:10.1177/019251301022005001. 103. Monserud M Attainment: A Perspective on Coresidence with Grandparents. Journal of Marriage and Family 2011;73(5):981 1000. doi:10.1111/j.1741 3737.2011.00858.x. 104. Perez L. Households of children of immi grants in South Florida. In: Portes A, ed. The New Second Generation New York: Russell Sage Foundation; 1996:108 118. 105. Bleakley A, Fishbein M, Hennessy M, Jordan A. How sources of sexual American Journal of Health Behavior 2009;33(1):37+. 106. Landale NS, Oropesa RS. Hispanic Families: Stability and Change. Annual Review of Sociology 2007;33(1):381 405. doi:10.1146/annurev.soc.33.040406.131655. 107. Updegraff KA, McHale SM, Whiteman SD, Thayer SM, Delgado MY. Adolescent sibling relationships in Mexican American families: Exploring the role of familism. Journal of Family Psychology 2005;19(4):512 522. doi:10.1037/0893 3200.19.4.512. 108. McHale SM, Updegraff KA, Shanahan L, Crouter AC, Killoren Differential Treatment in Mexican American Families. Journal of Marriage and Family 2005;67(5):1259 1274. doi:10.1111/j.1741 3737.2005.00215.x.

PAGE 370

354 109. Villarruel AM. Cultural influences on the sexual attitudes, beliefs, and norms of young La tina adolescents. J Specialists Pediatric Nursing 1998;3(2):69 79. doi:10.1111/j.1744 6155.1998.tb00030.x. 110. Whiteman SD, Zeiders KH, Killoren SE, Rodriguez SA, Updegraff KA. Sibling Influence on Mexican aviors: The Role of Sibling Modeling. Journal of Adolescent Health 2014;54(5):587 592. doi:10.1016/j.jadohealth.2013.10.004. 111. Averett SL, Argys LM, Rees DI. Older siblings and adolescent risky behavior: does parenting play a role? J Popul Econ 2011; 24(3):957 978. doi:10.1007/s00148 009 0276 1. 112. Argys LM, Rees DI, Averett SL, Witoonchart B. Birth Order and Risky Adolescent Behavior. Economic Inquiry 2006;44(2):215 233. doi:10.1093/ei/cbj011. 113. East PL, Khoo ST. Longitudinal pathways linking family factors and sibling relationship qualities to adolescent substance use and sexual risk behaviors. Journal of Family Psychology 2005;19(4):571 580. doi:10.1037/0893 3200.19.4.571. 114. Adler NE, Boyce T, Chesney MA, et al. Socioeconomic status and health: The challenge of the gradient. American Psychologist 1994;49(1):15 24. doi:10.1037/0003 066X.49.1.15. 115. Dehlendorf C, Marchi K, Vittinghoff E, Braveman P. Sociocultural determinants of teenage childbearing among Latinas in California. Matern C hild Health J 2010;14(2):194 201. doi:10.1007/s10995 009 0443 8. 116. Lareau A. Berkeley: University of California Press; 2011. 117. Link BG, Phelan J. Social Conditions As Fundamental Causes of Disease. Journal of Health and Social Behavior 1995;35:80 94. 118. Santelli J, Lowry R, Brener ND, Robin L. The association of sexual behaviors with socioeconomic status, family structure, and race/ethnicity among US adolescents. Am J Public Health 2000;90(10):1582 1588. 119. Waddell EN, Orr MG, Sackoff J, Santelli JS. Pregnancy risk among black, white, and Hispanic teen girls in New York City Public Schools. J Urban Health 2010;87(3):426 439. 120. Fomby P, Mollborn S, Sennott CA. Race/Ethnic differences in effects of family Journal of Marriage a nd Family 2010;72:234 253. doi:10.1111/j.1741 3737.2010.00696.x.

PAGE 371

355 121. Crissey SR. Sexual Behavior across the Border: A Comparison of Adolescents in Mexico to those of Mexican Origin in the United States. 2004. 122. Eitle TM, Wahl A MG, Aranda E. Immigra nt generation, selective acculturation, and alcohol use among Latina/o adolescents. Social Science Research 2009;38(3):732 742. doi:10.1016/j.ssresearch.2009.01.006. 123. Castro FG, Marsiglia FF, Kulis S, Kellison JG. Lifetime Segmented Assimilation Traj ectories and Health Outcomes in Latino and Other Community Residents. Am J Public Health 2010;100(4):669 676. doi:10.2105/AJPH.2009.167999. 124. Nagasawa R, Qian Z, Wong P. Theory of Segmented Assimilation and the adoption of marijuana use and delinquent behavior by Asian Pacific youth. The Sociological Quarterly 2001;42(3):351 372. 125. Portes A, ed. The New Second Generation New York: Russell Sage Foundation; 1996. 126. Mahler S, Pessar PR. Gender Matters: Ethnographers Bring Gender from the Periphe ry toward the Core of Migration Studies. International Migration Review 2006;40:27 63. doi:10.1111/j.1747 7379.2006.00002.x. 127. Connell R. Gender and Power Stanford, CA: Stanford University Press; 1987. 128. Wingood GM, DiClemente RJ. Application of the Theory of Gender and Power to Examine HIV Related Exposures, Risk Factors, and Effective Interventions for Women. Health Educ Behav 2000;27(5):539 565. doi:10.1177/109019810002700502. 129. West C, Don H. Zimmerman. Doing Gender. Gender & Society 198 7;1(2):125 151. 130. Ortner SB. Theory in Anthropology since the Sixties. Comparative Studies in Society and History 1984;26(1):126 166. 131. Beauvoir S. The Second Sex Vintage Books ed. New York: Vintage Books; 1989. 132. Butler J. Performative acts and gender constitution: An Essay in Phenomology and Feminist Theory. In: Feminst Theory Reader: Local and Global Perspectives New York and London: Routledge; 2003. 133. Carole V. Anthropology rediscovers sexuality: A theoretical comment. Social Science & Medicine 1991;33(8):875 884. doi:10.1016/0277 9536(91)90259 F. 134. Hondagneu Sotelo P. Immigration Berkeley Calif.: University of California Press; 1994.

PAGE 372

356 135. American Association of University Women. T he Simple Truth about the Gender Pay Gap. AAUW; 2014. http://www.aauw.org/files/2014/09/The Simple Truth_Fall.pdf. Accessed September 27, 2014. 136. Wall G, Arnold S. How Involved Is Involved Fathering?: An Exploration of the Contemporary Culture of Fathe rhood. Gender and Society 2007;21(4):508 527. 137. Gutmann M. Berkeley: University of California Press; 1996. 138. Hirsch JS. A Courtship After Marriage: Sexuality and Love in Mexican Transnational Families Berkeley: University of California Press; 2003. 139. Boe Latin American Perspectives 2008;35(1):16 30. doi:10.1177/0094582X07310843. 140. Parrado EA, Flippen CA. Migration and Gender among Mexican Women. American Sociological Review 2005;70(4):606 632. doi:10.1177/0003122 40507000404. 141. Cohn D, Livingston G, Wang W. After Decades of Decline, a Rise in Stay at Home Mothers Washington, D.C.: Pew Research Center; 2014:37. http://www.pewsocialtrends.org/files/2014/11/2014 11 14_remarriage final.pdf. Accessed November 18, 2 014. 142. between Latina adolescent daughters and mothers. In: Denner J, Guzmn BL, eds. Latina Girls: Voices of Adolescent Strength in the United States New York: New York Un iversity Press; 2006:29 43. 143. Denner J, Dunbar N. Negotiating Femininity: Power and Strategies of Mexican American Girls. Sex Roles 2004;50(5):301 314. doi:10.1023/B:SERS.0000018887.04206.d0. 144. Gallegos tural practices, constructing identities. In: Denner J, Guzmn BL, eds. Latina Girls: Voices of Adolescent Strength in the United States New York: New York University Press; 2006:44 58. 145. Garcia L. Respect Yourself, Protect Yourself: Latina Girls and Sexual Identity NYU Press; 2012. 146. Whitfield CL, Anda RF, Dube SR, Felitti VJ. Violent Childhood Experiences and the Risk of Intimate Partner Violence in Adults Assessment in a Large Health Maintenance Organization. J Interpers Violence 2003;18(2):16 6 185. doi:10.1177/0886260502238733.

PAGE 373

357 147. A Longitudinal Analysis. Journalism & Mass Communication Quarterly 1993;70(3):628 637. doi:10.1177/107769909307000313. 148. Schalet A. Sex, Love, and Autonomy in the Teenage Sleepover. Contexts 2010;9(3):16 21. doi:10.1525/ctx.2010.9.3.16. 149. Parrado EA, Flippen CA. Migration and Sexuality: A Comparison of Mexicans in Sending and Receiving Communities. Journal of Social Issues 2010;6 6:175 195. doi:10.1111/j.1540 4560.2009.01639.x. 150. Foner N, ed. Across Generations Immigrant Families in America New York http://site.ebrary.com/lib/cbuc/docDetail.action?docID=10313203. Accessed September 9, 2012. 151. Garca L. Love at first sex: Latina girls meanings of virginity loss and relationships. Identities 2009;16(5):601 621. doi:10.1080/10702890903172751. 152. Gilliam ML. The Role of Parents and Partners in the Pregnancy Behaviors of Young Latin as. Hispanic Journal of Behavioral Sciences 2007;29(1):50 67. doi:10.1177/0739986306295036. 153. academic success among adolescent Latinas. Environment and Planning 2000;32(4 ):635 654. 154. Culture, Health & Sexuality 2001;3(3):295 310. 155. Children of Immigrants Longitudi nal Study (CILS), 1991 2006. 2011. http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/20520. Accessed September 2, 2012. 156. Journal of Mixed Methods Research. SAGE 2012. http://www.sagepub.com/journals/Journal201775?prodId=Journal201775. Accessed May 5, 2012. 157. Leech NL, Onwuegbuzie AJ. A typology of mixed methods research designs. Quality & Quantity 2007;43:265 275. doi:10.1007/s11135 007 9105 3. 158. Teddlie C, Tashakkori A. A general typology of research designs featuring mixed methods. Research in the Schools 2006;13(1):12 28. 159. Onwuegbuzie AJ, Johnson RB. The validity issue in mixed research. Research in the Schools 2006;13(1):48 63.

PAGE 374

358 160. Harris K, Halpern C, Whitsel E, et al. Add Health: The national longitudinal study of adolescent hea lth: Research Design. 2009. http://www.cpc.unc.edu/projects/addhealth/design. 161. Sedano MV. Chicanismo: A Rhetorical Analysis of Themes and Images of Selected Poetry From the Chicano Movement. Western Journal of Speech Communication: WJSC 1980;44(3):17 7 190. 162. Agresti A, Finlay B. Statistical Methods for the Social Sciences Upper Saddle River, N.J.: Pearson Prentice Hall; 2009. 163. Upchurch DM, Lillard LA, Aneshensel CS, Li NF. Inconsistencies in Reporting the Occurrence and Timing of First Inter course among Adolescents. The Journal of Sex Research 2002;39(3):197 206. doi:10.2307/3813615. 164. Goldberg R, Adsera A, Tienda M. Gender, Nativity and Family Variations in the Timing of Sexual Initiation. In: New Orlreans, LA; 2013. 165. Jamieson LK, Wade TJ. Early Age of First Sexual Intercourse and Depressive Symptomatology among Adolescents. Journal of Sex Research 2011;48(5):450 460. doi:10.1080/00224499.2010.509892. 166. Schoen R, Landale NS, Daniels K. Family transitions in young adulthood. Dem ography 2007;44(4):807 820. 167. Myers D, Gao X, Emeka A. The gradient of immigrant age at arrival effects on socioeconomic outcomes in the U.S. International Migration Review 2009;43(1):205 229. doi:10.1111/j.1747 7379.2008.01153.x. 168. White MJ, Gli ck JE. The Academic Trajectories of Immigrant Youths: Analysis Within and Across Cohorts. Demography 2003;40(4):759 783. doi:10.1353/dem.2003.0034. 169. Rumbaut RG. Ages, Life Stages, and Generational Cohorts: Decomposing the Immigrant First and Second G enerations in the United States1. International Migration Review 2004;38(3):1160 1205. 170. Popkin BM, Udry JR. Adolescent Obesity Increases Significantly in Second and Third Generation U.S. Immigrants: The National Longitudinal Study of Adolescent Healt h. J Nutr 1998;128(4):701 706. 171. StataCorp. Stata Statistical Software: Release 11 College Station, TX: StataCorp LP; 2009. 172. Singer JD, Willett JB. Event Occurrence Oxford; New York: Oxfo rd University Press; 2003.

PAGE 375

359 173. Little RJA, Rubin DB. Statistical Analysis with Missing Data 2nd ed. Hoboken, N.J: Wiley; 2002. 174. Pew Hispanic Center. Demographic profile of Hispanics in Colorado, 2011. Pew 2014. http://www.pewhispanic.org/states/. Accessed June 8, 2014. 175. Patton M. Qualitative Research & Evaluation Methods 3rd edition, 4th printing. Thousand Oaks Calif.: Sage; 2004. 176. Maternowska C, Estrada F, Campero L, Herrera C, Brindis C, Vostr ejs MM. Gender, culture and reproductive decision making among recent Mexican migrants in California. Culture, Hlth & Sexuality 2010;12(1):29 43. doi:10.1080/13691050903108688. 177. Semhur. Map of Mexico, with states borders. 2008. 178. Castillo Mancill a J, Allshouse A, Collins C, Hastings Tolsma M, Campbell TB, MaWhinney S. Differences in Sexual Risk Behavior and HIV/AIDS Risk Factors Among Foreign Born and US Born Hispanic Women. Journal of Immigrant and Minority Health 2011;14(1):89 99. doi:10.1007/s 10903 011 9529 7. 179. Abu Lughod L. University of California Press; 2008. 180. Watson I. Life history meets economic theory: the experiences of three working class women in a local labour market. Work, Employment & Society 1993;7(3):411 435. 181. Danziger SK. Family life and teenage pregnancy in the inner city: Experiences of African American youth. Children and Youth Services Review 1995;17(1 2):183 202. doi:10.1016/0190 7409(95)00008 Z. 182. Hubbard G. The U sefulness of Indepth Life History Interviews for Exploring the Role of Social Structure and Human Agency in Youth Transitions. Sociological Research Online 2000;4. http://ideas.repec.org/a/sro/srosro/1999 84 3.html. Accessed May 6, 2012. 183. Schensul JJ LeCompte MD. SAGE; 1999. 184. Ulin PR. Reproductive Health Research Triangle Park, N.C.: Family Health International; 2002. 185. Strauss AL, Corbin J. Basics of Qualitative Research Sage Newbury Park, CA; 1990.

PAGE 376

360 186. Cubb in C, Santelli J, Brindis CD, Braveman P. Neighborhood context and sexual behaviors among adolescents: findings from the National Longitudinal Study of Adolescent Health. Perspect Sex Reprod Health 2005;37(3):125 134. doi:Article. 187. Cubbin C, Brindis CD, Jain S, Santelli J, Braveman P. Neighborhood poverty, aspirations and expectations, and initiation of sex. J Adolesc Health 2010;47(4):399 406. doi:10.1016/j.jadohealth.2010.02.010. 188. Donato KM, Gabaccia D, Holdaway J, Manalansan M, Pessar PR. A G lass Half Full? Gender in Migration Studies. International Migration Review 2006;40:3 26. doi:10.1111/j.1747 7379.2006.00001.x. 189. Berkman LF. Social epidemiology: social determinants of health in the United States: are we losing ground? Annual review of public health 2009;30:27 41. 190. Trejos Castillo E, Vazsonyi AT. Risky sexual behaviors in first and second generation Hispanic immigrant youth. J Youth Adolescence 2008;38(5):719 731. doi:10.1007/s10964 008 9369 5. 191. Manlove J, Ryan S, Franzett First Sexual Relationships. Perspectives on Sexual and Reproductive Health 2003;35(6):246 255. 192. Sandfort TG., Orr M, Hirsch JS, Santelli J. Long term health correlates of timing of sexual debut: re sults from a national US study. American journal of public health 2008;98(1):155. 193. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance 2012 Atlanta, GA: Department of Health and Human Services; 2013:174. 194. Romo Maternal Values Accompanying Contraceptive Use Advice to Young Latina Adolescent Daughters. Sex Res Soc Policy 2010;7(2):118 127. doi:10.1007/s13178 009 0001 6. 195. Gonzlez Lpez G. F athering Latina sexualities: Mexican men and the virginity of their daughters. Journal of Marriage and Family 2004;66(5):1118 1130. 196. Hardway C, Fuligni AJ. Dimensions of family connectedness among adolescents with mexican, chinese, and european backg rounds. Developmental Psychology 2006;42(6):1246 1258. doi:10.1037/0012 1649.42.6.1246. 197. Miller BC. Family Influences on Adolescent Sexual and Contraceptive Behavior. The Journal of Sex Research 2002;39(1):22 26.

PAGE 377

361 198. Glick JE, Ruf SD, White MJ, Go ldscheider F. Educational Engagement and Early Family Formation: Differences by Ethnicity and Generation. Social Forces 2006;84(3):1391 1415. 199. Sachs JD, Shatz HJ, Deardorff A, Hall RE. Trade and Jobs in U.S. Manufacturing. Brookings Papers on Economi c Activity 1994;1994(1):1 84. doi:10.2307/2534630. 200. McHale SM, Bissell J, Kim J Y. Sibling relationship, family, and genetic factors in sibling similarity in sexual risk. Journal of Family Psychology 2009;23(4):562 572. doi:10.1037/a0014982. 201. A bimanyi Ochom J. The better the worse: risk factors for HIV infection among women in Kenya and Uganda Demographic and Health Survey. AIDS Care 2011;23(12):1545 1550. doi:10.1080/09540121.2011.582477. 202. Manlove J, Ryan S, Franzetta K. Contraceptive U se and Consistency in U.S. Perspectives on Sexual and Reproductive Health 2004;36(6):265 275. 203. Minnis AM, Doherty I, vanDommelen Gonzalez E, Cheng H, Otero Sabogal R, Padian NS. Immigration and Sexual Part ner Risk Among Latino Adolescents in San Francisco. J Immigrant Minority Health 2010;12(6):900 908. doi:10.1007/s10903 010 9348 2. 204. Hirschman C. The educational enrollment of immigrant youth: A test of the segmented assimilation hypothesis. Demograph y 2001;38(3):317 336. 205. Coleman Minahan K, Krueger PM. Family structure, family socioeconomic status and sexual initiation in Mexican origin adolescents. In progress 2014. 206. Kao T Condom Use, Likelihood of Sexually Transmitted Infections. The Journal of School Nursing 2013;29(1):61 70. doi:10.1177/1059840512444134. 207. Glei DA. Measuring Contraceptive Use Patterns Among Teenage and Adult Women. Family Planning Perspectives 1999 ;31(2):73 80. doi:10.2307/2991642. 208. Relationships: The Role of Relationships, Partners, and Sexual Histories. Demography 2007;44(3):603 621. 209. Ryan S, Manlove J, Moor e KA. Science Says: The Relationship between Teenage Motherhood and Marriage The National Campaign to Prevent Teen Pregnancy; 2004. 210. Graham JW. Missing data analysis: Making it work in the real world. Annual review of psychology 2009;60:549 576.

PAGE 378

362 211 Coleman Minahan K. Immigrant generation, family structure, and contraceptive use in Mexican origin adolescents. 2014. 212. UCLA: Statistical Consulting Group. Stata FAQ: How to perform Sobel Goodman mediation tests in Stata? http://www.ats.ucla.edu/sta t/stata/faq/sgmediation.htm. Accessed October 8, 2014. 213. Driscoll AK. Community Opportunity, Perceptions of Opportunity, and the Odds of an Adolescent Birth. Youth & Society 2005;37(1):33 61. doi:10.1177/0044118X04267493. 214. Santelli J, Abraido Lan za AF, Melnikas AJ. Migration, acculturation, and sexual and reproductive health of Latino adolescents. J Adolesc Health 2009;44(1):3 4. doi:10.1016/j.jadohealth.2008.10.135. 215. Vasquez JM. Mexican Americans across Generations: Immigrant Families, Raci al Realities New York: New York University Press; 2011. 216. Bettie J. Berkeley: University of California Press; 2003. 217. Gonzlez Lpez G. Erotic Journeys Mexican Immigrants and Their Sex Lives Berkeley: University of California Press; 2005. http://0 search.ebscohost.com.s kyline.ucdenver.edu/login.aspx?direct=true&scope=site&d b=nlebk&db=nlabk&AN=131437. Accessed October 6, 2013. 218. Chavez LR, Hubbell FA, Mishra SI, Valdez RB. Undocumented Latina Immigrants in Orange County, California: A Comparative Analysis. Internation al Migration Review 1997;31(1):88 107. doi:10.2307/2547259. 219. Dreby J. Divided by Borders: Mexican Migrants and Their Children Berkeley: University of California Press; 2010. 220. Children of Global Migration: Transnational Families and Gendered Woes Stanford, Calif: Stanford University Press; 2005. 221. Suarez Orozco C, Todorova ILG, Louie J. Making Up For Lost Time: The Experience of Separation and Reunificat ion Among Immigrant Families. Family Process 2002;41(4):625 643. doi:10.1111/j.1545 5300.2002.00625.x. 222. Bartlett M, Shafer K, Seipel MMO. Association of Egalitarian Attitudes with Income, Poverty, and Entitlement Program Utilization. Journal of Human Behavior in the Social Environment 2013;23(8):918 930. doi:10.1080/10911359.2013.803453. 223. Villarruel AM, Jemmott JB, Jemmott LS, Ronis DL. Predicting Condom Use Among Sexually Experienced Latino Adolescents. West J Nurs Res 2007;29(6):724 738. doi: 10.1177/0193945907303102.

PAGE 379

363 224. Ehlers TB. Debunking Marianismo: Economic Vulnerability and Survival Strategies among Guatemalan Wives. Ethnology 1991;30(1):1 16. doi:10.2307/3773494. 225. idad. In: Denner J, Guzmn BL, eds. Latina Girls: Voices of Adolescent Strength in the United States New York: New York University Press; 2006:17 28. 226. Elliott S. Talking to Teens about Sex: Mothers Negotiate Resistance, Discomfort, and Ambivalence. S ex Res Soc Policy 2010;7(4):310 322. doi:10.1007/s13178 010 0023 0. 227. Ryan S, Franzetta K, Manlove J. Knowledge, Perceptions, and Motivations for Youth Society 2007;39(2):182 208. doi:10.11 77/0044118X06296907. 228. Else Quest NM, Hyde JS, DeLamater JD. Context Counts: Long Term Sequelae of Premarital Intercourse or Abstinence. The Journal of Sex Research 2005;42(2):102 112. 229. Foner N, Dreby J. Relations Between the Generations in Immig rant Families. Annual Review of Sociology 2011;37(1):545 564. doi:10.1146/annurev soc 081309 150030. 230. Ward LM. Understanding the role of entertainment media in the sexual socialization of American youth: A review of empirical research. Developmental Review 2003;23(3):347 388. doi:10.1016/S0273 2297(03)00013 3. 231. U A, Stueve A. Intimate Partner Violence Among Economically Disadvantaged Young Adult Women: Associations With Adolescent Risk Taking and Pregnancy Experiences. Perspectives on Sexual and Reproductive Health 2009;41(2):84 91. doi:10.1363/4108409. 232. Lopez B. Personal communication. 2013. 233. Pew Research Center. Second Generation Americans A Portrait of the Adult Children of Immigrants Washing ton, D.C.: Pew Research Center; 2013. 234. Lee D, Perreira KM, Harris KM. Making It in America: High School Completion by Immigrant and Native Youth. Demography 2006;43(3):511 536. doi:10.1353/dem.2006.0026. 235. Gibson MA. Accommodation without Assimil ation: Sikh Immigrants in an American High School Ithaca: Cornell University Press; 1988. 236. Romo LF, Kouyoumdjian C, Nadeem E, Sigman M. Promoting values of education in Latino mother adolescent discussions about conflict andsexuality. In:

PAGE 380

364 Denner J, G uzmn BL, eds. Latina Girls: Voices of Adolescent Strength in the United States New York: New York University Press; 2006:59 76. 237. Swidler A. Culture in Action: Symbols and Strategies. American Sociological Review 1986;51(2):273 286. doi:10.2307/2095 521. 238. Achievement Gaps: How Hispanic and White Students in Public Schools Perform in Mathematics and Reading on the National Assessment of Educational Progress: Highlights Washington, D.C.: U.S. Department of Education 239. Ross T, Kena G, Rathbun A et al. Higher Education: Gaps in Access and Persistence Study. Statistical Analysis Report. NCES 2012 046. U.S. Department of Education; 2012. http://eric.ed.gov/?id=ED534691. Accessed September 11, 2014. 240. Goldenberg C, Reese L, Gallimore R. Effects of Literacy Materials from School on American Journal of Education 1992;100(4):497 536. 241. Lee J S, Bowen NK. Parent Involvement, Cultural Capital, and the Achievement Gap among Element ary School Children. American Educational Research Journal 2006;43(2):193 218. 242. Hart B, Risley TR. The early catastrophe: The 30 million word gap by age 3. American Educator 2003;27(1):4 9. 243. Levitt P. The Transnational Villagers Berkeley: Univ ersity of California Press; 2001. 244. Dreby J, Stutz L. Making something of the sacrifice: gender, migration and Global Networks 2012;12(1):71 90. doi:10.1111/j.1471 0374.2011.00337.x. 245. Blumer H. Symbolic Interactionism: Perspective and Method University of California Press; 1986. 246. Singer M. Beyond the ivory tower: Critical praxis in medical anthropology. In: Understanding and Applying Medical Anthropology California: Mayfield Publishing Company; 19 98:225 239. 247. relationship to sexual outcomes. Journal of Adolescent Health 2000;26(4):268 278. doi:10.1016/S1054 139X(99)00096 8. 248. Jaccard J, Dittus PJ, Gordon VV. Ma ternal Correlates of Adolescent Sexual and Contraceptive Behavior. Family Planning Perspectives 1996;28(4):159 185. doi:10.2307/2136192.

PAGE 381

365 249. Kim C, Gebremariam A, Iwashyna TJ, Dalton VK, Lee JM. Longitudinal influences of friends and parents upon unprot ected vaginal intercourse in adolescents. Contraception 2011;83(2):138 144. doi:10.1016/j.contraception.2010.06.019. 250. Driscoll AK, Russell ST, Crockett LJ. Parenting Styles and Youth Well Being Across Immigrant Generations. Journal of Family Issues 2008;29(2):185 209. doi:10.1177/0192513X07307843. 251. Child Interactions in the Racial Classification of Adolescents. Sociological Forum 2009;24(3):658 688. 252. Harker K. Imm igrant Generation, Assimilation, and Adolescent Psychological Well Being. Social Forces 2001;79(3):969 1004. 253. Killoren SE, Updegraff KA, Christopher FS, Umaa Taylor AJ. Mothers, Fathers, Peers, and Mexican Jour nal of Marriage and Family 2011;73(1):209 220. doi:10.1111/j.1741 3737.2010.00799.x. 254. Crockett LJ, Brown J, Russell ST, Shen Y L. The Meaning of Good Parent Child Relationships for Mexican American Adolescents. Journal of Research on Adolescence 200 7;17(4):639 668. doi:10.1111/j.1532 7795.2007.00539.x. 255. Khurana A, Cooksey EC. Examining the Effect of Maternal Sexual Adolescent Risky Sexual Involvement. Journal of Adolescent Hea lth 2012;51(6):557 565. doi:10.1016/j.jadohealth.2012.03.007. 256. Pew Research Center. Abortion Viewed in Moral Terms: Fewer See Stem Cell Research and IVF as Moral Issues Religion & Public Life Project; 2013. ht tp://www.pewforum.org/2013/08/15/abortion viewed in moral terms/. Accessed December 9, 2014. 257. Haurin RJ, Mott FL. Adolescent Sexual Activity in the Family Context: The Impact of Older Siblings. Demography 1990;27(4):537 557. doi:10.2307/2061569. 258. Killoren SE, Roach AL. Sibling Conversations About Dating and Sexuality: Sisters as Confidants, Sources of Support, and Mentors. Fam Relat 2014;63(2):232 243. doi:10.1111/fare.12057. 259. Fine M, McClelland SI. Sexuality education and desire: Still mis sing after all these years. Harvard Educational Review 2006;76(3):297 338.

PAGE 382

366 260. Executive Actions on Immigration | USCIS. Official Website of te Department of Homeland Security 2014. http://www.uscis.gov/immigrationaction. Accessed December 18, 2014. 26 1. Fields J, Tolman DL. Risky business: Sexuality education and research in U.S. schools. Sex Res Soc Policy 2006;3(4):63 76. doi:10.1525/srsp.2006.3.4.63. 262. Portes A. Migration, Development, and Segmented Assimilation: A Conceptual Review of the Evi dence. The ANNALS of the American Academy of Political and Social Science 2007;610(1):73 97. doi:10.1177/0002716206296779. 263. Gans HJ. Comment: Ethnic Invention and Acculturation, a Bumpy Line Approach. Journal of American Ethnic History 1992;12(1):42 52. 264. Steele CM, Aronson J. Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology 1995;69(5):797 811. doi:10.1037/0022 3514.69.5.797. 265. Gonzales PM, Blanton H, Williams KJ. The Effects of Stereotype Threat and Double Minority Status on the Test Performance of Latino Women. Pers Soc Psychol Bull 2002;28(5):659 670. doi:10.1177/0146167202288010.

PAGE 383

367 APPENDIX A. APPENDIX FOR CHAPTER 2 Table A.1 : Discrepanci es In Reported Age at First Sex Discrepancy >=1 year Discrepancy >=2 years Discrepancy >=3 years Age at Wave I 1.23*** 1.26*** 1.21* (1.10 1.37) (1.11 1.42) (1.04 1.40) Male 2.03*** 2.42*** 2.06** (1.46 2.82) (1.66 3.53) (1.30 3.27) First gen 0.61 0.74 0.65 (0.28 1.31) (0.31 1.74) (0.24 1.79) Second gen 0.93 0.86 0.89 (0.60 1.44) (0.52 1.43) (0.47 1.68) Third gen Ref. Ref. Ref. Less than a high school degree 0.59* 0.69 1.02 (0.36 0.99) (0.38 1.24) (0.52 1.98) Lives with two biological parents 0.80 0.80 0.82 (0.53 1.21) (0.51 1.27) (0.47 1.44) Number of household members 1.07 1.08 0.98 (0.96 1.18) (0.97 1.21) (0.86 1.13) Speaks English at home 0.91 0.76 0.36** (0.58 1.44) (0.45 1.27) (0.19 0.68) Mom's age at participant's birth 0.97* 0.97 0.98 (0.93 1.00) (0.93 1.01) (0.94 1.03) Family income 1.47** 1.58* 1.37 (1.10 1.95) (1.11 2.24) (0.92 2.05) Dad's years of education 0.99 0.99 1.02 (0.92 1.07) (0.90 1.08) (0.92 1.13) Mom's years of education 0.97 0.93 0.94 (0.90 1.06) (0.85 1.02) (0.84 1.05) Father not working 2.19* 2.77** 2.78* (1.12 4.29) (1.31 5.87) (1.08 7.12) Unskilled father 1.08 1.29 1.56 (0.64 1.81) (0.70 2.38) (0.69 3.56) Skilled father 1.50 2.00* 2.61* (0.89 2.51) (1.09 3.65) (1.17 5.81) Professional father Ref. Ref. Ref. Mother not working 1.31 0.80 1.10 (0.77 2.23) (0.43 1.48) (0.49 2.48)

PAGE 384

368 Table A.1 : Discrepanci es In Reported Age at First Sex Unskilled mother 1.54 0.86 0.90 (0.86 2.78) ( 0.44 1.69) (0.38 2.17) Skilled mother 1.81* 1.71~ 2.43* (1.04 3.14) (0.93 3.14) (1.11 5.30) Professional mother Ref. Ref. Ref. Never married Age at first sex Constant 0.01** 0.00*** 0.01** (0.00 0.19) (0.00 0.10) (0.00 0.21) Observations 649 649 649 *** p<0.001, ** p<0.01, p<0.05, ~ p<0.10 Table A.2 : Discrepancies in R eported A ge at First B irth Discrepancy < 6 months Discrepancy >=1 year Discrepancy >1 year Discrepancy >2 year Age at Wave I 1.27 0 .99 0.80 1.10 (0.90 1.80) (0.69 1.42) (0.50 1.29) (0.57 2.13) Male 0.20** 1.96 1.04 0.39 (0.06 0.65) (0.58 6.66) (0.18 5.93) (0.03 4.51) First gen 1.02 Did not converge Did not converge Did not converge (0.16 6.56) Second gen 0.31 1.36 2.41 4.31 (0.08 1.30) (0.29 6.40) (0.23 25.33) (0.21 89.38) Third gen Ref. Ref. Ref. Ref. Less than a high school degree 1.14 0.53 0.60 1.46 (0.23 5.54) (0.05 6.20) (0.04 7.93) (0.08 26.50) Lives with two biological p arents 1.47 0.68 0.63 0.99 (0.44 4.95) (0.18 2.58) (0.11 3.56) (0.08 11.53) Number of household members 1.16 1.39* 1.74* 2.26* (0.87 1.54) (1.03 1.86) (1.14 2.66) (1.14 4.49) Speaks English at home 0.33 0.92 0.80 0.54 (0.08 1.31) (0.18 4.78) (0.08 7.64) (0.03 11.63) Mom's age at participant's birth 0.99 0.91 0.75* 0.73~ (0.90 1.08) (0.79 1.04) (0.59 0.96) (0.50 1.05) Family income 1.33 1.32 1.17 2.01 (0.53 3.33) (0.48 3.62) (0.31 4.46) (0.29 13.99) Dad' s years of education 0.93 1.03 0.78 0.86

PAGE 385

369 Table A.2 : Discrepancies in R eported A ge at First B irth (0.72 1.21) (0.77 1.38) (0.51 1.21) (0.49 1.50) Mom's years of education 1.01 0.98 1.33 2.07* (0.80 1.28) (0.74 1.31) (0.85 2.07) (1.02 4.22) Father not working 0.85 0.92 0.57 0.65 (0.13 5. 49) (0.09 9.66) (0.02 16.24) (0.03 16.00) Unskilled father 0.79 0.74 0.44 0.21 (0.17 3.70) (0.11 4.99) (0.03 6.36) (0.01 4.49) Skilled father 0.45 0.76 0.85 Did not converge (0.08 2.52) (0.11 4.98) (0.06 11.95) Professional fath er Ref. Ref. Ref. Ref. Mother not working 2.39 1.26 4.90 18.92 (0.22 25.47) (0.17 9.56) (0.18 130.18) (0.27 1,334.19) Unskilled mother 4.87 1.56 8.16 6.15 (0.48 49.48) (0.19 12.61) (0.27 249.52) (0.07 535.92) Skilled mother 2.4 1 2.30 9.12 8.04 (0.22 26.28) (0.32 16.54) (0.33 250.55) (0.11 607.78) Professional mother Ref. Ref. Ref. Ref. Never married 1.81 0.54 0.65 0.28 (0.58 5.61) (0.10 2.88) (0.09 4.71) (0.02 4.75) Age at first sex 0.99 0.86 0.66~ 0 .60 (0.80 1.22) (0.65 1.14) (0.41 1.05) (0.33 1.10) Constant 0.00 0.31 2,129.20 0.00 (0.00 10.80) (0.00 7,443.87) (0.00 5.11e+09) (0.00 9,893,299.82) Observations 333 307 307 263 *** p<0.001, ** p<0.01, p<0.05, ~ p<0.10 Table A .3 : Differences in C haracteristics by I mmigrant G eneration. First 1.5 Second Third Speaks English at home 89.90% 89.20% 94.40% 99.70% Bilingual 78% 70.50% 62% 7.20% Born a U.S. Cit i zen 1.80% 10% 99.70% 99.80% Lives with two biological parents 50.60% 57.50% 75.36% 45.06% Family income $15,233 $19,426 $25,859 $38,938 Mother's education 8.69 9.24 9.85 12.9 Father's education 9.29 9.9 10.07 12.35 Mother's employment 1.47 1.57 1.83 2.41 Father's employment 2.11 2.16 2.26 2.48 Age at first sex 19.62 1 7.82 18.67 18.69 Age at first birth 22.97 22.81 22.86 22.32

PAGE 386

370 Table A 4: Family Structure of First Generation Qualitative P articipants ID Age Origin Age at migration Two biological parents Grandparents Siblings 11 36 Chihuahua (small town) 17 no. Sing le mom w/migrant dad until 8yo, then two bio no 4 older brothers, 3 older sisters 12 27 Ciudad Juarez (rancho) 18 yes no One older sister, one younger brother, young sister 21 33 Chihuahua 6 mo no single mom until passed away then aunts no Middle, older sister, young brother 13 30 Ciudad Juarez (more urban) 18 yes no 2nd to last of six, 4 girls, 2 boys 14 34 Jalisco (small town) 15 no grandma and uncle yes 3rd of five, 2 boys, 3 girls 15 34 Ciudad Juarez (more urban) 19 yes no 3rd of six, 4 girls, 2 boys 16 39 Juarez (more urban) 10 No at 10 then moved with MOC to US and FOC visited no Four older sibs boys and girls 17 29 Jalisco (mom Zacatecas, dad Jalisco) 3 yes until FOC passed away 16yo no Older sister, younger brother, younger sister 18 31 Ciu dad Juarez (more urban) 9 No at 9 then MOC moved to US and remarried yes at various times Older brother, younger sister, two young half sisters

PAGE 387

371 Table A 4: Family Structure of First Generation Qualitative P articipants 19 31 Pueblo in Chihuahua 16 no, FOC moved to US when she was 4yo no Older sis, younger sis, younger bro (never lived w/youngest bro) 110 38 Chihuahua (urban) 10 No 8 or 9 yo when MOC left FOC for a year when MOC in US Youngest of five, all older brothers Table A 5: Family Structure of Second Generation Qualitative P articipants ID Age Childhood Origin migration Two parents Grandparents Siblings 22 28 Texas until 7yo then SW Denver FOC Coahuila MOC Aguascalientes yes no Two older brothers, one older sister 23 28 New Mexico until 2012 (small farming town) Chihuahua early 20s, settled early 30s yes no One older sister 24 27 Denver MOC Durango to Denver at 15 FOC born Texas, grew up in Juarez yes no Two older brothers 25 32 Denver then Commerce City Parents married in Juarez (urban), moved to US 19yo yes no Older sister, old brother 26 31 Jefferson County then SW Denver MOC migrated Chihuahua small town, bio dad Mexico no A few years only during childhood Older brother, young brother, younger sister 27 27 metro Denver MOC and FOC migrated Chihuahua young adults no no 3 younger sibs 28 37 Denver the n Commerce City Parents married in Juarez (urban), moved to US 19yo yes no Younger brother and sister 29 36 Denver Parents migrated from Juarez MOC 30 FOC 50 (MOC Michoacan FOC Durango) yes until FOC died when she was 15 no Older brother

PAGE 388

372 Table A 5: Family Structure of Second Generation Qualitative P articipants 210 34 Denver M OC from Denver, FOC Jalisco (rancho) 17yo Yes until she was 21 no Older brother and younger brother 211 27 Greely, moved around Denver at 15 yo POC from Chihuahua (rural) moved as adults No, 11yo when FOC deported no Older sister, two younger brothers Table A 6: Family SES of First Generation Qualitative P articipants ID employment 11 9th mail lady, store owner work when dad home 4th migrant worker, store owner. In US: construction for city 12 primary, enrolled secondary home but sold things at market primary, enrolled secondary police 21 unsure home, part time unskilled jobs unsure unsure 13 none home secondary butcher 14 secondary (carrera) secretary second ary migrant worker, cutting pine 15 elementary home secondary butcher 16 primary homemaker until moved to US and housekeeper primary bartender 17 primary homemaker and then housekeeper primary construction 18 primary maquiladora finished hs and som e college in Mexico maquiladora 19 primary fabrica secondary unsure 110 secondary and carrera (nurse cosmotology, baking) stand selling things at club in Mexico, hsd (stepdad hsd too) military and then worked at airport (stepdad trains) Table A 7: Fa mily SES of Second Generation Qualitative P articipants ID employment 22 home BS in ministry minister no degree until after kids grew, went back for hsd and ASD 23 cleaning 8th welding 3r d 24 9th home secondary. Then associates through rtd rtd technician

PAGE 389

373 Table A 7: Fa mily SES of Second Generation Qualitative P articipants 25 maybe 8th or 9th home hsd in mex sold things at flea market. self employed locksmith 26 3rd CNA in US, then janitor and lunch lady step dad: hsd in mex cook then janitor 27 3rd clea ning hotels most of childhood then CNA 6th construction 28 8th or 9th home hsd in mex meat packing. sold things at flea market. self employed locksmith 29 6th then Bible school home until FOC sick then packaged food for Airlines no school ever, illiter ate taco stand in LA then dishwasher in Denver 210 early hs, no degree taco bell, linen factory then home with kids did not go to school warehouse, staircases, then pastor 211 secundaria beef plant, agriculture secundaria beef plant, agriculture Tabl e A 8: Outcomes of First Generation Qualitative Participants ID Highest level of education Current employment First sex BC first sex First birth Gravida, parity current contraception 11 9th grade in Mexico dropped out in US, GED & MA cert Medical Assista nt 17 yes 18 G5P4 / BTL/ family complete 12 3rd year college prep, GED classes on weekends Stay at home Mom 18 no 19 G2P3 / IUD / family complete 21 completed 9th, GED classes Works at Dry Cleaners 14 no? 17 G4P4 / BTL / family complete 13 secondary (b eauty school finished), taking English and GED classes Stay at home Mom 18 yes 19 G3P3 / actively trying 14 8th Stay at home Mom 15 no 16 G5P5 / vasectomy / family complete 15 last year of prepa / English class Stay at home Mom 18 yes 20 G3P3 / IUD /fami ly complete 16 hsd and currently in college for BA Receptionist for non profit organization 17 no 20 G3P3 / OCPS / done 17 BA political science, minor Spanish Youth development at non profit 19 yes na G0 / nothing

PAGE 390

374 Tabl e A 8: Outcomes of First Generation Qualitative Participants 18 BA and currently working on MS glob al affairs full time college admissions 16 no (planb) na G0 / condoms 19 2nd year of secondary Cajera/chashier 17 yes 17 G2P2 / OCPs /family complete 110 GED and MA cert Medical assistant 18 (1 wk bf 19) no 20 G4P4 / IUD/family complete Table A 9: O u tcomes of Second Generation Qualitative Participants ID Highest level of education Current employment First sex BC first sex First birth Gravida, parity current contraception 22 BA Spanish and non profit administration, thinking about grad school Teache r/ degree in Spanish and non profit management 23 yes 28 G4P2/ BTL/family complete 23 MA Spanish and linguistics, BA Spanish Psychology Research assistant 19 yes na G0 /implant 24 BA Psychology minor biology Research assistant 20 yes 27 G1P1 / condoms 2 5 HSD & associates childhood edu & CAN Admin assistant 17 no 28 G2P2 /NFP 26 GED some college Admin assistant 14 no 16 G5P4/ BTL/ family complete 27 HSD and some college Student worker 16 yes 17 G3P3 /IUD 28 BA Business management Recently lost job at non profit 17 yes 30 G22012 /condoms/ family complete 29 CNA; medical assistant; is 2 1/2 yrs into MSW program Clinical assistant 18 no 24 G4P3 No current partner 210 HSD Receptionist 18 yes na G0/ hysterectomy d/t cancer 211 GED, medical assistant M edical assistant 14 yes 17 G3P2/ IUD Key: IUD: intrauterine device, BTL: bilateral tubal ligation, NFP: natural family planning, OCPS: oral contraceptive pills

PAGE 391

375 Table A 10 Qualitative Interview Guide most recently approved COMIRB version 11/1/13 Please tell me about your childhood and your family. school? For meals? Before school? family ie grandparent s. parents, siblings. parents? o did the chores in the house? Cooking, cleaning? celeb ration when you were a child. brothers? u? in school? up? you grew up ? brothers? not help you) achieve your goals? you? brothers? of values or lessons did you learn from your parents?

PAGE 392

376 Table A 10 Qualitative Interview Guide most recently approved COMIRB version 11/1/13 migration was described) grate? that changed? experience. family members if differ ent country of origin. the US? migrating to the U.S.? the U.S.? our life be different now if you had stayed in Mexico? *If Second Generation: migration to you? id they cross the border? who are migrants? Hispanic? r time? Please tell me about your experience with dating or relationships with boys/men as a teenager? id you do/ where did you go when dating? dating? (ie marriage, babies, school, c areer)?

PAGE 393

377 Table A 10 Qualitative Interview Guide most recently approved COMIRB version 11/1/13 money, etc. like? Sex, pregnancy, birth contro l, parenting What did you learn about sex? ? n about birth control? control? control? past? you want? want? Pregnancy #1 your next child? If any. Pregnancy #2 pregnancy?

PAGE 394

378 Table A 10 Qualitative Interview Guide most recently approved COMIRB version 11/1/13 en did you want your next child? If any. Pregnancy #3 pregnancy? birth? any. Pregnancy #4 pregnancy? you at the birth? any. future? want them to learn? u or will you talk to your children more likely to have a baby as a teenager than white /non Latina teenagers. What do you think about that? do you think many Latina adolescents are giving birth? (community, laws, politics, etc) your life right now? (is this where you want to be?) bled you to reach your goals or get where you are in life?

PAGE 395

379 Tables A 11 and A 12 examples of qualitative data displays for analysis Table A 11 Display of Gendered Division of Labor and Power among Qualitative P articipants ID GDL Differential tre atment sib GDP Differential treatment sib 11 VERY dad worked, mom home, dad did nothing in house. Mom took on male role and worked when dad in US (negative tone) MIGRATION: GDL and GDP improved w/mom in US, no more DV and mom didn't make tortillas YES, a finger in house (bothered her) EXTREME Dad violent and powerful against mom and kids, cheated on mom a lot, dad controlled mom and kids YES, GDP: could punish & control sisters; more material goods & allowance; stay out late; alcoho l; car; no curfew or rules; recognition for sports (really bothered her) Older sibs in general had power over her 12 VERY dad worked, mom home, dad did nothing to help in home (negative tone) Didn't have to do chores to earn going outside to play Dad p ower, things done his way (she excuses b/c he was raised that way) YES, brother received more materialistic stuff; parents "loved" him more (seemed to bother her) 21 grew up with single mom then single aunts who didn't marry until she was teenager but did say she knew men can't cook or wash dishes She and sister had more domestic responsibilities than brother but he worked outside home at 14 Knew that men and women were treated different and were separated Discipline and rules same for brother and sister 13 MODERATE dad worked, mom home, mom did most household work but dad helped (positive tone) Rules and discipline same for sibs Mom made decisions, sometimes divided money Rules and discipline same for sibs 14 MODERATE biomom worked but in home, grandma d id everything and uncle was sole provider (neutral) no sibs in home Uncle not grandma disciplined her more, grandma warned her not to let anyone hit her because she was hit by her husband no sibs in home but says uncle also watched out for brothers 15 MOD ERATE dad worked, mom home, mom did most household work but dad helped (positive tone) Rules and discipline same for sibs Rules and treatment same for brothers and sisters 16 VERY MODERATE, dad worked and mom home and dad didn't help much in home other t han to fix things in Mexico BUT mom worked once moved to US (negative tone) MIGRATION: Mom worked outside home Brothers didn't do much in house to help (didn't seem to bother her, brothers only) Mom moved to US against it but come; Mom hx of DV in previous marriage Didn't feel she was treated differently than brothers but mom was supposed to be and take care of her and dad spent time with brothers, not her 17 MODERATE both parents worked at times, mostly mom home. Dad did help w/cooking if mom wasn't available (positive tone) Brothers outside work, girls inside work, crocheting. Brother never washed a dish. (didn't mind) great marriage not a lot about decision making, migration decision was probably dads

PAGE 396

380 Table A 11 Display of Gendered Division of Labor and Power among Qualitative P articipants 18 MODERATE both pa rents worked when in Mexico, single mom in US when young She had primary responsibility for siblings and household. Brother assisted mom w/finances and decision making Dad made financial decisions and had more control "what he said goes" Mom more aggress ive than dad w/discipline Dad won't talk to her about politics or anything real. Championed education for brother more than sister. Mom didn't want her to move out until married though brother did 19 grew up with single mom She and sister had more respon sbilities in home than brother (seemed slightly annoyed but not much) Brothers very young, treated differently 110 Moderate both parents worked, mom did everything in home though dad fixed things (neutral tone), More chores for h er than brother (seemed slightly annoyed) Biodad powerful, violent, and mean Stepdad, loving Couldn't be in park after dark w/out her brother 22 MODERATE, dad pastor, mom worked on and off, mom most inside but dad made breakfast though mom wished he had helped more (negative tone) Boys outside chores, girls inside chores (didn't mind it) Dad power her actions reflected on dad parents more leniant w/brothers when they didn't follow rules;higher education expectations for her than brother; couldn't be home alone unless adult or brother was home. 23 MODERATE, both worked but mom took care of inside, dad would help when she wasn't home although sounds like mom wanted it that way (neutral) Sister only Dad head of household and mom does everything for him, but she still get what she wants example of women getting power within traditional roles Sister only 24 MODERATE, dad worked, mom home, mom inside, dad outside, did NOT argue about things in house not getting done Boys outside chores like oil chan ge but did laundry and iron. She actually had less responsibility in general b/c she was the baby Dad power, discussed stuff but dad had more of his way. dad strict with her and she was scared of him but she excelled b/c of it brothers controlled her datin g and interaction with boys and disciplined her. Brothers told to care of her because "she's a girl" Believes if she was a teen parent, parents would be more disappointed that she had sex but would have helped her raise kid more than they did w/brother. Sh e learned to be "good" after watching brothers struggle with dad and be punished. 25 MODERATE, dad worked, mom home. Mom inside work, dad outside and fixing things (neutral) "some machismo" in dad Girls inside work. She wasn't taught to fix things like b rother and wasn't allowed to help dad even though she wanted to. (bothered her) mom dependent on dad and when he died she had to learn to do things More dating reg and sex talk for girls. Brother played sports, she couldn't. brother treated differently as a boy, separated, more time spent w/him

PAGE 397

381 Table A 11 Display of Gendered Division of Labor and Power among Qualitative P articipants 26 VERY, both worked but mom did everything at home, stepdad refused to help and said his work is outside although mom does both (negative tone) Brothers not around a lot may have helped but mostly she did housework and cared for younger sister. Told women stay inside and take care of kids nad house (bothered her) EXTREME Dad powerful and violent. Dad made decisions growing up. Mom has more say now that she works too. It was dad's decision to kick her out of house w hen pregnant Brothers could be outside. No curfew for them. Women don't get tattoos. 27 VERY Dad only in picture for a while. Mom worked when not with dad. Mom did everything in house (negative tone) She had responsibility for house and younger sibs. Br other could go out and play. (bothered her) Dad power violent, kept mom from working Couldn't play with boys but brother could play with girls. Brother with dad, girls with mom. Brother could play and go out, she had to do housework 28 Moderate, dad work ed, mom home, mom inside work, dad helped sometimes (neutral tone) Brother didn't help as much around house (didn't bother her as much as favoritism) Brother allowed to have girls in room 29 Moderate equal Dad worked, mom home until dad got sick, then complete role reversal Brother didn't help around house and could do other things, mom wanted her to serve him (angered her and she refused) Equal decision making power, great relationship, both supported eachother 100% Brother allowed to go out and she co uldn't go out unless he went too 210 Moderate dad worked, were at home, although dad taught mom to cook (probably b/c he was migrant prior to their marriage and older) Had more responsibilities in house than brothers (preferred it th at way) Feels parents would compromise; though says dad could tell mom she couldn't go out or needed to take the kids. Power dad, wouldn't let her and husband sleep in same bed AFTER being married She couldn't leave house w/out brother per dad 211 Moderat e both parents worked, mom did everything in home, dad fixed things She and sister had more chores (neutral tone) Biodad more say than mom (neutral tone); Violent jealous controlling stepdad Table A. 12 Display of Migration Patterns, Family Structure, a nd Educational D iscourses among Qualitative P articipants. ID Type of migration Migration struggle Settlement struggle Two bio par Mom home Edu action Sacrifice Parents want more 11 Family stage No no/yes (at age add health) 12 Family unit (self & husband) Yes yes 21 Independent (mom took kids alone) migration struggle No no

PAGE 398

382 Table A. 12 Display of Migration Patterns, Family Structure, a nd Educational D iscourses among Qualitative P articipants. 13 Family stage No (only d/t migration, not divorce) yes 14 Independent (dad migrant worker, did not reunite) No no 15 Family stage Yes yes 16 Fa mily stage (migrated w/mom, left dad in Mex) no migration struggle ?settlement struggle No yes/no (by add health age) 17 Family stage migration struggle settlement struggle Yes (died) no yes yes 18 Family stage (mom migrated independentl y prior) mig ration struggle settlement struggle No no yes yes 19 Independent (went to live w/dad) No no yes 110 Family stage (mom left dad in Mex) migration struggle settlement struggle No no 22 Family stage (mom) Independent (dad) no migr ation struggle settlement struggle (mom only?) Yes yes yes 23 Family stage (both migrated independentl y prior) migration struggle settlement struggle Yes yes yes yes yes 24 Independent (mom) ?migration struggle for mom settlement struggle Yes y es yes yes 25 Family unit no migration struggle Yes yes yes yes 26 Independent (mom) migration struggle settlement struggle No no yes 27 Independent (mom) migration struggle settlement struggle No no 28 Family unit no migration s truggle Yes yes yes yes yes 29 Family stage (dad migrated indpendentl y prior) migration struggle settlement struggle Yes (died) yes/no (worked at add health age) yes

PAGE 399

383 Table A. 12 Display of Migration Patterns, Family Structure, a nd Educational D iscourses among Qualitative P articipants. 210 Independent (dad) migration struggle Yes yes yes 211 Family unit no migration struggle settlement struggle No no yes B. APPENDIX FOR CHAPTER 3 B1. Robustness summary for age at first sex models Residuals for the sexual initiation models were acceptable with a deviance of less than 2.5 except those with very early or very late ages of first sex (>25 or <=12) and a sum of squares deviance of less than 10 for all participants except for those with very late ages at first sex, over age 25. Residuals when dropping all participants who had sex within three years of Wave I were higher for the sum of squares deviance but lower for the general deviance, because most of the very early ages at first sex were dropped. Sixteen participants who had reporting having had sexual intercourse in at least one wave were missing on age at first sex. I was able to narrow the timing for first sex between waves they reported never having sex and the first wave they reported having sex. Eight of the sixteen had sex between either Wave I and Wave II or Wave II and Wave III and the mean age betwe en those two waves were used. For the other eight participants with longer intervals between known waves, I used mean matching for other participants with similar gender, immigrant generation, and age. A sensitivity analysis showed no increase in residuals using the manually imputed ages at first sex. Table B. 1 Hazards and Odds of Reproductive Health Outcomes by Decimal G enerations Hazard of sexual initiation Odds of contraceptive use at first sex Hazard of an adolescent birth Generation 1.25 (migrated age 13 17) 0.52*** 0.63 0.38*** (0.37 0.72) (0.26 1.50) (0.23 0.63) Generation 1.5 (migrated age 6 12.9) 0.63*** 0.41* (0.50 0.78) (0.19 1.10) (0.20 0.83) Generation 1.75 (migrated age before 6) 1.01 0.41* 1.01 (0.58 1.79) (0.18 0.91) (0.43 2.37) Generation 2 (U.S. born, both parents Mexican born) 0.58*** 0.9 0.43** (0.47 0.73) (0.58 1.40) (0.27 0.71) Generation 2.5 (U.S. born, one Mexican born parent, one U.S. born) 0.94 0.62

PAGE 400

384 Table B. 1 Hazards and Odds of Reproductive Health Outcomes by Decimal G enerations (0.74 1.18) (0.34 1.03) (0.35 1.10) Generation 3 (U.S. born to two U.S. born parents) Ref. Ref. Ref. Age 15.84*** Wave I age 0.91 Age 1.06*** (8.93 28.09) (0.81 1.02) (1.05 1.07) Age x age 0.92*** Age at first sex 1.09~ Age spline 0.89*** (0.91 0.94) (0.99 1.19) (0.85 0.93) Male 1.30** Male 1.03 Male 7.74*** (1.09 1.56) (0.75 1.41) (4.15 14.45) Marri ed 0.55* (0.35 0.86) Constant 0.00*** Constant 2.46 Constant 0.00*** (0.00 0.00) (0.28 21.68) (0.00 0.00) Person years 10,502 Observations 1,416 Person months 143,873 *** p<0.001, ** p<0.01, p<0.05, p<0.10 B.2 10 List wise delet ion, imputed predictors, and fully imputed models for contraceptive use at first sex. Contraceptive use and family structure models Family structure model differences (Models B.2, B.3, B.4) Listwise deletion models and imputed predictor models showed liv ing with two biological parents was less protective in second generation adolescents (OR 1.4) as compared to third generation adolescents (OR 2.64) although the significance was weak p<0.10). Table B.2 : Odds R atios (95% confidence intervals) for Contrace ptive Use at First S ex an d Family Structure; List wise deletion, no imputation Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 First generation 0.50*** 0.48*** 0.54** 0.49*** 0.49*** 0.50*** 0.49* (0.33 0.74) (0.32 0.72) (0.33 0.90) (0. 33 0.73) (0.32 0.74) (0.33 0.76) (0.22 1.06) Second generation 0.64* 0.53** 0.73 0.52** 0.50** 0.67 0.41** (0.40 1.02) (0.31 0.88) (0.41 1.29) (0.31 0.87) (0.29 0.86) (0.42 1.09) (0.18 0.90) Third generation Ref. Ref. Ref. Ref. Re f. Ref. Ref. Two biological parents 2.00*** 2.64*** 2.03*** 2.03***

PAGE 401

385 Table B.2 : Odds R atios (95% confidence intervals) for Contrace ptive Use at First S ex an d Family Structure; List wise deletion, no imputation (1.52 2.64) (1.67 4.18) (1.55 2.67) (1.55 2.67) First gen X two biological parents 0.72 (0.34 1.54) Second gen X two bio parents 0.53* (0.27 1.06) Third gen X two bio parents 1.00 (1.00 1.00) Grandparents 1.98* 1.49 (0.92 4.24) (0.63 3.52) First gen X grandparents 0.79 (0.06 11.27) Second gen X grandparents 2.09 (0.44 10.02) Third gen X grandparents 1.00 (1.00 1.00) Older sibling in the home 0.99 0.73 (0.71 1.38) (0.47 1.16) First gen X older sibling 1.10 (0.43 2.79) Second gen X older sibling 2.28* ( 0.99 5.27) Third gen X older sibling 1.00 (1.00 1.00) Wave I Age 0.96 0.99 0.99 0.99 0.99 0.96 0.96 (0.87 1.06) (0.89 1.10) (0.89 1.10) (0.89 1.10) (0.89 1.10) (0.86 1.07) (0.86 1.06) Age at first sex 1.07* 1.04 1.04 1. 04 1.04 1.04 1.05 (0.99 1.16) (0.96 1.13) (0.96 1.13) (0.96 1.13) (0.96 1.13) (0.96 1.13) (0.96 1.14) Male 0.88 0.84 0.85 0.85 0.85 0.86 0.88 (0.66 1.17) (0.62 1.14) (0.62 1.14) (0.63 1.16) (0.62 1.15) (0.64 1.16) (0.65 1.2 0)

PAGE 402

386 Table B.2 : Odds R atios (95% confidence intervals) for Contrace ptive Use at First S ex an d Family Structure; List wise deletion, no imputation Constant 1.25 0.97 0.89 0.88 0.92 2.11 2.41 (0.14 11.34) (0.10 9.67) (0.09 8.96) (0.09 8.53) (0.10 8.72) (0.19 23.37) (0.22 26.70) Observations 1,169 1,168 1,168 1,167 1,167 1,081 1,081 Table B.3 : Odds R atios (95% confidence interva ls) for Contraceptive Use at First S ex an d Family S tructure ; Imputed predictors only Model 1 Model 2 Model 3 Model 4 Model 5 Model 8 Model 9 First generation 0.50*** 0.48*** 0.53* 0.48*** 0.48*** 0.50*** 0.47~ (0.33 0.74) (0.32 0.72) (0.32 0.89 ) (0.32 0.72) (0.32 0.73) (0.34 0.74) (0.22 1.02) Second generation 0.64~ 0.53* 0.73 0.52* 0.50* 0.64~ 0.38* (0.40 1.02) (0.31 0.88) (0.41 1.30) (0.31 0.87) (0.29 0.86) (0.40 1.02) (0.17 0.85) Third generation Ref. Ref. Ref. Ref. Ref. Ref. Ref. Two biological parents 2.02*** 2.65*** 2.04*** 2.04*** (1.53 2.66) (1.67 4.19) (1.55 2.69) (1.55 2.69) First gen X two biological parents 0.74 (0.35 1.58) Second gen X two bio parents 0.53~ (0.27 1.06) Third gen X two bio parents 1.00 (1.00 1.00) Grandparents 1.98~ 1.49 (0.92 4.24) (0.63 3.51) First gen X grandparents 0.80 (0.06 11.46) Second gen X grandparents 2.10 (0.4 4 10.05) Third gen X grandparents 1.00 (1.00 1.00) Older sibling in the home 0.97 0.72

PAGE 403

387 Table B.3 : Odds R atios (95% confidence interva ls) for Contraceptive Use at First S ex an d Family S tructure ; Imputed predictors only (0.69 1.37) (0.46 1.13) First gen X older sibling 1.13 (0.45 2.87) Second gen X older sibling 2.32* (1.01 5.35) Third gen X older sibling 1.00 (1.00 1.00) Age at Wave I 0.96 0.99 0.99 0.99 0.99 0.96 0.96 (0.87 1.06) (0.89 1.10) (0.89 1.10) (0.89 1.10) (0.89 1.10) (0.87 1.06) (0.87 1.06) Age at first sex 1.07~ 1.04 1. 04 1.04 1.04 1.07~ 1.08~ (0.99 1.16) (0.96 1.13) (0.96 1.13) (0.96 1.13) (0.96 1.13) (0.99 1.16) (0.99 1.17) Male 0.88 0.84 0.84 0.85 0.84 0.88 0.91 (0.66 1.17) (0.62 1.13) (0.62 1.14) (0.62 1.15) (0.62 1.14) (0.66 1.17) (0.6 8 1.21) Constant 1.25 1.01 0.93 0.92 0.96 1.28 1.43 (0.14 11.34) (0.10 9.90) (0.09 9.20) (0.10 8.72) (0.10 8.92) (0.13 12.38) (0.15 13.96) Observations 1,169 1,169 1,169 1,169 1,169 1,169 1,169 Table B.4 : Odds R atios (95% confidence i ntervals) for Contraceptive Use at First S ex an d Family S tructure ; Fully imputed Model 1 Model 2 Model 3 Model 4 First generation 0.41* 0.43* (0.18 0.89) (0.16 1.16) (0.20 0.93) (0.16 1.02) Second generation 0.71 0.58 0.75 0.46* (0.47 1.08) (0.26 1.29) (0.49 1.16) (0.21 1.00) Third generation Ref. Ref. Ref. Ref. Number of siblings 1.07 1.04 (0.96 1.20) (0.84 1.29) First gen X siblings 0.99 (0.75 1.31) Second gen X siblings 1.10 (0.79 1.54) Third gen X siblings 1.00 (1.00 1.00) Presence of an older sibling 0.98 0.71 (0.68 1.39) (0.44 1.15) First gen X older sibling 1.16 (0.49 2.78)

PAGE 404

388 Table B.4 : Odds R atios (95% confidence i ntervals) for Contraceptive Use at First S ex an d Family S tructure ; Fully imputed Second gen X older sibling (0.94 5.46) Third gen X older sibling 1.00 (1.00 1.00) Age at Wave I 0.93 0.93 0.93 0.93 (0.83 1.05) (0.83 1.04) (0.82 1.04) (0.82 1.04) Age at first sex 1.09* 1.09* 1.09* 1.09* (1.00 1.17) (1.00 1.17) (1.01 1.18) (1.01 1.18) Male 1.00 1.00 1.00 1 .02 (0.75 1.33) (0.75 1.34) (0.75 1.33) (0.76 1.36) Constant 1.63 1.72 1.81 2.03 (0.18 15.07) (0.20 14.94) (0.19 17.31) (0.21 20.16) Observations 1,425 1,425 1,425 1,425 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Contraceptive use and family SES Parental education models Parental education differences (Models B.5 B.6 B.7) increases the odds of contraceptive use by 19% in the first generation as com pared to the education in the imputed model is 13% greater in the first generation versus the third and just out of significance. Table B.5 : Odds R atios (95% confidence i ntervals) for Contraceptive Use at First S ex an d Parental Education; L ist wise deletion, no imputations Model 1 Model 2 Model 3 Model 4 First generation 1.09 1.01 1.02 0.20 (0.63 1.89) (0.13 8.00) (0.49 2.11) (0.02 2.02) Second generation 0.8 5 1.11 0.74 0.69 (0.45 1.61) (0.09 14.09) (0.38 1.45) (0.07 7.07) Third generation Ref. Ref. Ref. Ref. Mother's education 1.07 1.07 (0.98 1.16) (0.96 1.19) First generation X mother's education 1.01 (0.83 1.23) Secon d generation X mother's education 0.98 (0.80 1.19) Third generation X mother's education 1.00 (1.00 1.00)

PAGE 405

389 Table B.5 : Odds R atios (95% confidence i ntervals) for Contraceptive Use at First S ex an d Parental Education; L ist wise deletion, no imputations Father's education 1.05 1.02 (0.95 1.17) (0.86 1.21) First generation X father's education 1.17* (0.97 1.42) Second generation X father's education 1.00 (0.84 1.20) Third generation X father's education 1.00 (1.00 1.00) Not working mother 1.88* 1.86* (0.98 3.62) (0.96 3.61) Unskilled mother 0.91 0.90 (0.51 1.61) (0. 50 1.60) Skilled/Professional mother Ref. Ref. Not working father 0.78 0.74 (0.32 1.90) (0.30 1.80) Unskilled father 1.23 1.19 (0.60 2.52) (0.59 2.43) Skilled/Professional father Ref. Ref. Wave I age 0.93 0.93 0 .91 0.91 (0.82 1.06) (0.82 1.06) (0.79 1.05) (0.79 1.05) Age at first sex 1.10* 1.10* 1.05 1.06 (0.99 1.22) (0.99 1.22) (0.94 1.18) (0.94 1.19) Male 0.85 0.85 0.96 0.95 (0.54 1.34) (0.53 1.34) (0.49 1.91) (0.48 1.91) Family Income 1.63*** 1.63*** 1.23 1.19 (1.14 2.33) (1.14 2.33) (0.60 2.52) (0.59 2.43) Resident Father 0.96 0.96 (0.55 1.68) (0.55 1.67) Constant 0.12 0.11 0.78 1.07 (0.00 3.24) (0.00 3.48) (0.02 29.32) (0.02 51.66) Observations 684 684 533 533 Table B.6 : Odds R atios (95% confidence intervals) for Contraceptive Use at First S ex an d Parental Education; I mputed predictors only Model 1 Model 2 Model 3 Model 4 First generation 0.66~ 0.71 0.74 0.17* (0.42 1.04) (0.12 4.04) (0.38 1.43) (0.03 0.94) Second generation 0.71 1.04 0.80 0.46 (0.41 1.23) (0.18 6.02) (0.44 1.44) (0.06 3.53) Third generation Ref. Ref. Ref. Ref. Mother's education 1.06~ 1.08

PAGE 406

390 Table B.6 : Odds R atios (95% confidence intervals) for Contraceptive Use at First S ex an d Parental Education; I mputed predictors only (0.99 1.14) (0.98 1.19) First generation X mo ther's education 1.00 (0.84 1.18) Second generation X mother's education 0.97 (0.83 1.12) Third generation X mother's education 1.00 (1.00 1.00) Father's education 1.05 0.99 (0.96 1.14) (0.87 1.11) First ge neration X father's education 1.15~ (0.99 1.34) Second generation X father's education 1.04 (0.89 1.22) Third generation X father's education 1.00 (1.00 1.00) Not working mother 1.48 1.48 (0.89 2.48) (0.88 2.47) Unskilled mother 1.14 1.12 (0.66 1.96) (0.66 1.90) Skilled/Professional mother Ref. Ref. Not working father 1.16 1.10 (0.53 2.51) (0.52 2.34) Unskilled father 1.09 1.05 (0.58 2.01) (0.58 1.89) Skilled/Professional father Ref. Ref. Wave 1 A ge 0.96 0.96 0.95 0.94 (0.85 1.07) (0.85 1.07) (0.84 1.06) (0.84 1.06) Age at first sex 1.06 1.06 1.05 1.05 (0.96 1.17) (0.96 1.17) (0.94 1.16) (0.95 1.17) Male 0.94 0.94 1.01 1.00 (0.69 1.28) (0. 69 1.29) (0.69 1.46) (0.69 1.46) Family Income 1.34* 1.34* 1.23 1.27 (1.01 1.79) (1.00 1.79) (0.91 1.66) (0.94 1.73) Resident Father 1.25 1.25 (0.85 1.83) (0.85 1.83) Constant 0.22 0.19 0.60 1.10 (0.01 4.64) (0.01 3.92) ( 0.03 11.90) (0.07 18.07) Observations 1,084 1,084 852 852

PAGE 407

391 Table B.7 : Odds R atios (95% confidence intervals) for Contraceptive Use at First S ex an d Parental Education; Fully imputed Model 1 Model 2 Model 3 Model 4 First generation 0.55 0.54 0.61 0 .22 (0.23 1.29) (0.08 3.53) (0.24 1.51) (0.04 1.36) Second generation 0.84 0.98 0.92 0.76 (0.52 1.38) (0.15 6.54) (0.51 1.66) (0.10 5.60) Third generation Ref. Ref. Ref. Ref. Mother's education 1.06 1.06 (0.99 1.13) (0.96 1.17) First gen X mother's education 1.00 (0.84 1.20) Second gen X mother's education 0.99 (0.83 1.16) Third gen X mother's education 1.00 (1.00 1.00) Father's education 1.03 0.99 (0.96 1.11) (0.89 1.11) First gen X father's education 1.11 (0.95 1.29) Second gen X father's education 1.01 (0.87 1.17) Third gen X father's education 1.00 (1.00 1.00) Not working mother (0.93 2.55) (0.94 2.54) Uns killed mother 1.10 1.10 (0.64 1.91) (0.65 1.87) Skilled/Professional mother Ref. Ref. Not working father 1.12 1.08 (0.52 2.41) (0.51 2.29) Unskilled father 1.01 0.99 (0.57 1.78) (0.57 1.71) Skilled/Professional fat her Ref. Ref. Age at Wave I 0.92 0.92 0.91 0.91 (0.81 1.05) (0.81 1.05) (0.80 1.03) (0.80 1.03) Age at first sex 1.08 1.08 (0.99 1.18) (0.99 1.18) (0.97 1.19) (0.98 1.19) Male 1.07 1.07 1.21 1.21 (0.77 1.47) ( 0.77 1.48) (0.85 1.72) (0.85 1.73) Family Income 1.27 1.27 1.26 1.28 (0.94 1.72) (0.94 1.72) (0.92 1.71) (0.93 1.77) Resident Father 1.12 1.12

PAGE 408

392 Table B.7 : Odds R atios (95% confidence intervals) for Contraceptive Use at First S ex an d Parental Education; Fully imputed (0.77 1.63) (0.77 1.63) Constant 0.41 0.39 1.02 1.48 (0.02 7.01) (0.02 7.30) ( 0.06 18.06) (0.10 22.96) Observations 1,333 1,333 1,047 1,047 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Family SES Parental employment differences (Tables B.8 B.9 B.10) Having a mother not working increased the odds of contraceptive use by 88% co mpared to unskilled and skilled and professional moms in the list wise model. This was weakly significant p<0.10 and not recovered in the imputed predictor model but was recovered in the fully imputed model. There was a significant interaction between moth employment and the first generation (lower odds for not working mothers versus professional mothers) in the listwise deletion model, but it was very unstable (OR for the first generation increased to 92.07) probably due to the small amount of skilled/ professional mothers in the first generation, especially after removing those missing on family income.) The list wise deletion model showed that in the second generation, having a not working father versus a professional father increased the odds of contr aceptive use in the second generation while it decreased the odds in the third generation. This was not significant in the imputed predictor model nor the full imputed model though the ORs are similar. Parental employment models Table B.8 : Odds R atios (95 % confidence intervals) for Contraceptive Use and Parental Employment; L ist wise deletion, no imputation Model 1 Model 2 Model 3 Model 4 First generation 1.09 92.07*** 1.02 0.95 (0.63 1.89) (8.30 1,021.33) (0.49 2.11) (0.33 2.71) Second gener ation 0.85 1.06 0.74 0.67 (0.45 1.61) (0.41 2.73) (0.38 1.45) (0.29 1.57) Third generation Ref. Ref. Ref. Ref. Not working Mother 1.88* 2.22* (0.98 3.62) (0.93 5.29) Unskilled Mother 0.91 1.18 (0.51 1.61) (0.59 2.35) Professional Mother Ref. Ref. First gen X not working mother 0.01*** (0.00 0.09) First gen X unskilled mother 0.01***

PAGE 409

393 Table B.8 : Odds R atios (95 % confidence intervals) for Contraceptive Use and Parental Employment; L ist wise deletion, no imputation (0.00 0.13) First gen X skilled/professional mother 1.00 (1.00 1.00) Second gen X not working m other 0.78 (0.19 3.23) Second gen X unskilled Mother 0.51 (0.16 1.63) Second gen X skilled/professional mother 1.00 (1.00 1.00) Not working Father 0.78 0.31* (0.32 1.90) (0.08 1.14) Unskilled Father 1.23 1.58 (0.60 2.52) (0.57 4.34) Professional Father Ref. Ref. First gen X not working father 1.91 (0.32 11.31) First gen X unskilled father 0.88 (0.22 3.48) First gen X skilled/professional father 1.00 (1.00 1.00) Second gen X not working father 6.46** (1.52 27.45) Second gen X unskilled father 0.65 (0.20 2.15) Second gen X skilled/professional father 1.00 (1.00 1.00) Mother's education 1.07 1.05 (0.98 1.16) (0.96 1.15) Fa ther's education 1.05 1.06 (0.95 1.17) (0.96 1.16) Age 0.93 0.93 0.91 0.92 (0.82 1.06) (0.82 1.06) (0.79 1.05) (0.79 1.06) Age*Age 1.10* 1.10* 1.05 1.04 (0.99 1.22) (1.00 1.22) (0.94 1.18) (0.92 1.17) Male 0.85 0.83 0.96 0. 92 (0.54 1.34) (0.52 1.35) (0.49 1.91) (0.45 1.86) Family Income 1.63*** 1.64*** 1.31 1.27

PAGE 410

394 Table B.8 : Odds R atios (95 % confidence intervals) for Contraceptive Use and Parental Employment; L ist wise deletion, no imputation (1.14 2.33) (1.13 2.37) (0.92 1.88) (0.90 1.80) Resident Father 0.96 0.98 (0.55 1.68) (0.56 1.73) Constant 0.12 0.12 0.78 1.05 (0. 00 3.24) (0.00 3.15) (0.02 29.32) (0.02 45.51) Observations 684 684 533 533 Table B.9 : Odds R atios (95% confidence intervals) for Contraceptive Use and Parental Employment; I mputed predictor model Model 1 Model 2 Model 3 Model 4 First g eneration 0.66~ 0.71 0.74 0.51 (0.42 1.04) (0.17 3.02) (0.38 1.43) (0.17 1.55) Second generation 0.71 0.98 0.80 0.85 (0.41 1.23) (0.39 2.47) (0.44 1.44) (0.37 1.98) Third generation Ref. Ref. Ref. Ref. Not working Mother 1.48 1 .70 (0.89 2.48) (0.85 3.43) Unskilled Mother 1.14 1.32 (0.66 1.96) (0.59 2.94) Skilled/professional Mother Ref. Ref. First gen X not working mother 0.66 (0.10 4.19) First gen X unskilled Mother 1.15 (0.22 5 .92) First gen X skilled/profess mother 1.00 (1.00 1.00) Second gen X unemployed mother 0.72 (0.21 2.40) Second gen X unskilled Mother 0.50 (0.13 1.88) Second gen X skilled/profess mother 1.00 (1.00 1.00) Not working Father 1.16 0.70 (0.53 2.51) (0.18 2.67) Unskilled Father 1.09 1.19 (0.58 2.01) (0.50 2.85) Professional Father Ref. Ref.

PAGE 411

395 Table B.9 : Odds R atios (95% confidence intervals) for Contraceptive Use and Parental Employment; I mputed predictor model First gen X not working father 2.15 (0.48 9.68) First gen X unskilled father 1.5 0 (0.41 5.52) First gen X skilled/profess father 1.00 (1.00 1.00) Second gen X not working father 2.35 (0.62 8.88) Second gen X unskilled father 0.65 (0.22 1.95) Second gen X skilled/profess father 1.00 (1.0 0 1.00) Mother's education 1.06~ 1.05 (0.99 1.14) (0.98 1.13) Father's education 1.05 1.05 (0.96 1.14) (0.97 1.14) Age at wave I 0.96 0.96 0.95 0.95 (0.85 1.07) (0.86 1.07) (0.84 1.06) (0.84 1.06) Age at first sex 1.06 1 .06 1.05 1.05 (0.96 1.17) (0.96 1.17) (0.94 1.16) (0.94 1.16) Male 0.94 0.93 1.01 0.99 (0.69 1.28) (0.68 1.27) (0.69 1.46) (0.68 1.45) Family Income 1.34* 1.35* 1.23 1.21 (1.01 1.79) (1.01 1.80) (0.91 1.66) (0.88 1.65) Resi dent Father 1.25 1.22 (0.85 1.83) (0.82 1.81) Constant 0.22 0.22 0.60 0.66 (0.01 4.64) (0.01 4.30) (0.03 11.90) (0.03 14.03) Observations 1,084 1,084 852 852 *** p<0.001, ** p<0.01, p<0.05, p<0.10 Table B.10 : Odds R atios (95% confidence intervals) for Contraceptive Use and Parental Employment; F ully imputed model Model 1 Model 2 Model 3 Model 4 First generation 0.55 0.68 0.61 0.39 (0.23 1.29) (0.18 2.61) (0.24 1.51) (0.10 1.44) Second generation 0.84 1.05 0.92 0.9 9 (0.52 1.38) (0.44 2.49) (0.51 1.66) (0.42 2.31) Third generation Ref. Ref. Ref. Ref.

PAGE 412

396 Table B.10 : Odds R atios (95% confidence intervals) for Contraceptive Use and Parental Employment; F ully imputed model Not working Mother 1.54~ 1.71 (0.93 2.55) (0.86 3.40) Unskilled Mother 1.10 1.29 (0.64 1.91) (0.59 2.85) Skilled/professional Mot her Ref. Ref. First gen X not working mother 0.61 (0.12 3.16) First gen X unskilled Mother 0.86 (0.17 4.28) First gen X skilled/profess mother 1.00 (1.00 1.00) Second gen X unemployed mother 0.82 (0.27 2.5 4) Second gen X unskilled Mother 0.59 (0.18 1.97) Second gen X skilled/profess mother 1.00 (1.00 1.00) Not working Father 1.12 0.68 (0.52 2.41) (0.20 2.31) Unskilled Father 1.01 1.08 (0.57 1.78) (0.45 2.60) Pr ofessional Father Ref. Ref. First gen X not working father 2.41 (0.48 12.11) First gen X unskilled father 1.70 (0.48 5.96) First gen X skilled/profess father 1.00 (1.00 1.00) Second gen X not working father 2.15 (0.58 7.98) Second gen X unskilled father 0.68 (0.23 2.01) Second gen X skilled/profess father 1.00 (1.00 1.00) Mother's education 1.06 1.05 (0.99 1.13) (0.98 1.12) Father's education 1.03 1.03 (0.96 1.11) (0.96 1.11) Age at wave I 0.92 0.92 0.91 0.91

PAGE 413

397 Table B.10 : Odds R atios (95% confidence intervals) for Contraceptive Use and Parental Employment; F ully imputed model (0.81 1.05) (0.81 1.05) (0.80 1.03) (0.80 1.03) Age at first sex 1.08 1.08 (0.99 1.18) (0.99 1.18) (0.97 1.19) (0.97 1.20) Male 1.07 1.07 1.21 1.19 (0.77 1.47) (0.77 1 .48) (0.85 1.72) (0.82 1.72) Family Income 1.27 1.28 1.18 1.18 (0.94 1.72) (0.94 1.73) (0.89 1.58) (0.88 1.59) Resident Father 1.12 1.11 (0.77 1.63) (0.76 1.61) Constant 0.41 0.39 1.02 1.04 (0.02 7.01) (0.02 6.40) (0.06 1 8.06) (0.05 19.85) Observations 1,333 1,333 1,047 1,047 *** p<0.001, ** p<0.01, p<0.05, p<0.10