Citation
Using socioeconomic indicators to predict the academic outcomes of African American students

Material Information

Title:
Using socioeconomic indicators to predict the academic outcomes of African American students
Creator:
Easton-Brooks, Donald
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
x, 138 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Prediction of scholastic success ( lcsh )
African American students -- Social conditions ( lcsh )
African American students -- Economic conditions ( lcsh )
Social status ( lcsh )
African American students -- Economic conditions ( fast )
African American students -- Social conditions ( fast )
Prediction of scholastic success ( fast )
Social status ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 122-131).
Thesis:
Education
General Note:
School of Education and Human Development
Statement of Responsibility:
by Donald Easton-Brooks.

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:
75391177 ( OCLC )
ocm75391177
Classification:
LD1193.E3 2006d E47 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

USING SOCIOECONOMIC INDICATORS TO PREDICT THE ACADEMIC OUTCOMES OF AFRICAN AMERICAN STUDENTS by Donald Easton-Brooks B.A., Greenville College, 1988 M.A., University of Colorado at Denver, 1995 A thesis submitted to the University of Colorado at Denver/Health Sciences Center in partial fulfillment of the requirements for the degree of Doctor ofPhilosophy School of Education and Human Development 2006

PAGE 2

This thesis for the Doctor of Philosophy degree by Donald Easton-Brooks has been approved by AlillDaVis Ph.D. Rodney Muth Ph"b. D hy Gamson-Wade Ph.D. Edward H. Freeman Ph.D.

PAGE 3

Easton-Brooks, R. (Ph.D., Educational Leadership and Innovation) Using Socioeconomic Indicators to Predict the Academic Outcomes of African American Students Thesis directed by Associate Professor A1an Davis ABSTRACT Much research indicates that socioeconomic status (SES) correlates with the academic outcomes of students. Related to African American students, recent fmdings suggest that wealth/assets also predicts the academic outcomes of these students. This study examined whether socioeconomic indicators (parents' education, parents' occupation, parents' income, and wealth) related to both African Americans and European Americans predicted the academic outcomes of these two groups differently. The sample of 1,302 African American and 6,362 European American public high school students drawn from the first and second year follow up of the public-use National Educational Longitudinal Study of 1988 (NESL:88) shows that SES predicts the academic achievement and academic engagement of African American students. In addition, this study found that wealth/assets accounts for more variance in the academic outcomes of African American students than of European American students. This abstract accurately represents the content of the candidate's thesis.)}ecommend its Signed_. AJanDavis

PAGE 4

DEDICATION PAGE For my loving wife (Lori), my comfort. For my sons Dawson and Devyn, as a testament of high academic success.

PAGE 5

ACKNOWLEDGMENT In 1848, Benjamin Roberts sued Boston Public Schools for not allowing his five-year-old daughter to attend the local primary school in their community. Despite a prior ordinance outlawing the exclusion of any child from the public schooling, Roberts unsurprisingly lost the case (Sarah Roberts v. The City of Boston, 1850). One hundred and four years later, the Brown family won a similar case (Brown v. The Board of Education of 1954). During the hundred years between these two cases, many African Americans fought to gain equitable educational opportunities. Since 1954, many continue to fight for equitable educational opportunities for all African Americans. So, I would like to acknowledge Linda Brown, Charlayne Hunter-Gault, Charles Hamilton Houston, WEB DuBois, Barbara Jordan, the Little Rock Nine, and the many African American pioneers who publicly challenge educational systems. These efforts helped young brothers and sisters to make it through the educational process. I would also like to acknowledge Martin Luther King Jr., Malcolm X, and the many that died so that young brothers or sisters could make it through the educational process. I honor those who came before me, and I respond by carrying the torch so that young brothers and sisters can make it through the educational process. In addition, I would also like to thank Dr. Alan Davis for his guidance through this dissertation process and to Dr. Donna Wittmer my advisor for both my master's and doctoral programs. You two have been supportive, compassionate, and a great

PAGE 6

inspiration. Additionally, I would like to thank those who served on my dissertation committee, Dr. Rodney Muth, Dr. Dorothy Garrison-Wade, and Dr. Edward Freeman. Thanks for supporting my work. Finally, thanks to my classmates and members of the Doctoral Students of Colors at the University of Colorado at Denver and Health Sciences Center. Thanks for helping me grow and for sharing your journey with me.

PAGE 7

TABLE OF CONTENTS FIGURES .................................................................................................................... IX TABLE ......................................................................................................................... X CHAPTER ONE INTRODUCTION ........................................................................................................... 1 The General Problem ............................................................................................. l The Background of SES Construction ................................................................... 5 Theoretical Framework .......................................................................................... 6 Methodology ........................................................................................................ 13 Overview of Chapters ........................................................................................... 15 CHAPTER TWO LITERATURE REVIEW .................... ................................... 16 The History of SES .............................................................................................. 16 Socioeconomic Status and the African American Community ............................ 28 Wealth and the African American Community .................................................... 31 Academic Outcomes and SES .............................................................................. 37 Engagement. ...................................................................................................... 37 Achievement ..................................................................................................... 45 Attainment. ........................................................................................................ 52 Wealth and Academic Outcomes ......................................................................... 58 CHAPTER THREE METIIODOLOGY ........................................................................................................ 64 Hypotheses ........................................................................................................... 65 Dataset .................................................................................................................. 65 Sample .................................................................................................................. 67 Dependent Variables ............................................................................................ 68 Engagement. ...................................................................................................... 68 Achievement ..................................................................................................... 74 Attainment. ........................................................................................................ 76 Independent Variables .......................................................................................... 78 Assets ................................................................................................................ 78 SES Indicators ................................................................................................... 82 Data Analysis ....................................................................................................... 84 Vll

PAGE 8

CHAPTER FOUR RESULTS ................................................................................................................... 90 Factor Analyses .................................................................................................... 91 Factor Analysis on socioeconomics related to African Americans .................. 91 Factor Analysis on socioeconomics related to European Americans ............... 94 Multiple Regressions ............................................................................................ 97 Academic Engagement for African American students .................................... 98 Academic Engagement for European American students ................................ 98 Academic Achievement for Mrican American students .................................. 99 Academic Achievement for European American students ............................. 101 Academic Attainment of African American students ..................................... 101 Academic Attainment of European American students ... Error! Bookmark not defined. Summary of Findings ......................................................................................... 104 CHAPTER FIVE CONCLUSION ........ 107 Limitations and Future Studies .......................................................................... 108 Limitation 1 ..................................................................................................... 108 Addressing Limitation 1 ................................................................................. 109 Limitation 2 ..................................................................................................... 110 Addressing Limitation 2 ................................................................................. 111 Limitation 3 ..................................................................................................... 111 Addressing Limitation 3 ................................................................................. 112 Implication of this Study .................................................................................... 112 REFERENCES ......................................................................................................... 116 APPENDIX APPENDIX A: vARIABLES AND CODES .................................................................... 129 NELS:88 Variables ............................................................................................ 129 New Composite Variable (NCV) ....................................................................... 131 viii

PAGE 9

LIST OF FIGURES Figures 1. Socioeconomic Indicators related to African Americans and the academic outcome of African American and European American students ..................... 10 2. Socioeconomic Indicators related to European Americans and the academic outcome of African American and European American students ..................... 11 ix

PAGE 10

2.1 2.2 2.3 2.4 2.5 2.6 2.7 3.1 3.2 3.3 3.4 3.5 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.6 4.10 411 5.1 5.2 AI LIST OF TABLES Table Percent of the education, income, and occupation criterion during 1980 .. Debt after college graduation .................................................... Percent of students who took part in extracurricular activities in 1992 .... Percentage by race participated in extracurricular activities in 1992 ....... Mean ACf scores for 2001 ...................................................... Percentage of reading and mathematic of 8th graders at or below level. .. Achievement scores between African Americans and European Americans ........................................................................ Background characteristics of the independent variables .................... Component matrix of the three types of engagement ........................ Reliability scale of the seven engagement components ...................... On-time graduation rates ......................................................... Back characteristics of the dependent variables ............................... Un-rotated component matrix for African Americans ....................... Rotated component matrix for African Americans ........................... Un-rotated component matrix for European Americans ..................... Rotated component matrix of European Americans ......................... Regression of academic engagement of African Americans ................ Regression of academic engagement of European Americans .............. Regression of academic achievement of African Americans ............... Regression of academic achievement of European Americans ............. Regression of academic attainment of African Americans .................. Regression of academic attainment of European Americans ................ Summarizing hypotheses ........................................................ Academic outcomes of African Americans and European Americans ..... Homeownership .................................................................... APPENDIX A: Socioeconomic Index .......................................... X 29 33 41 41 44 45 47 68 72 73 77 78 91 92 94 96 99 99 101 101 102 102 104 111 113 127

PAGE 11

CHAPTER ONE Introduction The General Problem This dissertation examines the relationship between socioeconomic indicators and the academic outcomes (engagement, achievement, and attainment) of African American public high school students. Historically, researchers measured the relationship between socioeconomic indicators and the academic performance of African American students using socioeconomic status (SES) as the major indicator of families' socioeconomic experiences. In national datasets such as the National Longitudinal Study, 1972 (NLS:72); High School and Beyond, 1980 (HS0:80); and National Educational Longitudinal Study, 1988 (NELS:88), SES represents an aggregate of socioeconomic indicators which includes father's income, mother's income, father's education level, mother's education level, father's occupational prestige, and mother's occupational prestige. While researchers use these aggregated socioeconomic indicators to examine the relationship between SES and the academic performance of African American students, they have presented inconclusive research findings (Adams & Singh, 1998; Battle, 1997; Carter, 2003; Craig, Connor, & Washington, 2003; Smith-Maddox, 1998). 1

PAGE 12

The selection of SES as a variable for assessing academic performance of African Americans derives from the Coleman, Campbell, Hobson, McPartland, Mood, Weinfeld, and York (1966) Equality of Education Opportunity Report. Coleman et al. explained that a variety of social factors could contribute to the gap between the educational opportunities of African American and European American students. The Equality of Education Opportunity Report led researchers to study how social influences related to historical struggles of African American public school students. Hedges and Nowell's (1998) exploration of various national datasets, ranging from 1965 to 1997, showed that the standardized test scores of African American students historically represented the highest proportion of students scoring in the lower 5th, lOth, and 25th percentile. Additionally, based on No Child Left Behind Legislation (NCLB) (200 1 ), which specifies high school graduation in four years, Swanson (2004) found that 50% of African American students graduated on time, 76% of European American students graduated on time, and 68% of the U.S. student population graduated on time. These studies demonstrate the need for research on variables that influence the academic outcomes of African American students. Researchers have used national datasets to analyze the relationship between various social factors and the academic outcomes of African American students. In general, researchers and policymakers have concluded that SES is a strong covariate or predictor of the academic performance of African American students. The fundamental justification for using SES as a criterion that correlates with the 2

PAGE 13

academic performance of African Americans stems from the belief that the variable SES explains how these students will perform in school (Battle & Coates, 2004). Based on this justification, research using the SES variable shows that students from higher SES households perform better academically (Duncan &Magnuson, 2005; Jencks & Phillips, 1998). However, when analyzing the relationship between SES and the academic performance of African American students, research findings are inconclusive. Some studies demonstrated a significant relationship between SES and the academic achievement of African American students (Adams & Singh, 1998; Lee, 1993). Other studies revealed no significant relationship between SES and the academic achievement of African American students (Battle, 1997; Ford, 1993). In contrast, Craig, Conner, & Washington's (2003) study showed that Africari American students from a lower SES community achieved higher test scores than did those African American students from a higher SES community. The dissimilarities in the findings between SES and the academic performance of African American students may be a result of the socioeconomic indicators used to measure SES. While SES may represent the socioeconomic experiences ofthe general population, Conley (1999), Oliver and Shapiro (1997), and Shapiro (2004) explain that the socioeconomic indicators that are more relevant to African Americans are debts, wealth, and assets. Conley (1999) argues that wealth impacts academic attainment more than SES. In addition, Orr (2004) argues that 3

PAGE 14

wealth has a significant impact the academic achievement of African American students. Further, the socioeconomic experiences of African Americans are a result of historic poverty experienced by this group. For instance, the African American population has historically been the one of the most impoverished ethnic groups in the United States (Alexis, 1998; Myers, Kim, & Mandala, 2004). Additionally, the U.S. Bureau of the Census (2002) showed that while the annual income of African Americans has gradually increased over time, their annual income is noticeably lower than that of European Americans. Kaufman, Crosnoe, & Elder (2004) also explain that, when controlling for education level, on average, European Americans earn more per year than African Americans. The difference in the socioeconomic experiences of African Americans and European Americans is also reflective of the difference in the two groups' wealth. Rothestein (2004) explains that given the same household income, the wealth of African Americans is equal to 12% ofthe wealth of European Americans. Often the wealth held by African Americans is in debt at a higher rate than the wealth held by European Americans. For instance, if an African American family and European American family owns a car or a house, the African American family will often owe more on the loans associated with these properties than will the European American family. According to Shapiro (2004), different types of wealth are associated with each group. While the wealth of many European American families comes from the 4

PAGE 15

wealth of their ancestors, the wealth of most African Americans comes from income, loans, or credit cards. This argument proposes that wealth for European Americans is a way of life. However, given the historical poverty experienced by. African Americans, wealth is difficult to obtain. The purpose of this study is to examine the impact wealth has on the academic performance of African American students. In this study used parents' assets as the measurement for wealth. Orr (2004) and Shapiro (2004) view assets as a form of wealth. For the purpose of clarification, assets consist of any item of economic value that a family can convert into cash. In this explanation, families' annual income does not represent assets (Orr, 2004; Shapiro, 2004). On the other hand, property owned, investments, stocks, bonds, trust funds, and savings accounts are forms of assets. The Background of SES Construction While wealth is a valuable part of a family's economic lifestyle, researchers do not often use wealth as an indicator of a family's SES. The socioeconomic indicators used to determine SES come from Duncan's (1961) Socioeconomic Index Scores (SEI). This index consists of occupational prestige, income, and education. Although Duncan ( 1961) was not the first to develop a SEI scale, his study presented the first examination of the statistical relationship between the socioeconomic indicators, education, income, and occupation. His study focused on males and the U.S. Census of 1949. Over time, researchers modified and utilized the SEI index as the criterion for evaluating socioeconomic status (Nakao & Treas, 1994; Stevens & 5

PAGE 16

Featherman, 1981). More recently, researchers have modified SES to include father's income, mother's income, father's education, mother's education, and father's and mother's occupation as measured by the SEI. Though the SEI index is a major component of the SES measure, research has yet to show ifthe index responds to the cultural experience of various U.S. populations. Additionally, the researchers have yet to show if the index has yet to account for the various historical and socio-cultural differences between dominant majority culture and caste-like minority cultures. Wooden (1933) proposes that given African Americans' oppressive history with poverty, employment, and education, their response to society and economics is different from that of European Americans. The difference in the economic experience of the two groups prompted this study to examine which socioeconomic indicators best explain the socioeconomic experience of African Americans. Afterwards, this study looked to understand the relationship between socioeconomic indicators and the academic outcomes of African American students. The purpose of these explorations were to examine whether socioeconomic indicators related to African Americans predicted the academic outcomes of African American students differently than the socioeconomic indicators related to European Americans. Theoretical Framework The theoretical framework for this study responds to the inclusive research findings on the relationship between SES and the academic outcomes of African 6

PAGE 17

Americans as well as to the argument that wealth is a major socioeconomic indicator of African Americans' socioeconomic economic experience. Although research shows inclusive findings when assessing the relationship between SES and the academic outcomes of African Americans, researchers show that the socioeconomic indicators income, education, and occupation statistically correlate. Duncan ( 1961) found that the average correlation among these socioeconomic indicators was 0.75. Nakao & Treas (1994) presented similar findings. Additionally, current literature explains that wealth accounts for a portion of African Americans' academic achievement and academic attainment (Conley, 1999; Orr, 2004). Orr (2004) explains that wealth accounts for a portion of the difference between European Americans and African Americans' academic achievement. Conley ( 1999) found that family wealth and types of assets highly correlate to the high school and college attainment of African American students. He further found that when assets are not present, income is the next greatest predictor of the academic attainment of African American students. Conley (1999) proposed that wealth, more than SES, allows families resources to provide their children with better educational opportunities. Conley (1999) also describes that with wealth and assets, parents can buy a home in a high-income neighborhood. He illustrates that in high-income neighborhoods, the elevated property taxes provide schools in these neighborhoods with better educational resources than schools in low-income neighborhoods. Additionally, wealth gives families the ability to provide extra necessities (e.g., 7

PAGE 18

computer, internet, tutors) that may contribute to the academic success of their children. By viewing wealth as a major socioeconomic indicator, this framework proposes that wealth provides a more relevant understanding of the relationship between the socioeconomic experience of African Americans and their academic outcomes than the relationship between the traditionally used SES composite variable and the academic outcomes of African American students. Additionally, this study supports Duncan's ( 1961) belief that income, education, and occupation present a more comprehensive explanation of the social and economic status of families. However, to examine how these variables respond to the socioeconomic experience of African Americans, this study disaggregates and assesses the socioeconomic indicators (income, education, and occupational prestige scores) used to construct the traditionally used SES variable as individual socioeconomic indicators. Disaggregating these values allowed this study the opportunity to understand how well income, education, occupation, and wealth predict the academic outcome of African American students. First, this study examined which socioeconomic indicators closely related to the socioeconomic experience of African Americans. The process provided information needed to understand the relationship between occupation, education, income, and wealth for the sampled African American population. This study used data to predict the relationship between these socioeconomic indicators and the 8

PAGE 19

academic outcomes of the sampled African American and European American students. These findings provided information on whether socioeconomic indicators related to African Americans predicted the academic outcomes of African American and European American students differently. Next, this study assessed which indicators related to the socioeconomic experience of the sampled European American population. This study used these findings to predict the relationship between these socioeconomic indicators and the academic outcomes of the sampled African American and European American students. The study then examined whether the socioeconomic indicators related to European Americans predicted the academic outcomes of African American and European American students differently. Based on the examination suggested above, this study proposes the following hypotheses: 1. Socioeconomic indicators related to African Americans predict academic engagement, academic achievement, and academic attainment for African American and European American students differently (Figure 1 ). 2. Socioeconomic indicators related to European Americans predict academic engagement, academic achievement, and academic attainment for both African American and European American students differently (Figure 2). 9

PAGE 20

Figure l: Socioeconomic IDdicaton retaied to the African American llld the aademic outcomes of Aftic:an American and EuropcaD America.D studcota Socioecooom.ic Indicators Component Variables Ac:ad.emlc Outcomes of African American Studems J r--A:-cad--:-llml--:.-c --,/ Outcomes of European AmcricaD Students AcU Cmic Achievemeol

PAGE 21

Figure 2: Soc;ioc:conomic IDdicaton n:lued to lhe European American and the academic outcomes of AJ'rlcan American and European American studmtl Socioeconomic Indicators Component Variables Academic Outcomes of African AmericaD Stude IllS Aclldllmlc Enpgc:ment European Attainmeot American Students ""'

PAGE 22

The hypotheses and theoretical framework proposed in this study addresses the methodological limitations presented in existing studies. Additionally, the hypotheses and theoretical framework examined the inconsistent approaches commonly used to assess the relationship between socioeconomic indicators and the academic outcomes of African Americans. Using socioeconomic indicators relevant to African Americans, this study examined whether these socioeconomic indicators best predict the academic outcomes of African Americans. The dependent variables in this study are academic outcomes as measured by academic engagement, academic attainment, and academic achievement. The independent variables in this study are the disaggregate SES indicators (annual household income, father's education, mother's education, father's occupation, and mother's occupation) and wealth as measured by assets. Methodology The public-use data file of the National Educational Longitudinal Study of 1988 (NELS:88) provided a sample needed to answer the research questions. The NELS:88 is a nationally representative sample of public and private school students surveyed in 1990, 1992, 1994, 2000, and 2004. The National Center ofEducational Statistics (NCES) and the U.S. Department of Education collected the data. The participants in the NESL:88 provided information related to individual background, family background, school experiences, school behaviors, and test-based test scores. 12

PAGE 23

To establish relevant socioeconomic indicators, separate principal component factor analyses with Varimax rotation on the seven socioeconomic indicators showed the common statistical relationship between the socioeconomic indicators. These factor analyses determined the level to which a common set of socioeconomic indicators explained the variation in socioeconomic experience of the sampled populations. The first factor analysis examined how the socioeconomic indicators collectively responded to the sampled African Americans. This analysis produced socioeconomic components related to African Americans (Hypothesis 1). The second factor analysis examined how socioeconomic indicators collectively responded to the sampled European Americans. This analysis produced socioeconomic components related to European Americans (Hypothesis 2). The socioeconomic components related to African Americans and European Americans provided data needed to run a series of multiple regression analyses. A series of analyses examined whether socioeconomic components related to the two groups predicted the academic outcomes of African American and European American students differently. One multiple regression analysis examined the relationship between the socioeconomic indicators and the academic outcomes of African American students (Hypothesis 1 ). The next multiple regression analysis examined the relationship between the socioeconomic indicators and the academic outcomes of European American students (Hypothesis 2). 13

PAGE 24

This study compared the findings from the series of analyses above to a series of multiple regression analyses on the relationship between the traditionally used SES variable, race, and academic outcomes. The analyses examined the relationship between SES, race and academic engagement; SES, race, and academic achievement; and SES and academic attainment. The purpose of these analyses was to understand whether socioeconomic indicators related to each the African American and European American populations predicted the academic gap between the two groups differently than the traditionally used SES variable. Overview of Chapters The chapters to follow provide additional supports for this dissertation. Chapter 2 reviews literature related to research findings, empirical evidence, and theoretical works associated with this current study. Chapter 3 includes the method for obtaining the descriptive data, the representative population, and data analysis used in the current study. Chapter 4 provides the results of the statistical analyses used to answer the research questions. Finally, Chapter 5 discusses the findings, the limitations of the study, validity concerns, and the need for further research. 14

PAGE 25

CHAPTER TWO Literature Review This chapter presents an overview of the history of SES; social status and the African American community; wealth and African Americans; African Americans' academic engagement, academic achievement, and academic attainment; and wealth and the academic outcomes of African American students. Based on the theoretical framework presented in Chapter 1, this chapter provides support for using socioeconomic indicators as relevant components for predicting the academic outcomes of African American students. The History of SES Socioeconomic status (SES) is a hierarchical scale that measures the social and economic status of a household (Johnson, 2000; Sirin, 2005). For years, researchers struggled with ways to operationalize SES. The struggle consisted of researchers' failure to fmd socioeconomic indicators that were non-subjective measures of SES. Before the 1940's, researchers used social and fmancial information from small urban and rural communities as indicators of SES (Powers, 1982). The researchers believed that small communities kept actual financial records and could provide common knowledge about the social status of individuals in their communities (Powers, 1982). 15

PAGE 26

Power ( 1982) explained that this approach represented biases because this method relied on subjective data. In smaller communities, people's views of SES came from their opinion of a family's social status in that community. For instance, if a doctor in a small community had a son who became a blacksmith in that same community, the status of the blacksmith would be dependent on his father's social status. If the people in the community valued the doctor, the community would then value the doctor's son. In turn, the social status of the blacksmith would be dependence on the social status of his father. The difficulty of developing a less subjective measurement of SES forced researchers to tum their attention to the variable income as reported to the U.S. Bureau of the Census. Researchers selected the variable income because the variable naturally presented itself as a hierarchical measure. Further, Duncan ( 1961) explained that income data reported to the U.S. Bureau of the Census provided a better account of economic status because census data presented a nationally representative sample. Additionally, census data provided researchers with a less subjective method for measuring SES. Researchers' attempts to understand the relationship between income and SES led them to discover the consistency between income and types of occupations. As researchers examined income, they saw a pattern between this variable and occupation (Duncan, 1961 ). However, the problem researchers faced was in 16

PAGE 27

converting types of occupations into a measurable hierarchical variable. Furthermore, researchers were unclear on how to deal with validity and reliability issues when ranking occupations. The problem researchers faced was developing a measure that related to the importance of an occupation while simultaneously and systematically accounting for SES. Powers (1982) explained that to deal with these issues, researchers proposed three solutions. The frrst solution was to continue to solicit public opinion on socioeconomic status by restricting opinions related to rating the prestige of occupational categories. The second solution was to use the income and the education as estimators of occupational success. The third solution was to accept both approaches as a method for measuring SES (Powers, 1982, p. 30). Powers (1982) explained that these solutions represented either a prestige of occupation model or a status of occupation model. He further described prestige of occupation as a judgment people generally had regarding the status of an occupation (p. 31 ). The status of an occupation referred to the educational requirements and economic rewards associated with an occupation. Although researchers defined the methods by which to measure SES, they were still uncertain of the criterion for developing a hierarchical scale for occupation. In 1925, George Counts developed the first hierarchical scale of occupations (Nakao & Treas, 1994; Powers, 1982). His scale represented a prestige of occupation model. Counts frrst created a list of 45 occupations he believed 17

PAGE 28

represented the entire range of careers (Power, p. 31 ). Counts then conducted a study by which he asked participants in six different groups to place the 45 occupations into nine categories. Counts instructed the participants to arrange the categories into a single list of occupations with the most respected occupations receiving the highest rankings and the least respected occupations receiving the lowest rankings. Counts then analyzed the participants' responses, ranked the 45 occupations, and developed a hierarchal scale of occupations. Researchers used Counts' occupational scale until the National Opinion Research Center (NORC) conducted the North-Hatt Prestige Study in 1947. The study analyzed the attitudes of 2,920 respondents on 90 prestige occupational titles. Under these titles, the researchers provided a list of 3,000 occupational categories. On a survey, respondents ranked the 90 occupational titles from excellent to poor (Haug & Sussman, 1971). The participants also rated the occupational categories. Occupational categories with an excellent rating received a score of 100 whereas occupational categories with a poor rating received a score of two. A mean score for each occupation represented an occupation prestige rating. NORC replicated the study in 1963. The scores from 1947 and 1963 have a correlation coefficient of0.99 (Hodge, Siegel, & Rossi, 1964). Nakao & Treas (1994) later replicated the study in 1989. 18

PAGE 29

The 1947 study was the first study to move from small communities' opinions of occupational status to a national opinion on occupational status. The study also was the first to provide researchers with a statistical rationale for using the prestige of occupation model. However, Powers (1982) proposed that the study did not adequately deal with issues of validity. For instance, although the sample was larger than previous studies, some argued that the sample was unrepresentative of the U.S. population (Power). Additionally, one can question whether this sample included a representative sample of minorities. Another concern was that the 1947 North-Hatt study duplicated similar types of occupations. For instance, on the survey, respondents could rank the occupational titles for both doctor and physician (Nakao & Treas, 1994). In addition, the study did not represent the nearly 450 occupational categories presented by the U.S. Bureau of the Census (Haug & Sussman, 1971 ). The duplications of similar occupations and limited occupational categories forced the U.S. Bureau of the Census to create the occupational classification system (Nakao & Treas, p. 3). This system aggregated similar job titles, and in the process, reduced its own occupational categories. Duncan ( 1961) further refined the occupational prestige model by developing the Socioeconomic Index (SEI). The rationale behind Duncan's model came from Edwards' social economic groupings (Duncan, p. 115). In his approach, Edwards believed that education played an important role in the social status of workers and 19

PAGE 30

that income played an important role in workers' economic status. From Edwards' work, Duncan proposed that most people engaged in an occupation had formal education. He felt that most people pursued education to engage themselves in a particular occupation. He explained, "A man qualifies himself for occupational life by obtaining an education" (Duncan, p. 116). He further described income as the result of occupational and educational achievement. Duncan concluded that the connection between occupation, education, and income represented SES. Duncan ( 1961) then used Edward's theory to develop his SEI index model. He based the SEI model on four criteria. Duncan used the occupational prestige ratings from the North-Hatt study as his first criterion. He used the North-Hatt study's variables because the study was the only study to present statistical findings on occupational rankings. From the North-Hatt study's findings, Duncan selected only those occupations that received "good" and "excellent" ratings (p. 117). He chose occupations with higher rates because he believed these occupations presented the most comprehensive representation of people's attitudes toward occupations. Additionally, Duncan used a scatter plot diagram (p. 119) to show that the occupations rated "good" and "excellent" had an exceedingly close relationship (p. 118) whereas the occupations rated "average," "somewhat below average," and "poor" were not as statistically close. 20

PAGE 31

Duncan's (1961) predictor variables consisted ofhigh school completion and a reported income of at least $3,500 on the 1950 U.S. Census (p. 119). Duncan based the predictor variables on an age-standardized criterion and on males. He established an age-standardization criterion because he believed an adjustment for age on income and education was necessary. He believed that the adjustment was necessary because younger workers earned less than older workers did. He felt that because younger workers earning lower incomes would alter occupational prestige scores. In addition, he believed that because younger workers' education levels were lower than older, the educational data of younger workers would also alter occupational prestige scores. Additionally, Duncan (1961) believed that occupations held by younger workers trained them for future occupations. He did not think that these workers' maturity levels allowed them to be completely committed to their occupation at the same level as older workers. Duncan believed that younger workers did not represent the true relationship between income, education, and occupation, so he selected those males who were 16 years of age or older. Duncan ( 1961) next used the criterion males as the focal population for his study. He reported that other studies showed that in a household where both husband and wife worked, the husband's occupation was a better predictor of the families' social status. Although Duncan felt that in a dual-income household the wife's 21

PAGE 32

income was valuable to the families' standard of living and overall income, he found that the ratio of males in the work force outnumbered females 2.5 to I. Duncan's (1961) rationale for using the income of at least $3,500 and an education level of at least high school completion followed his same rationale for selecting occupations rated "good" and "excellent". Duncan believed that higher responses (e.g., "good" and "excellent") collectively represented a positive relationship between occupation, income, and education. For instance, he explained that when assessing the central tendency of median income and median education, this measurement was insensitive to incomes and education levels in the top 2%. Based on Duncan's belief that those with higher education, income, and occupation would present a more realistic relationship between education, income, and occupation, he was more interested in incomes and education levels that represented the upper 50% of the U.S. population. However, the 1950 U.S. Census showed that the median income in 1949 was roughly $3,000 and the median education level was 9th grade. After Duncan established his criterion for the dependent variable (occupations) and the predictor variables (earns at least $3,500 annually of males 16 years of age and high school completion of males over 16 years of age), he needed to ensure that the occupations from the North-Hatt study matched occupations listed on the 1950 U.S. Census. Duncan found that of the 90 occupational titles in the North-22

PAGE 33

Hatt study, only 45 occupations were similar to the occupations listed in the 1950 U.S. Census. The list of the 45 occupation titles each received two percentile ratings. The first rating represented the percentage of males with an income of at least $3,500 and who scored "good" and "excellent" for each of the North-Hatt occupational titles. The second rating represented the percentage of males who completed at least high school and who scored "good" or "excellent" for each of the North-Hatt occupational titles. For example, if the teaching occupation showed that 48% of males had a "good" or "excellent" rating and they made at least $3,500 annually, then the teaching occupational prestige score would be 48. In addition, if91% of the male teachers scoring "good" or "excellent" completed at least high school, then the education rating for teachers would be 91. Duncan ( 1961) used these two ratings to develop SEI scores. He calculated SEI scores using the multiple regression equation, SEI = 0.59 (income rating)+ 0.55 (education rating)6.0. The value 0.59 represented a standardized score of the income ratings for all occupations. The value represented the linear regression of income rates on all of the occupations. Duncan used this standardization to weight the differences in particular occupations. For instance, because Duncan accounted for males at least 16 years of age, males in particular occupations were lower (i.e., newspaper boy) than males in other occupations (i.e., railroad conductor). Duncan 23

PAGE 34

standardized these rates to account for the noticeable difference between occupations. He then used this same concept to standardize the education rate. The value 0.55 represented a standardized score of the education rating for all occupations. Finally, the value 6.0 represented the regression coefficient of income on education for all occupations. For example, the equation for the SEI score for teachers was 0.59(48) + 0.55(91)6.0. Therefore, the SEI for teachers equaled 72. In Duncan's SEI model, the values of the SEI ranged from 2 to 96. Critics disagreed with the methods by which Duncan (1961) measured SEI. For instance, Hugh & Sussman (1971) suggested that along with the small sample size, only selecting the upper level responses (i.e., "good" and "excellent") from the NorthHatt study reduced the public's opinion of the value of occupations. The researchers also suggested that not relying on actual scores also reduced the reliability of the study. However, Duncan showed the internal consistency of occupation, income, and education for the 45 occupational titles at 0. 75. He also explained that occupational titles between the North-Hatt study and the 1950 U.S. Census showed moderate reliability (p. 131 ). Duncan ( 1961 b) further explained that the SEI was not universally valid (p. 139) and discouraged the use of the tool for measuring social stratification. He believed that the major reason the measurement was not universal was that the sample represented a homogeneous group consisting of males who completed high school 24

PAGE 35

and made at least $3,500 a year. His argument was that given the complexity and multidimensionality of social stratification a single index of SES cannot be suitable for all research purposes. He further explained that given the complex dynamics of a stratified structure, a single index would only reflect a selected aspect of the structure. Duncan encourages researchers to use their only judgment when applying the SEI model to their research problem. Despite its limitations, researchers continue to use SEI both nationally and internationally. Blishen, Carroll, & Moore (1981) used this model to develop Canada's socioeconomic index. Additionally, New Zealand adopted this scale in structuring a socioeconomic measurement tool (Davis, McLeod, Ransom, & Ongley, 1997). National datasets (e.g., NELS:88) use Duncan's SEI model to calculate SES. The primary reason researchers continuously use SEI is that few competing models present a SES measure combining occupation, income, and education. In a continuous effort to perfect the SEI model, Stevens & Featherman (1981) used 1970 U.S. Census data to revise Duncan's SEI model. Stevens & Featherman selected those who had at least one year of college and earned at least $10,000 a year. Stevens & Featherman arrived at this income amount by using the Consumer Price Index. Economists used this index to account for inflation over time. Stevens & Featherman also used those occupations marked "good" and "excellent" from the North-Hatt's 1947 study. Instead of selecting only males, Stevens & Featherman 25

PAGE 36

created two indexes. One index represented males whereas the other index represented the total labor force. The study also accounted for the adjustment of education and income between 1961 and 1979. Stevens & Cho (1985) also revised the SEI using data from the 1980 U.S. Census. They found no statistical significance in the occupational classification changes in the U.S. Census between 1970 and 1980. Later studies showed that researchers varied in their use of the North-Hatt's occupational prestige model and the Duncan's SEI model (Hauser & Warren, 1997; Hodge et al., 1964; Nakao & Treas, 1994). The Hodge et al. replication of the study in 1964 supported the occupational prestige model. They found that the responses from 651 participants yielded similar results to the North-Hatt study, with a correlation coefficient of0.99. The only changes they made were in giving scores of 100 to those occupations receiving excellent responses, 80 for those receiving good, 60 for those receiving average; 40 for those receiving somewhat below average, and 20 for those receiving poor ratings. This study showed slight changes in the ratings of occupations but not in the overall rankings of occupations. Nakao & Treas (1994) again replicated both the occupational prestige study and Duncan's SEI study in 1989. They randomly assigned 1,537 participants into 12 sample groups of 125 participants. Ten groups rated occupational prestige whereas the other two groups rated ethnic prestige. The groups rated 7 40 occupational categories as opposed to 204 categories in 1964. The group rated 40 occupational 26

PAGE 37

titles as opposed to 45 titles in 1964. They selected 740 categories because these categories closely resembled the 503 occupational categories on the 1980 U.S. Census. The researchers rated the occupational categories by reducing occupational categories into one homogeneous job task. For instance, college professor was representative of sociology teachers, psychology teachers, and other college teachers. Alarmingly, Nakao & Treas did not explain the results of those groups who rated ethnic prestige. Once Nakao & Treas (1994) categorized the occupational titles and categories, they developed a SEI using an equation similar to Duncan. Like Duncan, they wanted to control for age. Instead of selecting males who were at least high school graduates, they selected both males and females with at least one year of college. Instead of a minimum income of $3,500, Nakao & Treas selected those with a minimum income of$15,000 in 1979. They based their income criterion on Stevens & Featherman's (1981) income criterion. Nakao & Treas used these criteria to develop a revised SEI index. Nakao & Treas (1994) compared their findings to those of other studies on prestige of occupations and SEI models. They explained that the correlation between their study and other studies ranged from 0.86 to 0.97. They reported that while there were statistically significant changes in ratings between occupational prestige scores, the public's opinion of occupations remained consistent. They concluded that 27

PAGE 38

between 1964 and 1989 the public's opinion of occupational status did not change significantly. The SEI model has contributed to the way researchers measure SES today. For instance, the NELS:88 has developed theSES composite variable based on father's income, mother's income, father's education, mother's education, and father's and mother's occupation as rated by the SEI model. TheSES composite variable is represented by a z-score if the combined socioeconomic indicators. In the NELS:88 base-year study, the NCES used the Stevens & Featherman (1981) revised SEI model. In the first-year follow up study, the NCES used the Nakao & Treas (1994) revised SEI model. This 1980 revised SEI model used to calculate SES ignores African Americans' relationship to income. For instance, Nakao & Treas (1994), in the latest revised SEI model, set the income criterion at least $15,000. The 1980 U.S. Census found that the median income of African Americans was $13,002. These findings show that the SEI model did not include a large portion of the U.S. African American population. Therefore, the current SES measurement may not be the best method for determining African Americans' socioeconomics. Socioeconomic Status and the African American Community Researchers suggest that socioeconomic status is a composite variable measured by education, income, and occupation as rated by SEI. Data (U.S. Bureau 28

PAGE 39

of Census 1980) showed that during the year of the latest SEI revision, education, income, and occupation varied between African Americans, European Americans, and the general U.S. population. Table 2.1 illustrates that 53% of the African American population, 73% of the European American population, and 70% of the general population earned at least $15,000 a year. The U.S. Bureau of Census (1980) found that in relationship to the educational criterion for SEI, there was a slight difference between African Americans, European Americans, and the general population. Table 2.1 also illustrates that 46% of African Americans between 18 and 24 years of age completed at least one year of college whereas 51% of European Americans and 51% of the general population completed at least one year of college. Table 2.1 shows that 20% of African Americans held white-collar jobs and 36% of European Americans held white-collar jobs. The table also shows that while African American white-collar workers earned an average of $20,500, European American white-collar workers earned an average of $29,985. Additionally, 23% of African Americans worked blue-collar jobs and 25% of European Americans worked blue-collar jobs. On average, African Americans working blue-collar jobs made $19,558 whereas European Americans working blue-collar jobs made $23,229. Table 2.1 shows a slight difference in the education level, income, and earnings related to occupations between African Americans and European Americans. Additionally, these data show that as the income of European Americans increased, 29

PAGE 40

their occupation status (e.g., from blue-collar to white-collar) also increased. On the other hand, the table shows that as African Americans moved from blue-collar to white-collar jobs, their income only increased $1,000. The table supports a possible strong positive correlation between income, education, and occupation for European Americans. However, African American increase in occupational status and the slight increase in income may not support a possible strong positive correlation between their income and occupation. Table 2.1 Percent of the education, income, and occupation types during 1980 SEI Criteria Occupation Types At Least One At Least Year Of $15,000 Per College Year White Collar Blue Collar African Americans 45.9 52.8 20.1 22.3 European Americans 51.4 72.7 35.8 24.6 General Population 51.1 73.0 34.2 24.4 Source: Data extracted from the 1980 U.S. Census Report The difference in the relationship between income and occupation of African American and European Americans brings to question which socioeconomic indicators best demonstrate the socioeconomic experience of these two populations. Some researchers argue that the variables closely associated with African Americans' economic experience are debt and their limited ability to gain substantial wealth (Conley, 1999; Kennickell, 2003; Shapiro, 2004). This argument derives from African Americans' common experience with poverty. Shapiro (2004) explains that 30

PAGE 41

while the European American population has experienced generational wealth, African Americans have experienced generational poverty. Brooks-Gunn et al. (1997) reported that the African American population is more likely to live in poverty longer than European Americans. They explain that if an African American family and a European American family entered poverty simultaneously, after five years, 50% of African Americans remain in poverty, while only 22% of European Americans remain in poverty. They propose that availability of community and fmancial resources promote the difference in the two ethnic groups. The African American population tends to live nearest to impoverished communities {Ireland, Sharp, & Steinmetz, 2003) and impoverished communities provide residents with few ways to enhance their fmancial situation or financial resources. Additionally, over the past 50 years, African Americans have represented the highest impoverished ethnic population in the United States (Alexis, 1998; Myers, Kim, & Mandala, 2004). In 1959, 59% of African Americans lived at or below poverty (Dalaker & Proctor, 2000). By 1966, a little more than one-third of African Americans lived at or below poverty. By 1997, 30% of African Americans lived at or below poverty (Dalaker & Proctor; Iceland, 2003). Now, roughly a quarter of African American households live at or below poverty. Storr (personal conversation, 31

PAGE 42

November 13, 2004) argues that the trend of fewer African Americans experiencing poverty is reflective of new wealth in the African American community. Wealth and the African American Community This section presents the four concepts wealth, assets, liability, and net worth along with the trend of African American wealth. Conley (1999) describes wealth as the accumulation of a family's property, assets, and net worth. Assets represent items owned whereas net worth measures the difference in assets and liabilities/debts (e.g., items owed on) (Shapiro, 2004; Orr, 2003). For instance, if a family home (assets) is worth $100,000 and they owe $75,000 (liabilities/debts) on that home, their net worth is $25,000. Researchers often refer to net worth as wealth (Shapiro, 2004; Orr, 2003). Now, despite new wealth established by African Americans, their wealth is lower and different than that of European Americans' wealth. For instance, Conley (1999) explains that in 1865, African Americans owned 0.5% of the total U.S. wealth and by 1990 African Americans owned 1% of the total U. S. wealth. Shapiro (2004) further explains that while European Americans' wealth tends to come from generational wealth, African Americans' wealth tends to come from acquired wealth from purchasing assets. In fact, between 1998 and 2001, 30% of African Americans' assets were in debt whereas 11% of European Americans' assets were in debt. Assets in debt refer to assets in which loans are still outstanding on those assets (e.g., loans on a home, boat, car). 32

PAGE 43

Rothstein (2004) further explains that the difference in the wealth of African Americans and European Americans draws a parallel to the difference in the net worth ofthe two groups' incomes. He explains that given a similar annual income, the net worth of African Americans' annual income is 64% of the annual income earned by an European American person. These fmdings suggest that when both African Americans and European Americans earn $10,000, African Americans benefit from only $6,400 of this income. In other words, European Americans utilize more of their annual income than African Americans. While the income of African American goes toward paying debt, the income ofEuropean Americans helps to establish their wealth. As further illustrated by Rothstein (2004), given similar annual income, the mean wealth of African Americans is equal to only 12% of the mean wealth of European Americans. These data suggest that European Americans have about 36% more use of their income and 88% more use of their wealth than African Americans. Another difference in the wealth of European Americans and African Americans is that European Americans are more likely to receive financial support from other family members (e.g., parents or grandparents) in purchasing their first home or in paying for college. African Americans are more likely to pay for their own schooling by taking out student loans. African Americans are often the first 33

PAGE 44

generation to buy a home and become middle class; and they frequently have no capital support (Rothstein, 2004). Choy's (2000) study researching college students' debt four years after graduation found differences in the percentage of graduate school loans borrowed by African Americans and European Americans. She found that African Americans borrowed money for their undergraduate degree at a rate 5% higher that did European Americans (see Table 2.2). Her findings showed that African Americans borrowed money for their graduate studies at a rate 8% higher that did European Americans. Additionally, Choy (2000) reported that after graduating with a bachelor's degree, African Americans' mean annual salary was nearly $3,000 lower than European Americans' mean annual salary (see Table 2.2). These numbers suggest that upon graduation, African American students left school with more debt and with an average income less than their European American counterparts. This difference in debt owed and income earned reduces African Americans' ability to establish wealth in comparison to European Americans. Table 2.2 Debt after college graduation Percent who received loans Undergraduate Graduate level level African Americans 51% 38% European Americans 47% 29% Source: Data retrieved from Choy (2000, pp. 25 & 27) 34 Income after college undergraduate degree $31,449 $34,164

PAGE 45

Oliver and Shapiro (1997) presented another example of African Americans' relationship to wealth. They explained that European American college graduates earn, on average, $18,000 more than European American high school graduates (Oliver & Shapiro, 1997, p. 110). Whereas European American college graduates net worth is $34,000 more than that of European American high school graduates (Oliver and Shapiro, 1997, p. 110). For African Americans, Oliver and Shapiro (1997) explained that college graduates make, on average, $17,000 more than high school graduates, and that college graduates' net worth is $14,000 more than that of high school graduates (p. 11 0). These data demonstrate that as African Americans' and European Americans' education increased, their annual household income increased. Additionally, these data show that as European Americans graduate from college, the net worth (assets minus debt) between high school and college graduates increases at a much higher rate than income between high school and college graduates. On the other hand, as African Americans graduate from college, the difference in the net worth between high school and college graduates increases at a lower rate than income between European American high school and college graduates. These data show that debt has a greater impact on African Americans who graduate from college than the other three groups. 35

PAGE 46

In another display of African Americans' relationship with wealth, Shapiro (2004) explained the interaction between African Americans', Spanish-speakers', and European Americans' wealth and SES. In his study, Shapiro developed the concept asset poverty level (APL) (p. 37) to assess the difference between the groups' wealth. Shapiro computed the APL by multiplying the U.S. poverty line by three. The U.S. Bureau of the Census (2002) data showed that the poverty line for a family of four was $1,392 per month. When multiplying $1,392 by three, the APL for a family of four would be $4,176. Shapiro explained that the idea behind this concept was that the APL determined the assets a family needed to meet their basic needs over a temporary period. He considered families below the APL asset-poor (p. 38). Shapiro (2004) examined the relationship between APL and ethnicity by interviewing similar-income African American, European American, and Spanish speaking families from Boston, Los Angeles, and St. Louis. He found that in 1984, 67% of African American households lived below the APL (p. 38). In 1999, 54% of African American households were below the APL. Although the data show a positive trend for African Americans, the wealth gap between African Americans and European Americans was still noticeable. Shapiro further explained that while 52% of African American families lived below the APL, 26% of European Americans lived below the APL (p. 40). 36

PAGE 47

Orr (2003) refers to the assets described in the APL model as liquid assets. She explains that converting these types of assets into cash is easier than converting illiquid assets into cash. Orr explains that income produces liquid assets whereas illiquid assets are non-income produced. She explains that CD's, stocks, and savings accounts are products of liquid assets (p. 287) whereas Conley (1999) explains that trust-funds, inheritances, vehicles, and real estate are non-income produced illiquid assets. Oliver and Shapiro (1995) explain that the African American population is less likely than European Americans to acquire liquid assets. Conley (1999) further explains that liquid assets are useful when families are in crisis (e.g. unemployment, medical emergency). New Orleans citizens' experience with Hurricane Katrina is an example of the value of liquid assets. While families with liquid assets could purchase gas, meals, and hotel costs in another state, or obtain transportation needed to leave New Orleans, those families without liquid assets could not purchase gas or transportation. Conley explains that illiquid assets are also valuable because they provide for families' basic needs. For instance, a car or a home serves families' basic needs. He explains that illiquid assets provide families with psychological security and can function as a status symbol (Conley, p. 28). While liquid assets can help in times of crisis, illiquid assets represent status. For instance, the neighborhood the family's home resides in or the type of car the family drives can represent a status symbol. 37

PAGE 48

Academic Outcomes and SES Although findings support African Americans' common experience with wealth, researchers insist on using SES to explain the relationship between their economic status and academic outcomes. This section presents fmdings related to academic outcomes. As proposed in the introduction of this study, the factors academic engagement, academic achievement, and academic attainment represent academic outcomes. The following data describe the relationship between SES, academic engagement, academic achievement, and academic attainment. Engagement Finn's (1989) participation-identification model (p. 118) presents four levels of participation. Level-one participation (p. 128) suggests that students simply behave. This level represents students who perform basic classroom behaviors such as class participation, attentiveness in class, and following the teacher's directions. The second level of motivation, suggests that students initiate interactions with teachers and are willing to do extra schoolwork before, during, and after school. The third level of participation represents students' involvement in extracurricular activities. This level includes activities such as choir, athletics, and social clubs. The fourth level of participation refers to students who take part in school governance. These activities include students setting academic goals and playing a role in regulating disciplinary actions (Finn, p. 129). 38

PAGE 49

This current study measures engagement differently than Finn (1989). Whereas Finn measures levels of engagement, this study measures types of engagement. The rationale this study takes is based on the belief that just because a student participate the second level engagement does not mean the student also participates in first level engagement. For instances, just because a student participates in athletics, that does not mean that student participated in basic school activities. This study proposes that just because a student participates in higher-level engagement activities does not ensure that lower levels of engagement are completed. This approach leads this current study to measure types of academic engagement, not levels of academic engagement. Further, this current study presents not engaged as the first type of engagement. This type of academic engagement suggests that students are not involved in the any type of engagement. The second type of engagement reflects Finn's (1989) first level of participation. This type of engagement suggests that students participate in school at the basic level. Student participation includes attending school, doing homework, and being attentive in class. Finn's (1989) second level of participation focuses mainly on student motivation. Dev (1997) and Finn (1989) explain that students' motivation is highly linked to dropout and student success rates. Sometimes, teachers do not have control over students' motivation. While teachers may or may not provide an inviting 39

PAGE 50

classroom, students' motivation to engage can relate to factors not associated with school. For instance, Ogbu's (2003) Shaker Height study fmds more disengagement among African American students from both affluent and low-income communities than among European American students from similar communities. However, the African American students in this study are in the minority and few minority teachers are in the district. Here, the motivation of these students could be a result of how these students respond to their minority status and the low percentage of minorities on staff. As Finn & Voelkl (1993) found that African American students show a higher motivation when a greater percentage of minorities attend their school and when a greater percentage of African Americans work in the school. Another approach to students' motivation came from the Roscingo and Ainsworth-Darnell (1999) study on cultural capital (p. 1). They argue that the basis of students' motivation came from their investment in cultural capital experiences such as access to more cultured activities. The study proposes that because European Americans have more access to cultural capital, their engagement (motivation) is higher than that of African Americans. However, Davis argues that all human activities are cultured (email conversation, July, 27, 2005). He suggests that some cultural activities are reflective of mainstream culture, and some cultured activities are not reflective of mainstream culture. In response to Davis' (2005a) argument, it is 40

PAGE 51

uncertain whether mainstream cultured experiences motivate students differently than non-mainstream cultural experiences. This current study proposes that students' motivation in school can result from various factors. Therefore, this study does not focus on what schools are doing to educate African American students. This study is more interested in how African American students respond to school. Therefore, this study does not focus on what Finn (1989) describes as the second level participation. In this study, the third type of engagement refers to students' participation in extracurricular activities. This type of engagement is similar to Finn's (1989) third level of participation in that this type of engagement measures students' involvement in sports as well as social and academic clubs. This study's fourth type of engagement refers to students' engagement in activities that prepare them for college. This measurement includes students' involvement in college preparatory work, such as taking Advance Placement (AP) tests, Pre-Scholastic Assessment Tests (Pre-SAT), Scholastic Assessment Tests (SAT) or American College Tests (ACT). Studies (Johnson, Crosnoe, & Elder, 2001; Ainsworth-Damell & Downey, 1998) show that African Americans engaged in more basic school activities than European Americans and Spanish-speakers. Lee and Smith's (1995) work also finds a positive relationship between minority status and basic school participation. AinsworthDarnell & Downey conclude that while African American students spent 41

PAGE 52

less time on homework than European American students, African American students worked harder in class than European American students. Additionally, Johnson, Crosnoe, & Elder (2001) found that African Americans show less attachment to school than European American and Spanish-speaking students. On the other hand, their study show that African Americans engage in significantly more activities such as going to school, paying attention in class, and doing homework than the other two ethnic groups. Smerdon ( 1999) findings no significant difference between students' attendance, time spent on homework, and ethnicity. Finn and Voelke (1993) explain that while minority students attend school at a higher rate than European American students, they show less attachment to school. They also explain that African American students show less emotional connection to their school than did European American students. Data on the third type of engagement shows that students who engage in extracurricular activities perform better in school than students who do not engage in extracurricular activities. The National Center for Educational Statistics (NCES, 1995) reports that students who take part in extracurricular activities are more academically successful than did students who do not take part in extracurricular activities. The report (NCES) explains that students who take part in extracurricular activities are 20% more likely to have a 3.0 grade point average. The report (NCES) also shows that students who take part in extracurricular activities score higher on 42

PAGE 53

standardized reading and mathematic test than those who do not take part in extracurricular activities. Additionally, the report (NCES, 1995) presents that students who take part in extracurricular activities are 20% more likely to graduate with at least a bachelor's degree. Other findings from the report show that students from low SES households take part in extracurricular activities less than students from middle SES and high SES households. Table 2.3 demonstrates that while 80% of all high school students take part in extracurricular activities, students from low SES households take part in extracurricular activities 13% less than do students from high SES households. Table 2.3 also shows that students from low SES households who attend less affluent schools take part in extracurricular activities slightly more than do students from low SES households who attend affluent schools. Additionally, the table illustrates that students from low SES households who attend less affluent schools take part in vocational/professional clubs more often than do students from high SES households. The NCES (1995) report further shows no difference in the percentage of available extracurricular activities offered to students from more affluent and less affluent schools. Data from the NLES:88 first and second-year follow up survey shows that while African American and European American students participate in athletic activities at a similar rate, African Americans take part in performing arts activities 43

PAGE 54

(e.g., cheerleader, band/orchestra, plays, musicals) at a higher rate than do European American students. On the other hand, European Americans take part in school government/clubs activities at a slightly higher rate than African Americans. Table 2.3 Percentage of students who took part in extracurricular activities in 1992 Low Income High Income Less More Less More All Affluent Affluent Affluent Affluent students Schools Schools Schools Schools Any Activity 79.9 74.7 73.0 86.8 87.6 Academic Clubs 26.2 2.2 2.5 36.2 32.3 Vocational Clubs 2.8 29.2 25.6 16.0 ll.8 Source: Data retrieved from NCES (1995) Table 2.4 Percentage by race and participated in extracurricular activities in 1992 Activities African Americans European Americans Athletics 53.4 53.1 Performing Arts 59.0 43.1 School Government/Clubs 37.0 35.0 Source: Data retrieved from table 114 ofthe NCES's Digest ofEducation Statistics Tables and Figure In examining the role of extracurricular activities on African American students' high school experience, Jordan (1999) finds that African American students who take part in school-sponsored sports activities perform better in school and engage in academic activities at a higher rate than do those African American students who do not take part in extracurricular activities. Jordan (1999) also explains that when controlling for SES, the results do not change significantly. Jordan (1999) presents that 18% of African American students take part in team sports and 10% of African American students take part in individual sports by lOth grade. For reasons 44

PAGE 55

such as living in impoverished communities, lack of funding, and the lack of engagement, 82% to 90% of the African American students do not take advantage of school-sponsored sport activities. Data also show that African American students engage in the fourth type of engagement at a lower rate than European American students. NCES (2002) reports that in 2000 41% of Mrican American students took AP science courses compared to 43% of European American students. The data show that between 1984 and 2000, African American students took less AP examinations than Spanish-speaking and European American students. The percentage of African American students who took AP courses during this time grew from nine in every 1000 students to 53 in every 1000 students. However, African American students continue to take AP examinations at a much lower rate than European American and Spanish-speaking students. For instance, in 1984, European Americans took AP courses at a rate six times higher than African Americans. In 1997, European Americans took AP courses at a rate 3.5 times higher than African Americans (NCES, 2002). The Journal of Blacks in Higher Education (2005) explains that several barriers contribute to these African American students' low AP scores. The journal reports that while predominantly European American suburban schools prepare their students in lower grades to take AP courses, students from largely minority integrated communities do not receive adequate preparation to take AP courses. Additionally, 45

PAGE 56

the journal explains that many schools with large African American populations steer these students to vocational courses instead of AP courses. High School and Beyond (HS&B) data show that between 1980 and 1992, African Americans took part in vocational clubs at a higher rate than any other ethnic group. Another area related to the fourth type of engagement is college entry exams. In 2001, 11% of African Americans and 66% ofEuropean Americans took the SAT. African Americans scored lower than any other ethnic group on the verbal and mathematic sections of the SAT. African Americans scored 96 points lower on the verbal section and 105 points lower on the mathematics section than their European American counterparts (Hoffman, Llagas, & Snyder, 2003). Table 2.5 shows that African Americans also scored lower on the subject areas and the overall composite score of the ACT than any other ethnic group (Hoffman et al., 2003). As explained, research has found that while European American students show more attachment to schools, African Americans perform higher in the second type of engagement. African American and European American students took part in the third type of engagement at a similar rate. European Americans students show more involvement at the fourth type of engagement. Those African Americans who took part in extracurricular activities perform better in school than those African Americans who did not take part in extracurricular activities. 46

PAGE 57

Table 2.5 Mean ACT scores for 2001 Composite English Math Reading Science Scores Scores Scores Scores Scores African Americans 16.9 16.2 16.8 16.9 17.2 European Americans 21.8 21.3 21.3 22.2 21.8 Mexicans 18.5 17.5 18.7 18.6 18.8 Asian Americans 21.7 20.7 23.1 21.1 21.5 Native Americans 18.8 17.8 18.4 19.2 19.3 Source: Hoffman et al, 2003, p. 65 Achievement Hedges & Nowell (1998) examined the historical relationship between the academic achievement of African American and European American students (see Table 2.6). More recently, the National Assessment of Educational Progress (NAEP) data indicated that African American students' basic reading and mathematics scores were lower than that of European American, Native American, or Spanish-speaking students. Between 1992 and 2003, the percentage of African American eighth graders reading at or above the basic level increased (see Table 2.6). In addition, other ethnic groups score also increased. However, in comparison with other ethnic groups, African American eighth graders consistently scored lower. Whereas the percentage of European American students reading at or above the basic level increased 6%; Spanish speakers increased 3%; Native Americans increased 4%; and Asians increased 31%. This trend also is consistent with mathematic scores (NAEP, 2005b). In addition, between 1992 and 2003, the percentage of African American eighth graders at or above the basic mathematic level increased 6%. In comparison, the 47

PAGE 58

percentage of European American eighth graders at or above the basic mathematic level increased 12%; Spanish speakers increased 13%; and Asians increased 2%. Table 2. 7 illustrates that this pattern has been consistent across age groups. The table shows that reading achievement scores among nine-year-old European American students were 26 points higher than scores among nine-year-old African American students. In addition, 13 year-old European American students are 22 points higher than their African American counterparts, and 17 year-old European Americans are 29 points higher than African American students of the same age. When comparing African American students' mathematics achievement scores to European Americans' scores, Table 2.7 shows that European American 9 year-olds score 23 points higher, 13 year-olds score 37 points higher, and 17 year-olds score 28 points higher. Table 2.6 Percentage of reading and mathematics of 8th graders in at or above the basic level African Americans European Americans Spanish Speakers Native Americans Asian Americans Source: NCES (2005) Reading 1992 2003 33 42 71 77 44 47 53 59 41 72 Mathematics 1992 2003 20 39 68 80 35 48 76 78 Researchers present different findings in the relationships between SES and the academic outcomes of African Americans. Battle (1997) used the NELS:88 to examine the relationship between SES, marital status, and African American 48

PAGE 59

students' achievement. He shows that when controlling for SES, no significant relationship existed between academic achievement scores for students from divorced and married households. Additionally, Battle (1997) also presents that the academic achievement of African American students from divorced families do not benefit from increased SES. He proposes that African American students from married families do benefit from increased SES. These findings suggest that achievement test scores of students living in higher SES households are higher than achievement test scores of students living in lower SES households. In addition, the study presents that those students from low-income divorced households score significantly higher on standardized tests than students from low-income married households. Battle (1997) suggests that it is unknown if academic test scores among African American students from married and divorced households could be attributed to SES. Researchers present different findings in the relationships between SES and the academic outcomes of African Americans. Battle (1997) used the NELS:88 to examine the relationship between SES, marital status, and African American students' achievement. He shows that when controlling for SES, no significant relationship existed between academic achievement scores for students from divorced and married households. Additionally, Battle (1997) also presents that the academic achievement scores of African American students from divorced families does not increase when SES increases. He proposes that that the academic achievement scores 49

PAGE 60

Table 2.7 Achievement scores between African Americans and Euro,eean Americans Reading Mathematics 1971 1992 2004 1971 1992 2004 African Americans 9 Years-Old 170 185 200 190 208 224 13 Years-Old 222 238 244 228 250 262 17 Years-Old 239 261 264 270 286 285 European Americans 9 Years-Old 214 218 226 225 235 247 13 Years-Old 261 266 266 274 279 288 17 Years-Old 291 297 293 310 312 313 Source: NCES (2005) of African American students from married families do increase as SES increases. These findings suggest that achievement test scores of students living in higher SES households are higher than achievement test scores of students living in lower SES households. In addition, the study presents that those students from low-income divorced households score significantly higher on standardized tests than students from low-income married households. Battle (1997) suggests that it is unknown if academic test scores among African American students from married and divorced households could be attributed to SES. Smith-Maddox (1998) also used the NELS:88 data to study the various effects of culture on African American students' standardized mathematics scores. In their study, Smith-Maddox measured the cultural attitudes (aspirations) and behaviors (home habits and activities) of students, teachers, parents, and schools. One of the variables Smith-Maddox analyzed was the influence of SES on cultural dimensions 50

PAGE 61

and mathematics. Cultural dimensions represented the interaction between student, family, school, and teacher. Smith-Maddox found that SES had a significant positive effect on academic achievement. Out of all of the variables, SES produced the strongest relationship to academic achievement. She also assessed culturally relevant context by asking students if their teacher taught ethnicity in their history, government, or social studies classes. She found that when controlling for SES, the weakest relationship was between culturally relevant context and African Americans' academic performance. Adams & Singh (1998) used NELS:88 data to examine the interrelationship among African American students' prior achievement, their perceptions of their school environment, their awareness of quality of instruction, their aspirations, their motivation, their gender, their parent's aspirations for them, their parent's involvement in their education, and SES. They explain that a significant correlation between SES and these variables exist. They conclude that students from higher SES households score higher on standardized tests than students from low SES households. Adams and Singh discuss no other significant relationships. Adams and Singh (1998) explain that findings from other studies (Donovan, 1984; Fehrmann, Keith, & Reimers, 1987) show low correlations or indirect effects between SES and African American students' academic achievement. They argue that the contradictions in the fmdings resulted from the method used to measure academic 51

PAGE 62

achievement. They suggest that those studies that used self-reported grades as a measure of academic achievement showed no significant correlation between SES and academic achievement whereas the studies that used standardized tests as a measure of academic achievement showed significant correlation between SES and academic achievement. However, as explained earlier, Battle (1997) showed that when analyzing martial status, no significant correlation existed between SES and the standardized test scores of African American students. A recent study by Carter (2003) supports Adams and Singh's (1998) position. In this study, Carter used the NELS:88 data to assess the influence of family, community, and school on African American and Hispanic youth's academic achievement. Carter fmds that family SES is not predictive of students' grades. She also fmds that SES is predictive of students' standardized test scores. Her study suggests that the lower the student's income, the lower the student's standardized test score. As mentioned, some research suggests that when a significant correlation between SES and academic achievement is found, those who live in higher SES households perform academically better in comparison to those with lower SES households. However, Craig, Connor, & Washington's (2003) study shows the opposite effect. They examined the performance of 50 African American preschoolers and kindergartners on a reading comprehension test. The preschoolers 52

PAGE 63

resided in low-income households, and the kindergarteners resided in middle income households. They examined the two groups' oral language and cognitive skills at the end of the students' first and third grade years. The data show that no significant difference existed in the two groups' scores at the end of their first grade year. The data show that the preschoolers (students from a low-income household) score significantly higher than the kindergartners (students from middle-income households). Craig, Connor, & Washington explains that if those students in the kindergarten group had prior preschool experience and if the kindergarteners did not have prior preschool experience, this could explain the difference in the two groups' scores. Ogbu's (2003) ethnographic study of Shaker Heights High School examined engagement of African American students in an affluent middle-class community. He concluded that after six months of interacting with the parents, teachers, and students, African American students disengaged from school because of a lack of parental involvement, parental contact with the school, and student motivation. Ogbu (2003) suggests that although the achievement scores of African American students in the Shaker Heights School District are higher than the achievement scores of African Americans in the rest of the nation, a significant gap remain between their scores and the scores of European Americans in their district. Ogbu explains that given similar SES, the African American students in Shaker 53

PAGE 64

Heights score significantly lower on achievement tests than the European American students. He describes that European Americans from high-SES households do better than African Americans from high-SES households, and European Americans from low-SES households score significantly better than African Americans from low-SES households. Although Ogbu (2003) explains that African Americans in Shaker Heights scored higher than African Americans in the state of Ohio and across the county, he does not compare SES and achievement of Mrican Americans from Shaker Heights and African Americans from other communities. He does propose that, regardless of their community (e.g., suburban or urban), African American students from affluent communities struggle with the same achievement gaps as African American students from low-SES inner-city communities. In examining the relationship between SES, other social factors, and African Americans' academic achievement, findings are inconsistent on the nature ofthese relationships. When researchers control for SES by making the variable constant, inconsistency still occurs. Additionally, research on academic achievement explains that there is a significant difference in achievement scores, SES, and ethnicity. However, based on Johnson's (1992) argument, these findings could be a result of SES being more representative of European Americans' socioeconomic experiences and less representative of African Americans' socioeconomic experiences. 54

PAGE 65

Below, data describes the relationship between Afiican American students and educational attainment. Kaufman, Atl, & Chapman (2004) used data from the 2000 U.S. Census to show that 87% of the population between ages 18 and 24 completed high school, an equivalent credential, or aGED. Ninety-two percent of European Americans and 84% of African Americans between the ages 18 and 24 completed high school, an equivalent credential, or aGED. Current census data (2005) showed that between the ages of 18 and 19, 56% of the general population, 48% of African Americans and 58% of European Americans, completed high school, an equivalent credential, or a GED in2004. Swanson (2004) disagrees with these findings. He explains that some calculations and datasets tend to overestimate graduation rates. Swanson explains either that these datasets represent a small portion of the U.S. student population or that calculations for graduation rates do not consider migration and change in localized communities. In his study, Swanson (2004) used the Common Core of Data (CCD) to compute a national graduation rate. The CCD is a collection of state reported data produced by NCES to track enrollments, dropouts, graduation, and attainment of students from 9th, lOth, 11th' and 12th grades. Swanson (2004) used the CCD because it represented nearly 85% ofthe U.S. school districts and 93% of the U.S. public school population (p. 33). 56

PAGE 66

Swanson (2004) used the Cumulative Promotion Index (CPI) to estimate the four-year graduation rate of students from the CCD. The CPI estimated the graduation rate by assessing the enrollment of 9th, 1Oth, 11th, and 12th graders over a four-year period against diplomas earned. Based on NCLB (2001) and actual diplomas earned, the calculation did not consider those who earned an equivalent credential or aGED. Swanson's fmdings showed that 50% of African Americans, 76% of European Americans, and 68% of the U.S. student population graduated with a high school diploma within four years. Swanson (2004) explains that the use ofCPI to calculate graduation rates from CCD is more comprehensive than other analyses because the CCD represents 93% of the national student population. He explains that the Basic Completion Ratio (BCR) represented 99% of the national student population, the National Center for Education Statistics-G (NCES-G) represented 45% of the national student population, and the Serial Persistence Rate (SPR) represented 45% of the national student population. Swanson (2004) strongly believed that the CPI calculation presented the most efficient estimation of high school graduation rates. He compared the CPI method to the BCR, NCES-G, and SPR. Swanson found that the CPI method showed a national graduation rate of 68%. This rate was the same as the BCR ( 68%) but lower than the NCES-G (800/o) and the SPR (82%). The difference in these findings is a result of how the different methods calculate graduation rates. The BCR calculates graduation 57

PAGE 67

rates by dividing the number of students who received a high school diploma by the number of students who were in the 9th grade four years earlier. The BCR calculates graduation rates by dividing the number of students who received a high school diploma over the number of students who were in the 9th grade four years earlier. The NSEC-G calculates graduation rates by the number of students who received a high school diploma by the number of students who dropped out of the 12th grade. The SPR calculates graduation rates by the difference in dropouts over a four-year period. Swanson's CPI and the BCR do not account for dropout rates, but the NECS-G and the SRP do account for dropout rates in calculating graduation rates. In observing the dropout rates of the other approaches, Swanson's findings seem conservative or to underestimate the national graduation rates. For instances, Davis (2005) study of Denver Public schools found African Americans to graduate 6% higher than Swanson, while he found European American students to graduate at a similar rate as Swanson. A concern with Swanson's (2004) study is that not all states consistently report to CCD. Swanson explains that the problem with accounting for dropout rates is that states did not consistently report data (p. 36). However, Hoffman, of the NCES, explains that when states did not report data on dropout or enrollment, the report did not include findings for that state (personal phone conversation, August 22, 2005). Young (2003) used the CCD and presented data on dropouts and four-year 58

PAGE 68

completion by ethnicity and state. The data represented 38 states. This occurred because only 38 states reported data for four consecutive years. The report explained that to calculate a four-year graduation rate, states needed to report four consecutive years of data. Young's (2003) report did not present data for Texas, Washington DC, Florida, North Carolina, and South Carolina. The problem with not reporting these states' data is that these states represent a high percentage of the U.S. African American population. The 2002 U.S. Census found that Texas and Florida were two of the four states with two million or more African Americans. Between the years 2000 and 2003, more African Americans moved to Florida than to any other state. Additionally, 60% of the population in Washington DC is African American, and South Carolina has the second highest African American population. U.S. Census (2002) data also showed that in 2002, 55% of the African American population lived in the south. Not accounting for these states' data reduces the CCD accuracy in calculating actual dropout rates. These findings may explain why Swanson's (2004) CPI method did not account for dropouts when calculating graduation rates. Although Swanson (2004) did not account for dropout rates, some state and local level findings confirmed his graduation estimation rates. For instance, Helfand (2005) found that in the Los Angeles School District, the graduation rate for African Americans was 49% and the rate for European Americans was 67%. Davis (2005) 59

PAGE 69

found that in the Denver Public School District, the graduation rate for African Americans was 56% and European Americans' graduation rate was 75%. Davis analysis relied on a four-year cohort and omitted all students who transferred out of the district during that time. Asimov (2005) questioned Swanson's (2004) findings, because he did not account for transfers or dropouts. The number of students transferring in or out of schools and who graduate on time would underestimate or overestimate the graduation rate. These differences in estimations depended on the level in which schools or districts were experiencing student transfers. Because the NCLB (200 1) policy specifies graduation in four years and allows students to transfer from failing schools to passing schools, student transfers would affect a four-year cohort. Additionally, by not accounting for dropouts, again, the graduation rate could represent underestimated or overestimated rates. In understanding enrollment related to a four-year cohort, it seems that dropouts and transfers would affect the year-by-year change in enrollment. Hoffman (personal phone conversation, August, 22 2005) explains that the U.S. Department of Education is developing an Average Freshman Graduation Rate. She did not make clear if this rate would account for transfers but did explain that the rate will account for dropouts and graduation over a four-year span. These limitations in the findings led to uncertainty regarding the relationship between African Americans' graduation rates and SES. However, NELS:88 data 60

PAGE 70

showed that when rating SES by the lower 25%, middle 50%, and upper 25%, as SES increased, so did high school attainment. From these data, 31% of the bottom SES, 50% of the middle SES, and 74% of the top SES graduated with a high school diploma, GED, or certificate. NCES found that six years after expected graduation, 42% of those in the bottom SES had not enrolled and were not working toward a diploma or equivalency. Wealth and Academic Outcomes Conley (1999) explains that wealth impacts the academic outcomes of students because it provides families opportunities to advance their child's education. He explains that wealth provides families the opportunity to purchase homes in neighborhoods with higher property taxes. In turn, these communities' taxes lead to resources used to enhance students' academic careers. As mentioned earlier in this chapter, AP courses are more readily available to students in schools in affluent communities because the cost of AP courses can range for $25 to $400 per course (NCES, 1995). Additionally, wealth allows parents the opportunity to assist in the cost of educational materials, tutors, computers, and private schools. Three studies gained insights into the relationship between academic outcomes and wealth. Phillips, Brooks-Gunn, Duncan, Klebanov, & Crane ( 1998) failed to find an effect of wealth on academic achievement among five and six-year olds. While Conley (1999) found an effect of wealth on educational attainment for 61

PAGE 71

both high school students and college students, Orr (2003) found that wealth explained a portion of the difference between African Americans' and European Americans' (ages 5-14) academic outcomes. First, Philips et al. (1998) used the Children of the National Longitudinal Survey of Youth (CHLSY) to examine the family background, parenting practice, and gap between five and six-year old African American and European American students' Peabody Picture Vocabulary Test-Revised (PPVT-R) scores. Part of their study analyzed the effects of wealth on these students' scores. From respondents' answers to questions regarding the value of their assets and debt, they developed the variable, parental wealth (p. 141). They found that wealth had a small and non statistically significant effect on the gap between the scores of African American and European American students. Philips et al. concluded that though the study did not produce significant effects, examining the relationship between wealth and education on families with older children might produce different results. They explained that parents of older children could have more forms of wealth (e.g., savings) later in life than when their children were younger. Conley (1999) used the Panel Study of Income Dynamics (PSID) data collected by the Institute for Social Research at the University of Michigan to examine the relationship between wealth and African Americans' educational attainment. Conley performed three analyses. The first analysis examined the effects 62

PAGE 72

of being African American and high school graduation. This analysis consisted of five models, parental assets type, parental net worth, socioeconomic background (as measured by SEI), races and individual characteristics, and race only (Conley, p. 48, 68). The second analysis examined the relative effects of race and class on the odds of high graduation. The five models in this analysis were parental net wealth, parental income, parental occupational prestige, parental education, and being African American (Conley, p. 48, 71). The third analysis examined the effects of parental wealth on the odds of high graduation. The five models in this analysis were being African American, other illiquid assets, value of owned business, primary residence equity, and liquid assets (Conley, p. 49, 71). In the first analysis, Conley ( 1999) removed explanatory factors such as family structure and family income. He found no difference in high school educational attainment between African Americans and European Americans when measuring for the effects of being African American on the odds of high school graduation and parental asset types. He found that African Americans showed a non significant increased chance of high school completion than European Americans when examining the effects of being African American on the odds of high school graduation and parental net worth. Conley (1999) found that African Americans were 2.60 times more likely to complete high school than European Americans when examining the effects ofbeing African American on the odds of high school 63

PAGE 73

graduation and socioeconomic background (p. 68). He also found that the European Americans sampled were more likely to complete high school than did their African American counterparts when examining the effects of being African American on the odds of high school graduation. The second analysis showed that the strongest effect came from the relative effects of race and class on the odds of high school graduation and parental education. The effects of race and class on the odds of high school graduation and parental income showed a slight effect. This finding suggests that when considering race, class, and parental income, the higher the parental income, the higher high school attainment. The third analysis found that the stronger predictor of high school completion was parental wealth and liquid assets. Conley explained that the value of the families' business showed significance at 0.01. Conley (1999) concluded that given the strong effect of parental assets, parental wealth, and liquid assets, parents are better able to fund their students' education through assets than through income. He explained that when the analysis did not account for wealth, income represented the major factor for predicting educational attainment. Additionally, Conley believes that the findings showed that liquid assets mattered more than both wealth (net worth) and income. Finally, Orr (2003) used the National Longitudinal Survey of Youth (NLSY :79) dataset to assess the relationship between SES, wealth, and the academic 64

PAGE 74

gap between African Americans and European Americans ages 5 to 14. In examining this relationship, Orr used the mathematics portion of the Peabody Individual Achievement Test (PlAT) as the measure of academic achievement. To analyze the relationship between SES, wealth, and academic achievement, Orr developed four models. The first model analyzed only race on the PIA T Mathematics Assessment. Model 2 added the control for SES, measured by SEI, to the model. The third model added wealth (total net worth). Model4 added educational resources, cultural capital, self-esteem, and social capital. Of concern to this study are Models 1, 2, and 3. Orr (2003) found that African Americans scored statistically a significant 8.34 points lower than European Americans when analyzing the relationship between race and PlAT mathematics scores. Orr found that students from large families and/or with older mothers scored lower on the PlAT mathematics test when adding and controlling for SES. Orr also found that students scored higher on mathematics achievement tests when their parents had higher education and a more prestigious occupation. Data associated with Model 2 showed that the scores between African Americans and European Americans decreased to a statistically significant difference of 6.52. Orr (2003) added wealth and controlled for SES in Model3. Orr found that regardless of parental income, education, and occupation, students from families with little or no wealth scored lower on the PlAT mathematics test than those students in 65

PAGE 75

families with wealth. Orr further explained that wealth (0. 77) had a larger effect on PlAT mathematics scores than parental education (0.57) and occupation (0.58). Data from this model showed that the difference in the PIA T mathematics scores between African Americans and European Americans decreased to a statistically significant 4.98 points. However, the data showed that as the gap in wealth increased, so did the gap in PlAT mathematics scores. When analyzing PlAT Reading Comprehension and Reading Recognition scores, the researcher also found the same results between wealth and achievement. This study explained that a relationship between wealth and African American students presents significant findings. While Philips et al. (1998) found no significant difference in five and six-year old African American and European American students' PPVT-R, Orr (2003) explained that wealth accounts for some portion of the difference in PlAT mathematics scores between African American and European American students. Responses to both studies suggest that analyzing the relationship between wealth and students' academics may produce different results when analyzing older students. Conley ( 1999) explains that in relationship to African Americans, wealth, in particular liquid assets, showed to have a strong effect on high school attainment. 66

PAGE 76

CHAPTER THREE Methodology As presented in the theoretical framework and the literature review, when measuring the relationship between the aggregated SES composite variable and the academic outcomes of African Americans, researchers have presented inconsistent findings. The inconsistencies may be a result of using the SEI model to construct the SES variable. As presented, when developing the SEI model, researchers have historically ignored the socioeconomic experiences of African Americans. Yet, research supports a historical correlation between the SES composite variables, income, education, and occupation. Additionally, Conley (1999) and Orr (2003) argue that the socioeconomic indicator wealth is a more culturally relevant predictor of African Americans' academic attainment and achievement. In this chapter, this study explains the descriptive relationship between the NELS:88's African American and European American population's academic engagement, achievement, and attainment. The chapter also presents the descriptive relationship between the variables wealth, disaggregated SES, household income, parents' education, and parents' occupation for the NELS:88's African American population. Finally, factor analyses, multiple regression analyses, and comparison analyses assess the relationship between the academic engagement, achievement, and attainment of African American students and European American students. 67

PAGE 77

Hypotheses The following hypotheses concerning the relationship between the academic outcomes of African American and European American students were tested. 1. Socioeconomic indicators related to African Americans predict academic engagement, academic achievement, and academic attainment for African American and European American students differently. 2. Socioeconomic indicators related to European Americans predict academic engagement, academic achievement, and academic attainment for both African American and European American students differently. Dataset Data for this study came from the public-use data file of the National Educational Longitudinal Study of 1988 (NELS:88). The NELS:88 is a longitudinal survey study sponsored by the U.S. Department of Education's National Center for Education Statistics (NCES). The initial NELS:88 sample came from 24,599 8th graders enrolled in 1,052 public and private schools across the United States in 1988. Of the 24,599 students, the study randomly selected an average of 26 students from the 1,052 schools. During 1988, the base year, a parent survey, a teacher survey, a school administrator survey, and a student survey collected data on students' 8th grade experience. The participants provided information related to student background, family background, community, school experiences, self-reported test scores, and 68

PAGE 78

test-based test scores. NCES collected data on follow-up surveys in 1990, 1992, 1994, and 2000. Whereas this study used the NELS:88 data, this study considered two other national datasets: The National Survey of Youth of 1997 and the Common Core of Data. This study chose the NELS:88 data because the dataset provides variables that are more suitable for exploring the research questions proposed in this study. The NELS:88 provides information on students engagement, achievement, attainment and family background from eighth grade through college. The National Longitudinal Survey of Youth of 1997 (NLSY:97) data, collected by the U.S. Department of Labor's Bureau of Labor Statistics, provides information on those living in the United States in 1997 who were born during the years 1980 through 1984. Though the dataset provides valuable information on debt, assets, family, and academic achievement, the dataset does not provide information on students' academic attainment and engagement. The NYLS:97 subsets provide information on students' academic achievement, however, the measurement used PlAT to assess students between 5 and 14 years of age. The Common Core ofData (CCD) collected by the NCES, is a comprehensive national database collected annually from all public elementary and secondary schools. Swanson (2004) explains that the CCD dataset represents 93% of the students and 85% ofthe school districts in the United States. The CCD provides excellent data for analyzing enrollment, dropouts, and high school 69

PAGE 79

attairunent; however, the dataset excludes information needed to examine students' engagement and achievement scores. For this study, the NELS:88 presents a more comprehensive dataset. Sample The sample for this study represents students who were in 1Oth grade in 1990 and who attended public school from lOth grade to 12th grade. This sampled selection follows the suggestion of Orr (2003) and Philips, Brooks-Gunn et al. (1998) that when assessing the relationship between wealth and academic outcomes, using an older student population may produce different results than using younger sample groups. The researchers' rationale was that families might have more established wealth during their students' high school years than during their students' elementary and middle school years. This study also chose to assess students in public school settings because the 2002 U.S. Census data showed that 97% of African Americans attended public schools. After removing those participants who attended private school, the NELS:88 first-year follow up in 1990, produced a sample of 1,302 African American and 6,362 European American participants. This sample provided information needed to analyze data on achievement and engagement of 1Oth graders. The same sampled group from the first-year follow-up study appeared in the second-year follow-up study in 1992. 70

PAGE 80

The second-year follow up data provided information needed to analyze data on 12th graders' high school attainment. Dependent Variables The three dependent variables used in this study are student engagement, academic achievement, and academic attainment. Descriptive data for these are included in table 3.1. Engagement This study used a principal component factor analysis and a reliability analysis on the three types of academic engagement to create a composite engagement scale. The study developed the composite engagement scale based on academic engagement of the sampled African American student population. This study used the scale as a way of understanding which of the 17 engagement variables best represent African American students' school participation. A reliability analysis of the engagement variables with a matrix component score of .30 or higher assessed the internal consistency reliability of the composite scale. The 17 engagement variables are representative of three types of academic engagement: basic school engagement, extracurricular activity engagement, and college preparatory engagement. Basic school engagement assesses students' basic school behavior, such as completing homework, attending school, and being attentive in class. Extracurricular activity engagement assesses students' participation in 71

PAGE 81

extracurricular activities, such as playing sports, taking part in clubs, playing in the band, and participating in other activities not directly associated with academic work. College preparatory engagement assesses students' interest in education beyond high school. This type of engagement measures whether students have taken SAT, ACT, or AP tests. This study assessed engagement based on teachers' and students' answers to questionnaires from the first and second-year follow-up studies. To obtain basic engagement, this study compared the answers of teachers on questions, "How often the student does homework," "How often the student is absent," and "How often the student is attentive in class." The responses to the question, "How often the student is absent" presented a code of 1 as, "The student was never absent." Whereas the responses to the other questions related to the first type of engagement presented a code of 1 as ''Never, the student never did homework and was never attentive in class." The question, "How often the student is absent," was reverse coded (see Appendix A). Reverse coding the responses provided consistency between responses related to engagement. The responses, after recoding, showed a range of the least engaged to the most engaged. Seventy-five percent (n=955) of the African American students and 81% (n = 5,667) of the European American students responded to questions related to basic school engagement. 72

PAGE 82

Table 3.1: Background characteristics of the independent variables (N = 7,664) African American European American (n = 1,302) (n = 6,362) Dependent Variable n % n % Type 1 Engagement 599 46.3% 3,183 50.0% Homework Never 18 1.4% 43 0.7% Rarely 67 5.1% 247 3.9% Some of the time 133 10.2% 599 9.4% Most of the time 239 18.4% 1,365 21.5% All of the time 142 10.9% 929 14.6% 595 45.7% 3,191 50.2% Attentiveness Never 9 0.7% 26 0.4% Rarely 37 2.8% 128 2.0% Some of the time 128 9.8% 591 9.3% Most of the time 276 21.2% 1,620 25.5% All of the time 145 11.1% 826 13.0% 592 45.5% 3,190 50.1% Absences All the time 3 0.2% 6 0.1% Most of the time 28 2.2% 97 1.5% Some of the time 155 11.5% 934 14.7% Rarely 357 27.4% 1,869 29.4% Never 49 3.8% 284 4.5% Type 2 Engagement Participation in Team Sport 317 24.3% 2,695 42.4% Participation in Individual Sport 511 39.2% 2,180 34.3% Participation in Cheerleading 403 30.9% 1,591 25.0% Participation in Music Group 506 43.0% 2,130 33.5% Participation in Plays 437 33.6% 1,890 29.7% Participation in Government 463 35.6% 1,850 29.1% Participation in Honor Society 441 33.9% 2,051 32.2% Participation in Yearbook Club 454 34.9% 2,141 33.6% Participation in Service Club 443 34.2% 1,838 28.9% Participation in Academic Club 528 40.5% 2,426 38.1% Participation in Hobby Club 380 29.2% 1,571 24.7% Participation in FT AIFHAIFF A 548 42.1% 2,080 32.7% 73

PAGE 83

African American European American (n = 1,302) (n = 6,362) Dependent Variable n % n % Type 3 Engagement Took SAT 504 38.7% 2,317 36.4% Took ACT 342 26.3% 2,231 35.1% TookAPTest 109 8.3% 829 13.0% Achievement Standardized Test Scores 505 38.8% 1,059 16.6% Lowest 25th Percentile 293 22.5% 1,463 22.9% Low Percentile 216 16.6% 1,773 27.9% High Percentile 105 8.1% 2,060 32.4% Highest 25th Percentile Attainment Graduated w/diploma in 1992 770 59.1% 4,581 72.0% Other accreditations in 1992 6 0.5% 24 0.4% Did not graduate in 1992 204 15.7% 725 11.4% In order to discover the needed information to obtain the extracurricular engagement variable, this study compared the response of students on a series of questions regarding interscholastic sports activities and activities not directly associated with academics. The series of questions asked, "Please mark one response for each type of interscholastic activity in which you have participated in this school year." NCES surveyors asked students to answer whether they took part in team sports, individual sports, cheerleading, pompom, and drill team. These variables translated into extracurricular engagement (see Appendix A). 74

PAGE 84

Additionally, students answered a series of questions, "Please mark one for each activity in which you have participated in this school year." This series of questions related to students' participation in band related activities, drama/plays/musical, student government, national honor society, yearbook/newspaper committee, service clubs, academic clubs, hobby clubs, and FT AIFHAIFF A. These variables also translated into extracurricular engagement (see Appendix A). Related to sport activities, clubs, and talent activities, it was unclear if the response "multiple responses" indicated both "participated" and "did not participate." This study treated this response as missing. Missing values received a mean score for wither the African American or European American student population. For instance, if an African American student selected "multiple responses" to a question on a sport activity, that student received a mean score of the African American students on that particular question. After recoding the responses related to sports activities, clubs, and talent activities, the combined values made up the variable, extracurricular engagement (see Appendix A). Seventy-five percent (n = 962) of the African American students and 82% (n = 5,716) of the European American students responded to questions related to extracurricular engagement. This study used responses from the series of questions, "Have you taken or are you planning to take any test this year?" This series of questions related to SAT, 75

PAGE 85

ACT, and AP tests and was used to establish the college preparatory engagement variable. Students responded either "o plan to take either test," or they gave the month and year they took a test or planned to take a test. NCES recorded the students' responses as 90 = 1990, 91 = 1991, 92 = 1992, 93 = 1993, and 4 = refused. If a student refused to answer or stated that he or she did not take or plan to take either one of the test, he or she received a code of not engaged (see Appendix B). If a student stated that he or she took either the SAT, ACT, or AP test in 1990, 1991, 1992, or 1993, the student received a code of engaged (see Appendix A). Table 3.1 shows that a slightly higher percentage of African American students took the SAT, and a higher percentage of European American students took the ACT and AP test. This study then used the 17 variables that make up the three types of engagement to develop the composite engagement scale. This study developed the scale by first running a factor analysis on the 17 variables that make up the three types of academic engagement. The study conducted the analysis of the variables for the African American student population. Eleven variables had a component matrix score of 0.30 or higher (see Table 3.2). After removing participation in team sports, music, and plays, the remaining eight variables showed a reliability alpha of0.73 (see Table 3.3). These eight variables formed the composite engagement scale. This scale helped determine the relationship between socioeconomic indicators and the academic engagement of African American students. 76

PAGE 86

Table 3.2 Component Matrix of the three types of engagement TyPes of Engagement Component Attendance in school 0.53 Homework 0.83 Attentive in class 0.87 Participation in team sports -0.48 Participation in individual sports Participation in cheerleading Participation in music group 0.53 Participation in plays and musicals 0.34 Participation in student government Participation in honor society 0.62 Participation in year books Participation in service club Participation in academic club 0.54 Participation in hobby club Participation in FT AIFFH/FHA Taken the SAT 0.55 Taken the Act 0.31 Taken AP test 0.34 Table 3.3 Reliability scale of the seven engagement components Variables Scale of mean Scale of Corrected Homework Attendance Attentive Honor Society Academic Club SAT ACT AP Test oc = 0.73 Achievement if item deleted variance if items-total 23.57 23.60 25.21 30.57 30.58 30.18 30.25 30.29 item is deleted correlation 7.35 0.70 8.77 0.74 11.95 0.45 13.60 0.47 14.72 0.16 14.50 0.35 14.26 0.37 14.05 0.41 Alpha if items deleted 0.65 0.61 0.69 0.70 0.73 0.72 0.71 0.71 This study assesses achievement using standardized test scores and student grade point average. However, Carter (2004) argues that both standardized test scores 77

PAGE 87

and grade point average present biases when measuring students' academic achievement. She explains that standardized tests present a culturally biased method of assessing students' achievement. Carter further explains that grades could reflect more of the teacher's view of his or her own ability to teach instead ofthe students' level of achievement. Carter's validity concerns are well noted and understood. However, given the limited options available, this study used the standardized test as the variable for assessing academic achievement. The grade point average variable consisted of self-reported data that may not reflect students' actual grade point average. Therefore, this study did not select grade point average as a valid variable for assessing students' academic achievement. To assess standardized test scores, this study uses the NELS:88's standardized test composite variable. This variable represents a combination of students' standardized mathematics and standardized reading scores. To obtain these scores, administrators and teachers administered reading and mathematics tests developed by the Educational Testing Services. The tests consisted of multiple-choice questions. The reading test consisted of 21 questions with four to five passages. The reading test questions measured students' comprehension of words, identification of figures of speech, interpretation of author's perspective, and evaluation of the passage. The mathematic test consisted of word problems, graphs and geometric figures, equations, and quantitative comparisons. Researchers who have used the NELS:88 consider the 78

PAGE 88

tests unbiased, reliable, and valid (Battle, 1997; Roscingo & Ainsworth-Darnell, 1999). Rock & Pollack (1995) also found the reading and mathematics tests to have statistical reliability. They showed the reliability ofthe reading test at 0.80 for the base-year study, 0.86 for the first-year follow-up study, and 0.85 for the second-year follow-up study. The authors also found the reliability ofthe mathematics test at 0.89 for the base-year study, 0.93 for the first-year follow-up study, and 0.94 for the second-year follow-up study. Rock and Pollack's data also showed a correlation between mathematics and reading scores of0.71 for the base-year study, 0.76 for the first-year follow-up study, and 0. 74 for the second-year follow-up study. NELS:88 presented that 1,119 (88%) of African Americans and 6,355 (91 %) of European Americans provided test scores. Table 3.1 shows that a higher percent of African American students scored in the low and lowest quartile, a higher percent of European American students scored in the high and highest quartile. Attainment As mentioned in the review of literature, researchers use different methods to assess high school graduation. To comply with the NCLB (2001) defmition of graduation in four years, this study proposes that the most accurate method for obtaining on-time graduation rates is by gathering data directly from students' transcripts. This method reduces errors that could occur when calculating estimations 79

PAGE 89

of dropout rates and graduation rates. Findings demonstrate that to obtain dropout or graduation rates, researchers use graduations, dropouts, transfers, and enrollment as a way to obtain graduation rates. Using several variables can help researchers account for more and different types of errors. For instance, multivariate statistics suggest that when assessing several variables, this method allows researchers to account for more error (Hair, Anderson, Tatham, & Black, 1998). However, this method assumes that the data collection process presents data with a reduced chance of errors. By using transcript data, this study reduced the possibility of error because this study retrieved students' graduation status directly from their transcripts. In this study, NELS:88 transcript data determined high school attainment. However, Swanson (2004) has concerns that the NCES data accounts for only a small portion of the U.S. student population and that the data does not appropriately represent large school districts (p. 34). He argues that because the NCES data accounts only for a small sample of the U.S. student population, NCES's data do not accurately represent high school attainment. However, the NELS:88 represents a national randomly sampled population. According to Gay, Mills, & Airasian (2003), the general definition of a random selection is a sample that equally represents a given population. Additionally, random sampling is an adequate method for measuring a statistical phenomenon in a population. 80

PAGE 90

Given the NELS:88 sampled population, this study uses NELS:88's transcript-indicated outcome variable as the method for measuring academic attainment. The variable provided data on whether students graduated in 1992, whether students received their GED, or whether students dropped out of school (see Appendix B). To determine graduation in 1992, this study developed three graduate statuses. If a student's transcript presented the student as still enrolled, dropped out, aged out, or left for health reasons, this study viewed the student as having not graduated. If a student's transcript showed the student received a diploma in special education, a certificate of attendance, or a GED, the student received either an alternative diploma. If a student's transcript showed that the student graduated in the spring 1992, graduated another time during 1992, or graduated before 1992, this study viewed the student as graduating on time (see Appendix B). The NELS:88 showed transcript data for 80% (n=1,023) of the African Americans and 81% (5,605) European Americans. Additionally, transcript data showed that 59% of African Americans and 72% of European Americans graduated with a high school diploma by their senior year. Table 3.4 shows that various findings indicators a noticeable difference in the graduation rate between African American and European American students. The table also presents that transcript data suggest a 13% difference between the two groups. On the other hand, Davis (2005) shows a 19% difference, Helfand (2005) an 18% difference, and Swanson (2004) a 26% 81

PAGE 91

difference. The results from the transcript data can stem from one of two validity concerns. Either, the findings presented in others' work cannot account for unexplained error (i.e. data collection) or the NELS:88 sampled population is an unrepresentative sample of U.S. public school students. Table 3.4 On-time graduation rates Populations African Americans European Americans Transcript Data 59% 72% Davis (2005) 56% 75% Independent Variables Helfand (2005) 49% 67% Swanson (2004) 50% 76% The independent variables are wealth as measured by parents' assets. Additionally, the disaggregated SES variables income, father's education, mother's education, father's occupation, and mother's occupation represent the other independent variables in this study. Table 3.5 provides descriptive data on the independent variables in this study. Assets As mentioned, Orr (2003) and Philips, Brooks-Gunn et al. (1998) explain that the relationship between wealth and academic outcomes might vary with the age of the child. They explain that parents of older children may have more forms of wealth (e.g., savings) than when their children were younger. NELS:88 data found that as children grew older, more families reported assets. In the base year, 1988, 30% (306) of the parents reported savings set aside for their child's education. In the second-year 82

PAGE 92

Table 3.5 Background Characteristics of the dependent variables (N = 7,664) African American European American (n = 1,302) (n = 6,362) Independent Variable n % n % Assets used for child's education after high school Savings Account 467 35.9% 2,693 42.3% Mortgage 132 10.0% 648 10.2% Trust funds 70 5.4% 620 9.7% Insurance Policy 244 18.7% 1,191 18.7% U.S. Savings Bonds 178 13.7% 1,015 16.0% Investments 207 15.9% 1,302 20.5% Annual household income Less than $15,000 504 38.7% 815 12.8% $35,000-$74,999 421 32.3% 2,042 32.1% $75,000 and More 329 25.3% 3,241 50.9% Father Occupation (Prestige Score) Clerical Worker (56.68) 35 2.7% 269 3.9% Craftsperson (27 .41) 124 9.7% 1,016 14.6% Farmer (28.00) 5 0.4% 183 2.6% Homemaker (N/ A) 3 0.2% 10 0.1% Laborer (7.33) 114 8.9% 358 5.1% Manager (67.73) 61 4.8% 688 9.9% Military (N/ A) 26 2.0% 85 1.2% Machine Operator ( 19 .18) 266 2.8% 1,362 19.6% Professional I (7 .21) 35 2.7% 450 6.5% Professional II (7 .21) 28 2.2% 337 4.8% Sm. Bus Owner (49.70) 14 1.1% 263 3.8% Protect Service (38.00) 19 1.5% 158 2.3% Sales (54.42) 27 2.1% 549 7.9% School Teacher (7.21) 20 1.6% 144 2.1% Service Worker (15.90) 81 6.3% 172 2.5% Technical Worker (61.40) 18 1.4% 189 2.7% Other (N/A) 186 14.5% 351 5.0% 83

PAGE 93

African American European American (n = 1,302) (n = 6,362) Independent Variable n % n % Education 163 12.7% 773 11.1% Not Finish H.S. 335 26.2% 1,941 27.9% Graduated H.S. 102 8.0% 683 9.8% Junior College 69 5.4% 506 7.3% College Least than 4 Y rs 117 9.1% 986 14.2% Graduated College 41 3.2% 549 7.9% Masters Degree 49 2.4% 366 5.3% Ph.D., M.D., Etc. Mother's Occupation (Prestige Score) Clerical Worker (56.68) 223 17.4% 1,522 21.9% Craftsperson (27.41) 27 2.1% 143 2.1% Farmer (28.00) 2 0.2% 34 0.5% Homemaker (N/ A) 123 9.6% 1,186 17.1% Laborer (7.33) 32 2.5% 102 1.5% Manager (67.73) 39 3.0% 279 4.0% Military (N/ A) 10 0.1% Machine Operator (19 .18) 135 1.6% 436 6.3% Professional I (7 .21) 70 5.5% 435 6.3% Professional II (7.21) 8 0.6% 49 0.7% Sm. Bus Owner (49.70) 6 0.5% 116 1.7% Protect Service (38.00) 5 0.4% 16 0.2% Sales (54.42) 30 2.3% 306 4.4% School Teacher (7.21) 44 3.4% 400 5.8% Service Worker (15.90) 332 26.0% 1,365 19.6% Technical Worker (61.40) 14 1.1% 115 1.7% Other (N/A) 82 6.4% 209 3.1% Education Not Finish H.S. 190 14.9% 708 10.1% Graduated H.S. 351 27.4% 2,400 34.5% Junior College 145 11.3% 781 11.2% College Least than 4 Y rs 98 7.7% 570 8.2% Graduated College 123 9.6% 969 13.9% Masters Degree 65 5.1% 480 6.9% 84

PAGE 94

African American (n = 1,302) European American (n = 6,362) Independent Variable n % n % Ph.D., M.D., Etc. 24 1.9% 116 1.7% follow-up, 59% (702) of the parents reported savings set aside for their child's education. Additionally, parents reported more assets. Seven percent of the African American parents in the base year reported assets above $5,000, whereas 14% of the parents in the second-year follow-up reported assets above $5,000. As explained in the literature review, debt, limited ability to obtain meaningful wealth, and assets are variables that influence African Americans' economic status. In this study, availability of assets is one of the variables used to assess economic status and academic achievement. As described by Orr (2003) and Conley (1999), liquid assets have more impact on families' ability to obtain immediate financial support. This study used these asset types as a form. The NELS:88 collected wealth data using the parent questionnaire. The variable wealth comes from responses related to families' access to assets. To establish the variable, assets, parents responded to the questions, "Which ofthe following sources of money did you or will you use to cover current educational expenses for the 1991/1992 school year?" and "Which of the following sources of money will you use to cover your teenagers' future educational expenses?" The answers ranged from, "Used savings/assets for teen's education," "Used 2nd 85

PAGE 95

mortgage for teen's education," "Used alimony/child support for teen's education," and "Used trust fund for teen's education." Parents responded either yes or no to each of the questions (see Appendix B). One thousand (78%) African Americans and 5,552 (80%) European Americans reported some form of assets. As presented by Conley (1999) and Orr (2003), liquid and illiquid assets play a role in students' education. To understand these types of assets, this study viewed, "Used savings/assets for teen's education," "Will you use trust fund for teen's education," "Will you borrow from your U.S. bonds," "Used other forms of savings," "Will you set aside money for your teen's future education," and "Will you use savings/assets for teen's future education" as liquid assets. This study viewed, "Will you use a 2nd mortgage for teen's education," "Will you use insurance policy for teen's education," "Will you use investments in stocks/real estate for teen's education," and "Will you re-mortgage property for teen's education" as illiquid assets (see Appendix B). Table 3.4 illustrates that European Americans reported having more liquid assets than did African Americans, while African Americans had a similar amount of illiquid assets (e.g. second mortgage and insurance policy). SES Indicators Using the NELS:88 dataset to assess the relationship between educational performance and SES, researchers historically relied on the SES composite variable. As mentioned, this SES variable represents an aggregate of father's income, mother's 86

PAGE 96

income; father's education, mother's education, father's occupation, and mother's occupation. As further explained in the theoretical framework in chapter one, researchers have found a strong correlation between those socioeconomic indicators (occupation, education, and income) that make up the SES composite variables (Duncan, 1961; Hauser & Warren, 1997; Hodge et al., 1964; Nakao & Treas, 1994). These findings lead to this study's examination of the relationship between SES indicators and African American students' academic outcomes. In examining socioeconomic status and its relationship with academic outcomes, theSES components father's income, mother's income, father's education, mother's education, father's occupation, and mother's occupation were disaggregated and assessed as separate socioeconomic indicators. However, because the NELS:88 dataset provides only household income, it was impossible to determine father's income and mother' income. Therefore, this study did not disaggregate household income by father's income and mother's income. NELS:88 data show that 1,114 (87%) of the African Americans and 6,117 (88%) of the European Americans responded to questions on income. The median African American annual income ranged from $25,000 to $34,999 while the median European American annual incomes ranged from $35,000 to $49,000. Table 3.4 shows that half of European American population reported an annual salary of at least $75,000, while a quarter of the African American population reported an annual year income of at least $75,000. 87

PAGE 97

NCES used the question, "What is your occupation?" to computed the occupation variables for fathers and mothers. Each parent provided one of the following responses: "Clerical Worker," "Craftsperson," "Farmer," "Homemaker," "Laborer "Manager "Military "Machine Operator "Professional I ' ' "Professional II," "Small Business Owner," "Protect Service," "Sales," "School Teacher," "Service Worker," ''Technical Worker," ''Never worked," and "Other." These variables presented information needed to develop the mother's prestige occupation variable and father's prestige occupation variable. Using the Nakao & Treas' (1994) latest SEI model, this study coded the occupation categories to their corresponding occupational prestige scores. Appendix A provides a comprehensive view of how this study converted the occupational categories into occupational prestige scores (See Appendix B). The mean occupational prestige score for African American fathers is 32.4 (SD = 21. 7) and the mean occupational prestige score for European American fathers is 40.9 (SD = 21.9). Additionally, mothers' occupational prestige scores were higher than fathers' occupational prestige scores. African American mothers' mean occupational prestige score is 37.5 (SD = 22.9), and European American mothers' mean occupational prestige score is 44.3 (SD = 22.0). These mean scores closely resemble Conley and Yeung's (2005) findings, which showed the occupational 88

PAGE 98

prestige scores for their sampled African Americans at 26.9 and a mean occupational prestige score for their sampled European Americans at 42.9. Related to parents' education status, the NELS:88 provided data both on mother's and father's education level. Each parent answered the question, "What is your highest level of education?" Each parent response ranged from, "never completed high school" the Ph.D., M.D., etc. (see Appendix B). The mean level of education for African American fathers is 2.9 (SD = 1.7) and the mean education level for European American fathers is college graduation, 3.4 (SD = 1.8). The mean level of education for African American mothers is 2.9 (SD = 1.6), and the mean level of education for European American mothers is 3.1 (SD = 1.6). Data Analysis As presented in chapter one, researchers have historically shown a statistical correlation between the socioeconomic indicators income, education, and occupation (Duncan, 1961; Hauser & Warren, 1997; Hodge et al., 1964; Nakao & Treas, 1994). Additionally, researchers propose that wealth, in the form of assets, influences the academic achievement of African American students (Conley, 1999; Oliver & Shapiro, 1997; Orr, 2003; Shapiro, 2004). To better understand the relationship between the socioeconomic indicators that make up SES, wealth as measured by assets, and the academic outcomes of African Americans, this study analyzed household income, father's education, 89

PAGE 99

mother's education, father's occupation, and mother's occupation and the variables liquid and illiquid assets. This study conducted separate principal component factor analyses with Varimax rotation on the seven socioeconomic indicators. Based on the findings from these analyses, this study ran a series of multiple regressions to determine whether the computed socioeconomic components produced by the separate factor analysis predicted the three academic outcomes (engagement, achievement, and attainment). The separate factor analyses extracted the seven socioeconomic indicators and presented the variables as smaller intercorrelated components. The factor analyses revealed whether there is a latent principal factor underlying the seven SES-related variables discussed above and which combinations of variables account for independent variance. By including wealth, this study analyzed whether wealth loads on the same principal factor as other variables, whether wealth constitutes a separate factor from income, and whether these factor structures are different for African American and European American students. The first factor analysis focused on the sampled African American population, the next factor analysis focused on the sampled European American population, and the third factor analysis focused on the African American and European American population. In the factor analysis process, the analysis groups variables by extracting those variables with the greatest amount of variance from the sampled populations. This 90

PAGE 100

study reduced the variables household income, father's education, mother's education, father's occupational prestige score, mother's occupational prestige score, liquid assets, and illiquid assets into smaller interrelated variables for each sampled population. After the factor analysis grouped the first intercorrelated components, the analysis grouped the next set of variables by extracting those variables with the next greatest amount of variance from the variables not in factor one. The extraction process continued until the factor analysis extracted all the variables. The extraction process continued until the factor analysis extracts and groups all of the variables related to the selected sampled population (Anderson & Black, 1998; George & Mallery, 2001). For instance, when looking at the African American population, the factor analysis statistically selected which of the seven socioeconomic indicators best explained the highest amount of variance in this population. Then the analysis presented the remaining variables that explained the next highest amount of variance in this population, and so on. This study used this same process extracting variables for the sampled combined African American and European American population. By reducing the variables to smaller commonly intercorrelated components, these smaller variables explained general interactions between socioeconomic indicators and the sampled populations. The first analysis provided variables related to the sampled African Americans' group response to assets, liquid assets, illiquid assets, household income, father's education, mother's education, father's occupation, 91

PAGE 101

and mother's occupation. The second analysis provided variables related to the sampled European Americans' group response to assets, liquid assets, illiquid assets, household income, father's education, mother's education, father's occupation, and mother's occupation. The third analysis provided variables related to the combined African American and European American sampled group response to the sample variables. The purpose is to understand which socioeconomic indicators explain the socioeconomic experience of the sampled African American population and which socioeconomic indicators explain the experience of the sampled combined African American and European American population. Each factor loadings received z-scores. These z-scores are standardized scores of the interconnected components. Further, the z-scores provide data needed to run multiple regression analyses. This study produced three series of multiple regression analyses. These analyses showed how predictor variables relate to criterion variables. The ftrst series of the multiple regression analyses examined the relationship between the socioeconomic indicators related to the sampled African American population (predictor) and the three academic outcomes (criterion). The second series of multiple regression analyses examined the relationship between socioeconomic indicators related to the sampled European American population (predictor) and the academic outcomes (criterion). 92

PAGE 102

More specifically, the first series of multiple regressions measured the relationship between factor loadings and academic engagement. In this series, the first multiple regression analysis measured the relationship between the factor loadings related to African Americans and the academic engagement of this group. The next analysis measured the relationship between the factor loadings related to European Americans and the academic engagement of African American students. The third analysis measured to relationship between the factor loadings related to European Americans and the academic engagement of European American students. The last analysis measured the relationship between factor loadings related to African Americans and the academic engagement of European American students. The second series of multiple regressions measured the relationship between factor loadings related and the academic achievement. The first analysis measured the relationship between the factor loadings related to African Americans and the academic achievement of African American students. The next analysis measured the relationship between the factor loadings related to European Americans and the academic achievement of African American students. The third analysis measured to relationship between the factor loadings related to European Americans and the academic achievement of European American students. The final analysis measured the relationship between factor loadings related to African Americans and the academic achievement of European American students. 93

PAGE 103

The last series of multiple regressions assesses the relationship between factor loadings and the academic attainment. The first multiple regression analysis measured the relationship between the factor loadings related to African Americans and the academic attainment of this group. The next analysis measured the relationship between the factor loadings related to European Americans and the academic attainment of African American students. The third analysis measured to relationship between the factor loadings related to European Americans and the academic attainment of European American students. The final analysis measured the relationship between factor loadings related to African Americans and the academic attainment of European American students. The series of multiple regression analyses on factor loadings and the academic outcomes provided fmdings that explain whether these socioeconomic components related to African Americans predict the academic outcomes of African American and European American students differently. These findings also showed which socioeconomic components best predict the academic outcomes of both African Americans and European Americans. These fmdings will provide a greater understanding on the relationship between socioeconomic indictors and the academic performance of African American students. 94

PAGE 104

CHAPTER FOUR Results The primary purpose of this study was to compare the relationship between wealth and individual socioeconomic components of socio-economic status to educational outcomes for African American and European American students. The study presented two hypotheses to test. First, this study tested whether socioeconomic indicators related to African Americans predict academic engagement, academic achievement, and academic attainment for African American and European American students differently. Second, this study tested whether socioeconomic indicators related to European Americans predict academic engagement, academic achievement, and academic attainment for both African American and European American students differently. This study first used two principal component factor analyses with varimax rotation to understand the relevancy of the socioeconomic indicators in explaining the socioeconomic experience of both African Americans and European Americans. Afterwards, this study conducted a series of multiple regression analyses on the relationship between the socioeconomic components produced by the factor analyses and the academic outcomes of African American and European American students. The purpose of the factor analysis is to provide empirical evidence for the combining 95

PAGE 105

of particular socioeconomic indicators rather than relying on the historical combination of income, occupational status, and educational level traditionally used in sociological research. In particular, this analysis allows for the examination of the relationship of wealth to other variables that have historically been associated with socioeconomic status, and to see whether distinct patterns of relationship emerge for African American and European American families. Factor Analyses This study conducted separate factor analyses for African Americans and European Americans following a three-step procedure. The first step demonstrated the appropriateness of the independent variables (socioeconomic indicators) for conducting a factor analysis. The second step provided data on the un-rotated component matrix, which provided an overall view of the portion of variance each socioeconomic indicator explained. The third step provided data on the rotated component matrix, which produced models that explained the correlation of the socioeconomic indicators. Factor Analysis on socioeconomics related to African Americans The first step of the factor analyses consisted of assessing whether the variables, father's education, mother's education, father's occupational prestige score, mother's occupational prestige score, and income were adequate for running a factor analysis. A Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO) and a 96

PAGE 106

Bartlett's test showed that the socioeconomic indictors related to African Americans were adequate for conducting a factor analysis. The analysis produced a KMO of .832; values greater than .8 are meritorious, and are adequate for conducting a factor analysis (George and Mallery, 2001, p. 242). The Bartlett's test was significance (p < .000). These findings indicate that the socioeconomic indicators related to the sampled African American population are not an identity matrix and are acceptable for conducting a factor analysis. The next step of this analysis examined the un-rotated principal component matrix for the variables that traditionally combine to measure SES and the six variables that represent forms of assets. Table 4.1 shows that, with exception of insurance policies, U.S. savings bonds, and second mortgage, the other indicators loaded at least .20 on the first component. These fmdings indicate that except for U.S. savings bonds, second mortgage, and insurance policies, the other indicators that combine to make up SES and investments are all associated with a single principal factor accounting for the largest amount of shared variance among the variables. The third step of this analysis examined the rotated component matrix. An orthogonal rotation (Varimax) yielded three completely independent (uncorrelated) factors (see Table 4.2). The first factor loading consists of the socioeconomic indicators that combined to make up the SES variable. The second factor loading represents assets 97

PAGE 107

generated directly from income. Finally, the third factor loading represents assets not generated directly from income. Table 4.1 Un-Rotated Component Matrix of African Americans Socioeconomic Indicators Principal Component Father's Education .731 Mother's Education .774 Annual Income .694 Father's Prestige Score .716 Mother's Prestige Score .648 U.S. Savings Bonds Investments in Stocks/Real Estate Trust Funds Savings Account 2nd Mortgage Insurance Policy .434 .358 .200 .175 Table 4.2 Rotated Component Matrix of African Americans Socioeconomic Indicators Component 1 2 3 Father's Education .791 Mother's Education .816 Annual Income .687 Father's Prestige Score .734 Mother's Prestige Score .632 U.S. Savings Bonds Investments in Stocks/Real Estate Trust Funds Savings Account 2nd Mortgage Insurance Policy .587 .625 .413 -.485 .688 .833 .515 Two separate reliability analyses measured the internal consistency of each of the first two factor loadings. These analyses determined whether the items within each of the factor loadings measured the same thing. The first analysis measured the internal consistency of the items (father's education, mother's education, father's 98

PAGE 108

occupation, and mother's occupation) that make up theSES component. The analysis yielded a Cronbach's alpha of0.75. The second analysis measured the internal consistency of the items (U.S. savings bonds, investments, and savings account) with a rotated value ofO.SO or greater for income generated assets. The second analysis measured the internal consistency of the items (U.S. savings bonds, investments, and savings. The analysis yielded an alpha of 0.44. The third loading, non-income generated assets consist of one variable with a rotated value of 0.50 or higher, second mortgage, so a Cronbach analysis was not necessary. Next, this study created three separate composite variables from the three factor loadings. The first factor loading, the SES composite variable derived from converting father's education, mother's education, father's occupation, and mother's occupation into z-scores (centered on the grand mean of the data set); then adding together the four separate z-scores into a single score. The second factor loadings represented the composite variable income generated assets. This composite variable consists of the combined z-scores for U.S. savings bonds, investments, and savings accounts. The third factor loading, second mortgage represented the composite variable, non-income generated assets. Though the composite variable consisted of only the variable second mortgage, this study converted the values from the variables into a z-score. The purpose of converting the variable into a z-score was to make the values of variable consistent with the other two composite variables. 99

PAGE 109

Factor Analysis on socioeconomics related to European Americans The KMO for the sampled European American population was .842. This statistic was similar to the KMO for sampled African American population. The Bartlett's test with a significance of .000, and indicates that the socioeconomic indicators related to the sampled European American population are not an identity matrix and are acceptable for conducting a factor analysis. Table 4.3 presents findings for the un-rotated principal component matrix. The table shows that only the variables father's education, mother's education, father's occupation, and mother's occupation loaded with a value equal to or higher than .20 on the first component. These findings indicate that the socioeconomic indicators that combine to make up SES are all associated with a single principal factor accounting for the largest amount of shared variance among the variables. Table 4.3 Un-Rotated Component Matrix of European Americans Socioeconomic Indicators Principal Component Father's Education .800 Mother's Education .727 Pu!nuallncome .688 Father's Prestige Score .714 Mother's Prestige Score .500 U.S. Savings Bonds Investments in Stocks/Real Estate Trust Funds .211 Savings Account .155 2nd Mortgage Insurance Policy .130 100

PAGE 110

Table 4.4 shows that the rotated procedure of the factor analysis produced three completely independent (uncorrelated) factors. The first factor loading consists of the socioeconomic indicators that combined to make up the SES variable. The second factor loading represents liquid assets. Finally, the third factor loading represents illiquid assets. The assets for African Americans (see Table 4.2) loaded different that assets for European Americans (see Table 4.4). While assets for African Americans represented income-generated assets and non-income-generated assets, European American's assets represents liquid assets and illiquid assets. As described in chapter two, Orr (2003) explains that CD's, stocks, and savings accounts are products of liquid assets. Whereas, Conley (1999) explains that trust-funds, inheritances, vehicles, and real estate are illiquid assets. Regarding trust funds, the Federal Deposit Insurance Corporation (FDIC) (2006) explains that trust funds can be considered income if the funds are living trust-fund. This type of trust fund is set up by parents for their child while they (the parents) are still alive. If a child draws from this fund, the withdrawal is consisted income. Related to this study, when asked how they will support their child's college education, one of the answers presented by parents was that they would establish a trust-fund. In this study, this type of trust-fund is concerned a living trust-fund that can be immediately assets. Therefore, trust-fund is considered a form of liquid assets. 101

PAGE 111

Table 4.4 Rotated Component Matrix of European Americans Socioeconomic Indicators Component 1 2 3 Father's Education .829 Mother's Education .759 Annual Income .660 Father's Prestige Score .727 Mother's Prestige Score .510 U.S. Savings Bonds .542 Investments in Stocks/Real Estate .691 Trust Funds 713 Savings Account .516 2nd. Mortgage .601 Insurance Policy .652 Two separate Cronbach analyses measured the internal consistency of each of the second and third factor loadings. The first factor loading represents the same SES composite variable found for the sampled African American population. A Cronbach analysis measured the internal consistency of the items (investments, trust funds, and savings accounts) that represents liquid assets. The analysis yielded an alpha of0.47. A Cronbach analysis also measured the internal consistency of the items (U.S. savings bonds, second mortgage, and insurance policy) that represents illiquid assets. The Cronbach analysis yielded an alpha of 0.23. Next, this study created two separate composite variables for the second and third factor loadings. The first factor loading, the liquid assets composite variable consisted of converting investments, trust funds, and savings accounts into z-scores. Again, the separate z-scores combined to create a single score. The second factor loadings represented the composite variable illiquid assets. This composite variable 102

PAGE 112

consists of the combined z-scores for U.S. savings bonds, second mortgage, and insurance policy. Multiple Regressions This study examined two series of multiple regression analyses. The first series of multiple regressions examined the relationship between the three factor loadings related to African Americans (SES, income-generated assets, and non income generated assets) and the academic outcomes of both African American and European American students. The second series of multiple regression analyses examined the related between the three factor loadings related to European Americans (SES, liquid assets, and illiquid assets) and the academic outcomes of both African American and European American students. These multiple regression analyses also examined the statistical influence of each of the factor loadings on three academic outcomes of both the African American and European American students. Academic Engagement for African American students This study conducted two separate multiple regression analyses on socioeconomic indicators and academic engagement of African American students. The first multiple regression analysis with the enter method examined the relationship between the factor loadings related to African Americans and the academic outcomes of African American students. The second multiple regression analysis with the enter method examined the relationship between the factor loadings related to European 103

PAGE 113

Americans and the academic engagement of African American students. The analysis on the factor loadings related to African Americans and the academic engagement of African American students was not significant (R = 0.14) (see Table 4.5). The analysis on the factor loadings related to European Americans and the academic engagement of African Americans was also non-significant (R = 0.11 ). Table 4.5 shows that income generated assets was the greatest predictor of the academic engagement of African American students (j) = 0.11 ). Academic Engagement for European American students The first multiple regression analysis showed a significantly weak positive relationship between the factor loadings related to African Americans and the academic outcomes of European American students (R = 0.07) (see Table 4.6). The multiple regression analysis also showed a significantly weak positive significant relationship between the factor loadings related to European Americans and the academic engagement of European American students (R = 0.07). SES was the greatest predictor of the academic engagement of European American students (j) = 0.06) (see Table 4.6). Income generated assets was the second greatest predictor of the academic engagement of European American students. Further, income generated assets predicted the academic engagement of African American students 5% greater than it predicted the academic engagement of European American students. Academic Achievement for African American students 104

PAGE 114

Table 4.7 illustrates a significantly positive relationship between the factor loadings and the standardized test scores for African Americans (R = 0.49). The table also shows that when examining the factor loadings related to African Americans and standardized test scores, SES was the greatest predictor of the standardized test scores of African Americans (f3 = 0.45). Income generated assets was the next greatest predictor of the standardized test scores of African American students (f3 = 0.15). Table 4.7 also show that the multiple regression analysis between the factor loadings related to European Americans and the standardized test scores of African American students had the same regression as the factor loadings related to European Americans and the standardized test scores of African American students (R = 0.49; p<.Ol). While SES related to both African Americans and European Americans was the greatest predictor of the standardized test scores of African Americans (f3 = 0.45), income generated assets was the next greatest predicted of the standardized test scores of African Americans (f3 = 0.15). In addition, liquid assets also predicted the standardized test scores of African American students (f3 = 0.13). Academic Achievement for European American students Table 4.8 presents a significantly positive relationship between the factor loadings related to African Americans and the standardized test scores of European Americans (R = 0.38). In addition the table shows that relationship between the factor loadings related to European Americans and the standardized test scores of European 105

PAGE 115

Table 4.5 Regression of the academic engagement of African Americans Variables R Standardized Factor Loadings Related to African Americans (N=341) Regression SES Income Generated Assets Non-Income Generated Assets Factor Loadings Related to European Americans (N=338) Regression SES Liquid Assets Illiquid Assets p<.05*, p<.Ol ** 0.14 0.11 Table 4.6 Regression of the academic engagement of European Americans Beta 0.02 0.11* 0.08 0.03 0.04 0.09 Variables R Standardized Factor Loadings Related to African Americans (N=2,206) Regression SES Income Generated Assets Non-Income Generated Assets Factor Loadings Related to European Americans (N=2,281) Regression SES Liquid Assets Illiquid Assets 0.07** 0.07* 106 Beta 0.06* 0.04* 0.01 0.05* 0.03 0.02

PAGE 116

American students was also significant (R = 0.38). The table further presents that SES was the greatest predictor of the standardized test scores of European American 0.38). However, assets did not significantly predict the standardized test scores of European American students. Academic Attainment of African American students Table 4.9 provides data on the relationship between the factor loadings and on-time graduate of African American students. The analysis found a small significantly positive relationship between the factor loadings related to African Africans and the on-time graduation of African American students (R = 0.21) and the factor loadings related to European American students and the on-time graduation of African American students (R = 0.20). SES was the only predictor of on-time graduation of African American students. Academic Attainment of European American students Table 4.10 illustrates a small positive relationship between the factor loadings related to African Americans and the on-time graduate of European American student (R = 0.21) and a small positive relationship between the factor loadings related to European Africans and the on-time graduation of European American students (R = 0.20). SES was the greatest predictor of the on-time graduation of European American students, and liquid assets were the second greatest predictor of the on-time graduation of European American students. 107

PAGE 117

Table 4. 7 Regression of the academic achievement of African Americans Variables R Standardized Factor Loadings Related to African Americans (N=262) Regression SES Income Generated Assets Non-Income Generated Assets Factor Loadings Related to European Americans (N=259) Regression SES Liquid Assets Illiquid Assets 0.49** 0.49** Table 4.8 Regression of the academic achievement of European Americans Beta 0.45** 0.15** -0.04 0.45** 0.13* -0.01 Variables R Standardized Factor Loadings Related to African Americans (N=1,782) Regression SES Income Generated Assets Non-Income Generated Assets Factor Loadings Related to European Americans (N=1,771) Regression SES Liquid Assets Illiquid Assets 0.38** 0.38** 108 Beta 0.38** 0.05 0.03 0.38** 0.02 0.02

PAGE 118

Table 4.9 Regression of the academic attainment of African Americans Variables R Standardized Factor Loadings Related to African Americans (N=274) Regression SES Income Generated Assets Non-Income Generated Assets Factor Loadings Related to European Americans (N=271) Regression SES Liquid Assets Illiquid Assets 0.21 ** 0.20* Table 4.10 Regression of the academic attainment of European Americans Beta 0.20** 0.01 -0.10 0.19** 0.15 -0.06 Variables R Standardized Factor Loadings Related to African Americans (N=1,911) Regression SES Income Generated Assets Non-Income Generated Assets Factor Loadings Related to European Americans (N=l,891) Regression SES Liquid Assets Illiquid Assets 109 0.20** 0.20* Beta 0.20** -0.04 0.00 0.20** -0.04* -0.02

PAGE 119

Summary of Findings The two hypotheses for this study proposed first that socioeconomic indicators related to African Americans would predict the academic outcomes for African American and European American students differently. Secondly, that the socioeconomic indicators related to European Americans would predict the academic outcomes for African American and European American students differently. For understand the relationship between socioeconomic indicators and the academic outcomes of the two populations, two separate factor analyses show that SES loaded the same for both ethnic groups, assets loaded differently. African Americans' assets represented assets that were income generated and non-income generated. On the other hand, European Americans' assets reflect liquid and illiquid assets. Orr (2003) presents that income generated assets are often viewed as liquid assets. In this study income generated assets are described as those assets that are produced by income. Whereas non-income generated assets represent assets generated by other sources, such as a loan (e.g. second mortgage) or an established account (e.g. trust-fund). After concluding the factor analyses, the series of multiple regressions produces findings for the two hypotheses (see Table 4.11 ). The findings from the table below shows that while the socioeconomic indicators related to African Americans did predict the academic engagement and academic achievement of African American and European American students differently, the tables also shows 110

PAGE 120

that the socioeconomic indicators related to African Americans predicted the academic attainment of African American and European American students the same. Table 4.11 further illustrates that the socioeconomic indicators related to European Americans predicted the academic engagement, achievement, and attainment of African American and European American students differently. Table 4.11 Summarizing hypotheses Academic Outcomes Hypothesis 1 Engagement Achievement Attainment True True False Hypothesis 2 True True True The findings in this study further explain that SES is the greatest predictor of the academic achievement and attainment of African American students. In addition, the findings explain that SES is the greatest predictor of all of the academic outcomes of European American students. Related to wealth/assets, the findings presents that income generated assets is the greatest predictor of the academic engagement of African American students, and that income Generated Assets and Liquid Assets predicts a significant portion of the academic achievement of African American students. On the other hand, income generated assets and liquid assets does not predict a significant portion of the academic achievement of African American students The conclusion is that these findings presented support that the socioeconomic indicators related to African Americans and European Americans predict the 111

PAGE 121

academic outcomes of the two groups differently than the traditionally. Further, while SES was the greatest predictor of the academic achievement of African Americans and European American students, assets predict the academic outcomes of African Americans more than the variable predicts the academic outcomes of European Americans. 112

PAGE 122

CHAPTER FIVE Conclusion Researchers use socioeconomic status (SES) to the stratify individuals into social classes so that they (researchers) can easily understand how individuals in this class system function in various social settings or social institutions. However, when examining the relationship between social class and these social environments, it is easy to lose track of the role culture plays on this relationship. As explained in this study, African Americans' cultural and historical experiences with socioeconomic indicators (e.g. education, occupation, and income) is different that European Americans' cultural and historical experiences with these indicators. Therefore, relying on a variable such asSES to explain the two groups' relationship to education may not fully present the influence culture has on the academic outcomes of the two groups. The approach supports Duncan's ( 1961 b) argument that research cannot expect a single socioeconomic index to predict the social stratification of a complex and multidimensional population of people. As presented in this study, Conley ( 1999) and Orr (2003) propose wealth as a variable that explains a portion of the academic experience of African American students. Their argument suggests that wealth in the form of net worth (Conley, 1999) and assets (Shapiro, 2004) has a greater influence on the economic experience of African Americans than other economic indicators. Further, Orr (2003) findings show 113

PAGE 123

that wealth (liquid assets) correlated positively with the standardized test scores of African American students, while Conley's (1999) findings show that wealth is positively associated with the high school and college graduation rate of African American students. Conley (1999) foWld that when examining wealth, in some cases, African American students graduate from high school at a higher rate than European American students. This current study found that while SES was the strongest predictor of the academic achievement and academic attainment of African American students, income-generated assets also contributed significantly to the prediction of the academic achievement and academic engagement of these students. Unlike Conley's (1999) findings, this current study did not find a significant relationship between wealth and the high school attainment of African Americans. More striking, this current study did not find that assets contributed significantly to the prediction of the academic engagement European American students. However, the fmdings show that assets predicted a small significant portion of these students' academic attainment. Limitations and Future Studies Although the findings in this current study are similar to the findings from other studies, there are limitations associated with this study. The study presents three limitations of this study and explains how future research can address these limitations. 114

PAGE 124

Limitation 1 The first limitation is the measure of wealth (assets). In the NELS:88 public used dataset, parents were asked a series of questions related to which type of assets would be used to fund their child's education beyond high school. Even though these questions suggest that parents had or would have had access to various forms of assets, it is uncertain the extent of these assets. For instance, it is difficult to determine the amount of the assets held by the families and the percent of the assets the parents would use to fund their child's school. In addition, the dataset did not provide the debt held by the parents. By understanding amount assets held by the family and the amount of debt owed by the family, this study could have gained a more accurate understanding of the family's wealth. In Conley's (1999) study, one of the ways he measured wealth was by the net worth of the family. He describes the net worth as assets minus liabilities (debt). Addressing Limitation 1 A more effective dataset for this study would provide comprehensive data on assets, debt, academic engagement, academic achievement, and academic attainment. As mentioned chapter three, other national dataset did provide a more comprehensive data on economic and academic. The National Longitudinal Survey of Youth of 1997 (NLSY :97) dataset provides valuable information on debt, assets, family, and academic achievement, the dataset does not provide information on students' 115

PAGE 125

academic attainment and academic engagement. The Panel Study of Income (PSID) dataset provides the most comprehensive data on wealth, however, the dataset only provides data for attainment and standardized test scores through age 12. The Common Core of Data (CCD) dataset provides excellent data for analyzing enrollment, dropouts, and high school attainment; however, the dataset excludes information needed to examine students' engagement, achievement scores, and wealth. Limitation 2 The second limitation to this current study is the use of single-year income. Yeung and Conley (2004) explain that more scholars acknowledge that single-year income does not explain the overall economic level of the family. They explain that a more appropriate measure of income should be multiple-year income. Yeung and Conley (2004) explain that family's income can change as their jobs change, and that this activity affects the income of families with lowest salaries. In addition, the NELS:88 data shows that total household income changed significantly between the base year (1988) and the second-year follow-up (1992). This study did not measure multiple-year income because the interest of this study is to examine the inter-correlation (part of the factor analysis) between income, education, and occupational prestige. However, the NELS:88 dataset only provides the education and occupation of parents during at one of it survey periods. It is likely 116

PAGE 126

that the inter-correlation between income, education, and occupational prestige would look different if this study had used multiple-year data. In addition, if families' multiple-year income were higher than their single-year income, there would be no checking if occupational prestige and education level for multiple-years are also higher. Addressing Limitation 2 This study shows that SES and wealth together correlated significantly with the academic outcomes of African American students. Future studies that examine the relationship between SES, wealth, and academic outcomes for this population should make use of multiple-year values of SES and multiple-year wealth. Another approach to understanding the relationship between these variables is to examine whether the change in SES and wealth overtime can predict a change in academic outcomes over the same period. These two approaches will allow researchers to examine the relationship of wealth and SES to academic outcomes over time, and degree to which the change in SES and wealth result in a change in academic outcomes. Finally, while single-year SES and single-year wealth gets a snap shot in the economic experience of families, multiple-year SES and multiple-year wealth provides an insight into the financial pattern of families. 117

PAGE 127

Limitation 3 While this study explains the relationship between the socioeconomic indicators and the academic outcomes of African American and European American students, this study did explains the academic gap between the two populations. Though the study shows how socioeconomic indicators influence the academic of outcomes of the two groups, the study did not account whether a factor loadings related to the two groups explain the gap in the two groups academic performance. As presented in table 5.1, data from the NESL:88 shows that the academic engagement, achievement, and attainment of European American student are higher than that of African American students. 5.1 Academic outcomes of African Americans and European Americans Academic Outcomes African Americans European Americans Engagement (mean) 29.41 45.15 Achievement (mean) 32.51 52.09 Attainment 59% 72% Addressing Limitation 3 To further examine wealth, SES, and the academic outcomes of African Americans, future studies should utilize statistical techniques that require a more in depth examination into the relationship between these variables. Studies using path analysis or structural equation modeling may determine possible causal relationships between wealth, SES, and academic outcomes. In additional, studies using hierarchical linear modeling can estimate the variance associated with assets both 118

PAGE 128

within races and between races. These types of studies can provide a more in depth insight into the role wealth on the education of African American students. Implication of this Study Despite the limitations of this study, the fmdings in this study suggest that while SES does play a significant role in understanding the academic outcomes of African Americans, wealth also plays an important role in understanding the academic outcomes of African American students. These findings have important implications on the field of education and educational research. First, the findings imply that to understand the relationship between education and social indicators associated with cultural groups, it is important to take into account the multidimensional characteristics associated with the cultural groups. As this study shows, socioeconomic indicators related to one cultural group (African Americans) do not necessarily explain how another cultural group (European Americans) will respond to those same socioeconomic indicators. In addition, this study implies that while there are common elements (e.g. SES) associated with different cultural groups, there are also cultural relevant elements (e.g. wealth) that specifically explain each cultural group's response to the same social experience (e.g. education). Secondly, this study implies that wealth has an important effect on the academic outcomes of African American students. While wealth likely increases the chances of higher academic success among African American students, those students 119

PAGE 129

in families without wealth are likely to experience lower academic outcomes. Alarming to the African American communities is that the median African American family earning less than $15,000 annually has no assets (Conley, 1999). Even more alarming, is that while African American homeownership has increased between 1994 and 2005, only 48% of African Americans own home (see Table 5.2). In addition, the percent of African American homeowners is lower than other U.S. ethnic groups. It is not possible in a correlational study such as this one to establish a causal relationship among variables. Nonetheless, this study provides support for the likelihood that there is a causal relationship between wealth and educational outcomes for African American students. The advantage wealth/assets provides these students are in the form of educational resources. While families with wealth/assets are better able to provide their children with educational materials (e.g. computers, books) necessarily to obtain higher levels of academic success, families without wealth/assets may not be able to provide their children these advantages. In addition to providing their children with educational resources, families with wealth/assets are more able to afford a home in an upscale neighbor with better schools than those families who do not have wealth/assets. So long as the large discrepancy between the wealth of African Americans and European Americans persists, the academic outcomes of African American students are likely to remain lower. On the other hand, 120

PAGE 130

those African American students in families with wealth/assets will continue to produce higher academic outcomes that those African Americans in families without wealth/assets. 5.2 Homeownership 1994 2005 %Increased African American 42% 48% 6% European American 68% 73% 5% Hispanic/Latino 41% 50% 9% Native American 52% 58% 6% Asian/Pacific 51% 60% 9% Source: U.S. Census 2004: Table 20. Homeownership Rates by Race and Ethnicity of Householder: 1994 to 2005. Though this study shows that wealth has an influence over the academic outcomes of African American students, the findings in this study are not to downplay the importance of race and racial differences found in the U.S. educational system. This study, however, shows that despite the strides African Americans have made in income, education, and occupation, African American families must focus on increasing their wealth/assets. 121

PAGE 131

REFERENCES Ainsworth-Damell, J. W., & Downey, D. D. (1998). Assessing the oppositional culture explanation for racial/ethnic: Differences in school performance. American Sociological Review, 63(4), 36-553. Asimov, N. (2005, April 2005). Study puts Oakland dropout rate at 52% mayor decries crisis district questions research accuracy. San Francisco Chronicle, p. A-1. Adams, C., & Singh, K. (1998). Direct and indirect effects of school learning variables on the academic achievement of African American 1Oth graders. Journal of Negro Education, 67(1 ), 48-66. Alexis, M. (1998). Assessing 50 years of African-American economic status, 19401999. The American Economic Review, 88(2), 368-375. Anderson, C. (1994). Black labor European American wealth: The search for power and economic justice. Bethesda, MD: PowerNomics Corporation of America. Banks, J., & Banks, C. A. (Eds.) (2004). Handbook of research on multicultural education, 2nd edition. San Francisco, CA: Jossey-Bass. Battle, J. (1997). The relative effects of married versus divorced family configuration and socioeconomic status on the educational achievement of African American middle-grade students. Journal of Negro Education, 66(1 ), 2942Battle, J., & Coates, D. L. (2004). Father-only and mother-only, single parent family status of Black girls and achievement in grade twelve and at two-year post high school. Journal of Negro Education, 73(1). 392-407. Blishen. B. L., Carroll. W. K., & Moore C. (1987). The 1981 socioeconomic index for occupations in Canada. Rev Soc A nth 1987 (24), 465-88. Brody, G. H., Stoneman, Z., Flor, D., McCrary, C., Hastings, L., & Conyers, 0. ( 1994 ). Financial resources, parent psychological functioning, parent co caregiving, and early adolescent competence in rural two-parent African American families. Child Development, 65, 590-605. 122

PAGE 132

Braddock, J. H., & Dawkins, M.P. (1993). Ability grouping, aspirations, and attainments: Evidence from the national educational longitudinal study of 1988. Journal ofNegro Education, 62(3), 324-336. Brooks-Gunn, J., Duncan, G. J. & Maritato, N. (1997). Poor families, poor outcomes: The well-being of children and youth. In J. Brooks-Gunn, & G. Duncan (Eds.), Consequences of growing up poor. (pp. 1-17). New York: Russell Sage Foundation. Carter, T. P. (2003). Does it take a whole village? Influences of the family, community and school on African American and Hispanic youth's academic achievement (Doctoral dissertation, University of Missouri-Columbia, 2003). ProQuest Information and Learning Company, UMI Number 3091909. Choy, S. P. (2000). Debt burden four years after college. Washington, DC: U.S. Department of Education. National Center for Educational Statistics. Coleman, J. S., Campbell, E. Q., Hobson, C. J., McPartland, J., Mood, A.M., Weinfeld, F. D., & York, R L. (1966). Equality of Educational Opportunity. Washington, DC: U.S. Government Printing Office. Collins, P. H. (2003). Black feminist thought: Knowledge, consciousness, and the politics of empowerment. NY: Routledge. Conley, D. (1999). Being Black, Living in the Red: Race, Wealth, and Social Policy in America. Berkeley, CA: University of California Press. Cook, P. J., & Ludwig, J. (1998). "The burden of'acting White': Do Black adolescents disparage academic achievement? In C. Jencks, and M. Phillips. (Eds.) The Black-White test score gap (pp. 375-400). Washington, D.C: Brookings Institute. Craig, H. K., Connor, C. M., & Washington, J. A. (2003). Early positive predictors of later reading comprehension for African American students: A preliminary investigation. Language, Speech, and Hearing Services in Schools, 34, 31-43. 123

PAGE 133

Diaz-Gimenez, J., Quadrini, V., & Rios-Rull, J-V. (1997). Dimensions of inequality: Facts on the U.S. distributions of earnings, income, and wealth. Federal Reserve Bank of Minneapolis Quarterly Review 21(2), 3-21. Dalaker, J., & Proctor, B. (2000). Poverty in the United States: 1999: Current population reports, series P-60 No. 21. Washington, DC: US. Government Printing Office. Davis, A. (2005, May 20). DPS numbers tell dismal story; for solutions, listen to kids. The Rocky Mountain News retrieved on December 13, 2005 from http://www.rockymountainnews/drmn/ news_columnists/Article/0,1299,DRMN_86_3790872,0.html Davis, P., McLeod, K., Ransom, M., and Ongley, P. (1997). The New Zealand socioeconomic index of occupational status (NZSEI). Wellington, New Zealand: Statistics New Zealand. Dev, P. C. (1997) Intrinsic motivation and academic achievement: What does their relationship imply for the classroom teacher? Remedial and Special Education, 18(1), 12-19. Donovan, R. (1984). Path analysis of theoretical model of persistence in higher education among low-income Black youth. Research in Higher Education, 21(3), 243-259. Duncan, G. L., & Duncan, K. A. (2005) Can family socioeconomic resources account for racial and ethnic test score gaps? Future ofChildren 15(1), 35-54. Duncan, 0. D. (1961). A socioeconomic index for all occupations (pp. 109-138). In J. Reiss Jr. (Ed.), Occupations and social status. NY: Free Press of Glencoe. Duncan, 0. D. (196lb). Properties and Characteristics of the Socioeconomic Index (pp. 139-161). In J. Reiss Jr. (Ed.), Occupations and social status. NY: Free Press of Glencoe. Ensign, J. (2002). Subsequent educational and professional attainment of black and European American students from two segregated schools. Journal of Negro Education, 71(4), 331-346. 124

PAGE 134

Fehnnann, P. G., Keith, T., Z., & Reimers, T. M. (1987). Home influence on school learning: Direct and indirect effects of parental involvement on high school grades. Journal of Educational Research, 80(6), 330-337. Federal Deposit Insurance Corporation (FDIC) (2006). New rules for revocable living trust accounts. Retrieved on April29, 2006 from http://www.fdic.gov/deposit/deposits/deposit/index.html. Finn, J.. D. (1989). Withdrawing from school. Review of Educational Research, 59, 117-142. Finn, J.D., & Volkl, K. E. (1993). School characteristics related to student engagement. The Journal of Negro Education, 62 (3). 249-268. Ford, D. (1993). Black student's achievement orientation as a function of perceived family achievement orientation and demographic variables. The Journal of Negro Education, 62(1), 47-66. Franke, R.H. and Kaul, J.D. (1978). The Hawthorne experiments: First statistical interpretation. American Sociological Review, 1978, 43, 623-643. Garibaldi, A.M. (1997). Four decades of progress. And decline: An assessment of African American educational attainment. Journal of Negro Education, 66(2), 105-12. Gay, L. R. & Airasian, P. (2003). Educational research: Competencies foe analysis and application, 6th edition. Upper Saddle River, NJ: Prentice Hall. George, D. and Malley, P. (2001). SPSS for windows step by step: A simple guide and reference 1.0 update, 3rd edition. Boston, Ma: Allyn and Bacon. Grantham, T. C. and Ford, D. Y. (2003). Beyond self-concept and self-esteem for African-American students: Improving racial identity improves achievement. High School Journal, 87(1),18-29. Greif, G. L., Hrabowski, F. A., & Maton, K. I. (2000). African American mothers of academically successful sons: Familial influences and implications for social work. Children and Schools, 22(4), 232-245(14). 125

PAGE 135

Hair, J. F. Jr., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis, 5th edition. Upper Saddle River, NJ: Prentice Hall. Hedges, L.V. & Nowell, A. (1998). Black-European American test score convergence since 1965. In C. Jencks, and M. Phillips. (Eds.) The Black European American test score gap. (pp. 103-145). Washington, D.C: Brookings Institute. Helfand, D. (2005, March 24). Nearly halfofBlacks, Latinos dropout, school shows. L.A. Times. Retrieved on December 13,2005 from http://www.latimes.com/news/local/la-me graduate24mar24,0,6007 681, full.story?coll = !a-home-headlines. Haug, M. R. & Sussman, M. B. (1971 ). The indiscriminate state of social class measurement. Social Forces, 49(4), 549-563. Hauser, R. M. & Warren, J. R. (1996). Socioeconomic indexes for occupations: A review, update, and critique. Madison, WI: Center for Demography and Ecology. Hodge, R. W., Siegel, P.M., & Rossi, P. H. (1964). Occupational prestige in the United States, 1925-63. American Journal of Sociology, 70(3), 286-302. Hoffman, K., Llagas, C., & Snyder, T. (2003). Status and trends in the education of Blacks. Washington, DC: U.S. Department of Education, National Center for Education Statistics. Iceland, J., Sharpe, C., & Steinmetz, E. (2003). Class differences in African American residential patterns in U.S. metropolitan areas: 1990-200. Paper presented at the annual meetings of the Population Association of America, Minneapolis, Minnesota, May 1-3, 2003. Iceland, J. (2003). Poverty in America. Berkley, CA: University of California Press. Jacob, E. & Jordan, C. (Eds.). (1993). Minority education: Anthropological perspectives. New Jersey: Ablex Publishing Corporation. 126

PAGE 136

Jencks, C., & Phillips, M. (1998). The black-white test score gap: An introduction. In C. Jencks and M. Phillips (Eds.), The black-white test score gap (pp. 1-51). Washington, DC: Brookings Institution. Johnson, A. G. (2000). The Blackwell dictionary of Sociology : A user's guide to sociological language. Malden, MA; Blackwell Publishers, Inc. Johnson, K. J., Crosnoe, R, Elder, G. H. (2001). Student's attachment and academic engagement: The role of race and ethnicity. Sociology of Education, 74 (4), 318-34. Johnson, S. T. (1992). Extra-school factors in achievement, attainment, and aspiration among junior and senior high school-age African American youth. Journal of Negro Education, 61(1), 99-119. Jordan, W. J. (1999). Black high school students' participation in school-sponsored sports activities: Effects on school engagement and achievement. Journal of Negro Education, 68(1), 54-71. Journal of Blacks in Higher Education (2004). College-bound Black students are making inroads in advanced placement tests. Retrieved on September 8, 2005 from http:/ /jbhe.com/features/44 _advanced _placement.html. Kaufman, P., Atl, M. N. & Chapman, C. (2004). Dropout rates in the United States: 2001 (NCES 2005-046). U.S. Department of Education, National Center for Education Statistics. Washington, DC: U.S. Government Printing Office. Kennickell, A. B. (2003). A rolling tide: changes in the distribution of wealth in the U.S., 1989-2001. Retrieved on November 12, 2004 from http:/ /ww. federalreserve.gov/pubs/oss/oss2/scfindex.html. Lee, C. D. (1993). Culturally responsive pedagogy and performance-based assessment. Journal of Negro Education, 67(3), 268-279. Lee, J. (2002). Racial and ethnic-achievement gap trends: Reversing the progress toward equity? Educational Researcher, 31(1), 3-12. 127

PAGE 137

Lee, V. E., & Smith, J. B. (1995). Effects of high school restructuring and size on early gains in achievement and engagement. Sociology of Education, 68( 4 ), 241. LadsonBillings, G. ( 1994 ). The dream keeper: Successful teachers of African American children. San Francisco: Jossey-Bass. Mayer, S. E. (1997). What money can't buy: Family income and children's life chances. Cambridge, MA: Harvard University Press. Myers, S. L. Jr., Kim, H., & Mandala, C. (2004). The effect of school poverty on racial gaps in test scores: The case of the Minnesota basic standards tests. Journal of Negro Education, 73(1), 81-98. Nakao, K. & Treas, J. (1994). Updating occupational prestige and socioeconomic scores: How the new measures measure up. Sociological Methodology, 24, 1-72. National Assessment of Educational Programs (NAEP) (2004a). Reading assessment. Washington DC: U.S. Department of Education, Institute of Education, and National Center for Educational Statistics. National Assessment of Educational Programs (2004b). Mathematics assessment. Washington DC: U.S. Department of Education, Institute of Education, and National Center for Educational Statistics. National Center for Education Statistics (2002). The condition of education in the brief. (NCES 2002-011). Washington DC: U.S. Department of Education: Office of Educational Research and Improvement. National Center for Education Statistics (1995). Extracurricular participation and student engagement. (NCES 95-741). Washington DC: U.S. Department of Education: Office of Educational Research and Improvement. Ogbu, J. U. (2003). Black American students in an affluent suburb: A study of academic disengagement. New Jersey: Lawrence Erlbaum. Ogbu, J. U. (1978) Minority education and caste: The American system n cross cultural perspective. NY: Academic Press. 128

PAGE 138

Oliver, M. L., and Shapiro, T. M. (1997). Black wealth, white wealth: A new perspective on racial inequality. Routledge NY: Routledge. Orr, J. A. (2003). Black-white differences in achievement: The importance of wealth. Sociology of Education 76( 4), 281-304. Phillips, M., Brooks-Gunn, J., Duncan, G.J., Klebanov, P.K., and Jencks, C. (1998). Family background, parenting practices, and the BlackWhite test score gap. In C. Jencks, & M. Phillips. (Eds.) The Black-European American test score gap. (pp. 103-145). Washington, D.C: Brookings Institute. Phillips, M., Brooks-Gunn, J., Duncan, G.J., Klebanov, P.K., & Jencks, C. (1993). Do neighborhoods influence child and adolescent development? TheAmerican Journal of Sociology, 99(2), 353-393. Power, M.G. (Ed.) (1982). Measures of socioeconomic status: Current issues. Boulder, CO: Westview Press, Inc. Roscingo, V. J., & AinsworthDarnell. (1999). Race, cultural capital, and educational resources: Persistent inequalities and achievement returns. Sociology of Education, 72(3), 158-178. Rothstein, R. (2004). Class and schools: using social, economic, and educational reform to close the black-European American achievement gap. Washington DC: Economic Policy Institute. Shapiro, T.M. (2004). The hidden cost of being African American: How wealth perpetuates inequality. NY: Oxford University Press. Sirin, R. S. (2005). Socioeconomic status and academic achievement: A meta-analytic review of researcher. Review of Educational Research, 75(3), 417-453. Smartmoney (2005). Debt management: Debt and your taxes. Retrieved on July 28, 2005 from http:/ /www.smartmoney .com/debt/advice/index.cfm?story = debttaxes. Smerdon, B. A. (1999). Engagement and achievement: Differences between African American and European American high school students. Research in Sociology of Education and Socialization, 12, 103-134 129

PAGE 139

Smith-Maddox, R. (1998). Defining culture as a dimension of academic achievement: Implications for culturally responsive curriculum, instruction, and assessment. Journal of Negro Education, 67(3), 302-317. Sowell, T. (1975). Race and economics. NY: David McKay. Stevens, G. & Cho, J. H. (1985). Socioeconomic indices and the new 1980 census occupational classification scheme. Social Science Research, 14, 142-168. Stevens, G. & Featherman, D. L. (1981). A revised socioeconomic index of occupational status. Social Science Research, 10, 364-395. Swanson (2004). Sketching a portrait of public high school graduation: Who graduates? Who doesn't? In G. Orfield. (Ed.) Dropouts in American: Conforming the graduation rate crisis. (pp. 13-40). Cambridge, MA: Harvard Education Press. U.S. Bureau of Census (2005). 2000 current population survey. Washington DC: U.S. Government Print Office. U.S. Bureau of Census (2002). 2000 census of the population. Washington DC: U.S. Government Print Office. U.S. Bureau of Census (1980). 1979 census of the population. Washington DC: U.S. Government Print Office. Wilson, K. R, and Allen, W. R. (1987). Explaining the educational attainment of young black adults: Critical familial and extra-familial influences. Journal of Negro Education, 56(1), 64-76. Woodson, C. C. G. (1933). The Miseducation of the Negro. Trenton NJ: African World Press, Inc. Yan, W. (1999). Successful African American students: Role of parent involvement Journal of Negro Education, 68, 5-22. 130

PAGE 140

Young, B. (2003). Public high school dropouts and completers from the common core of data: School year 2000-2001 (NCES 2004-310). U.S. Department of Education, National Center for Education Statistics. Washington, DC: U.S. Government Printing Office. 131

PAGE 141

APPENDIX Appendix A: Variables and Codes This appendix describes the variables used in this study. The variables come directly from the National Educational Longitudinal Study of 1988 (NELS:88) dataset. Data from the NELS:88 first-year and second-year follow provided information needed to conduct this current study. The NELS:88 variables present the variables taken from the NELS:88 dataset. The new composite variables are items developed using the NELS:88 variables 132

PAGE 142

NELS: 88 Variables Category Demographics Race-ethnicity Socioeconomic status Socioeconomic composite Wealth/ Assets Started a savings account Bought an insurance policy Made investments in stocks/real estate Bought U.S. savings bonds Established another form of savings Planned to re-mortgage property/take loan Used savings/assets for teens ed Used 2nd mortgage for teens ed Used trust fund for teens ed Income Father's Education and Occupation What is respondent's father's occupation Father's highest level of education Mother's Education and Occupation Academics What is respondent's mother's occupation Mother's highest level of education Engagement How often the student is absent How often the student does homework How often the student is attentive in class Participated on a team sport at school Participated in individual sport at school Participated in cheerleading/pompom Participated in school music group Participated in school play or musical Participated in school government Participated in academic honor society 133 Variables BYRACE F1SES F1SESQ F2P79A F2P79B F2P79D F2P79C F2P79G F2P791 F2P92B F2P92C F2P92G F2P74 F2N7 BYS34A F2N5 BYS34B F1T1 16 F1T1 15 F1T1 18 F2S30AA F2S30AB F2S30AC F2S30BA F2S30BB F2S30BC F2S30BD

PAGE 143

Category Academics (Continues) Engagement Participated in school yearbook, newspaper Participated in school service clubs Participated in school academic clubs Participated in school hobby clubs Participated in school FT A, FHA, FF A Year Respondent took the SAT Year Respondent took the ACT Year Respondent took AP test Achievement Standardized Test Scores (reading and math) Attainment Transcript-indicated outcome 134 Variables F2S30BE F2S30BF F2S30BG F2S30BH F2S30BI F2S44BYR F2S44CYR F2S44DYR F22XCOMP F2RTROUT

PAGE 144

New Composite Variable (NCV) Demographics Race NELS:88: F2RACE 1 = Asian, Pacific Island 2 =Hispanic 3 =Black, Non-Hispanic 4 =White, Non-Hispanic 5 = American Indian NCV: Race Recoded 3 = African American 4 = European American The Occupational Prestige Scores of Mothers and Fathers NELS:88: F2N5 and F2N7 1 = Clerical Worker 2 = Craftsperson 3 =Farmer 4 = Homemaker 5 =Laborer 6 =Manager 7 =Military 8 = Machine Operator 9 = Professional I 10 = Professional II 11 = Small Business Owner 12 = Protect Service 13 =Sales 14 =School Teacher 15 = Service Worker 16 =Technical Worker Absences NELS:88: F1T1 16 1 =Never 2 =Rarely 3 = Some of the time 4 = Most of the time 5 = All of the time Type 2 Engagement NELS:88: F2S30AA 135 NCV: Occupational Prestige Scores 56.68 =Clerical Worker 27.41 = Craftsperson 28.00 =Farmer 0.00 = Homemaker 7.3 3 = Laborer 67.33 = Manager 0.00 = Military 19.18 = Machine Operator 70.21 = Professional I 70.21 = Professional II 49.70 =Small Business Owner 38.00 =Protect Service 54.24 =Sales 70.21 = School Teacher 15.90 =Service Worker 61.40 =Technical Worker NCV: Absences Reverse Coded 5 =Never 4 =Rarely 3 = Some of the time 2 = Most of the time 1 = All of the time NCV: Team Sport

PAGE 145

1 = School doesn't have 2 =Didn't participate 3 = Junior Varsity 4 =Varsity 5 = Captain/Co-Captain NELS:88: F2S30AB 1 =School doesn't have 2 =Didn't participate 3 = Junior Varsity 4 =Varsity 5 = Captain/Co-Captain NELS:88: F2S30AC 1 = School doesn't have 2 =Didn't participate 3 =Junior Varsity 4 =Varsity 5 = Captain/Co-Captain NELS:88: F2S30BA 1 = School doesn't have 2 = Didn't participate 3 = Participated 4 = Officer/Leader NELS:88: F2S30BB 1 = School doesn't have 2 = Didn't participate 3 = Participated 4 = Officer/Leader NELS:88: F2S30BC 1 =School doesn't have 2 =Didn't participate 3 = Participated 4 = Officer/Leader NELS:88: F2S30BD 136 1 & 2 receded to 1 = did not participate 3, 4, & 5 receded to 2 =participated NCV: Individual Sport 1 & 2 recoded to 1 = did not participate 3, 4, & 5 receded to 2 =participated NCV: Cheerleading 1 & 2 receded to 1 = did not participate 3, 4, & 5 receded to 2 =participated NCV: Music Club 1 & 2 receded to 1 = did not participate 3 & 4 receded to 2 = participated NCV: Plays/Drama 1 & 2 receded to 1 = did not participate 3 & 4 receded to 2 = participated NCV: Student Government 1 & 2 recoded to 1 = did not participate 3 & 4 receded to 2 = participated NCV: Honors Society

PAGE 146

1 =School doesn't have 2 =Didn't participate 3 = Participated 4 = Officer/Leader NELS:88: F2S30BE 1 =School doesn't have =Didn't participate 3 = Participated 4 = Officer/Leader NELS:88: F2S30BF 1 =School doesn't have 2 = Didn't participate 3 = Participated 4 = Officer/Leader NELS:88: F2S30BG 1 =School doesn't have 2 =Didn't participate 3 = Participated 4 =Officer/Leader NELS:88: F2S30BH 1 = School doesn't have 2 =Didn't participate 3 = Participated 4 = Officer/Leader NELS:88: F2S30BI 1 = School doesn't have 2 =Didn't participate 3 = Participated 4 = Officer/Leader SAT NELS:88: F2S44BYR 90 = 1990 137 1 & 2 recoded to 1 = did not participate 3 & 4 recoded to 2 = participated NCV: Yearbook Club 1 & 2 recoded to 1 = did not participate 3 & 4 recoded to 2 = participated NCV: Service Club 1 & 2 recoded to 1 = did not participate 3 & 4 recoded to 2 = participated NCV: Academic Club 1 & 2 recoded to 1 = did not participate 3 & 4 recoded to 2 = participated NCV: Hobby Club 1 & 2 recoded to 1 = did not participate 3 & 4 recoded to 2 = participated NCV: FT AIFHAIFF A 1 & 2 recoded to 1 = did not participate 3 & 4 recoded to 2 = participated NCV: SAT 97 recoded to 1 = did not take

PAGE 147

91 = 1991 92 = 1992 93 = 1993 97 =Refused ACT NELS:88: F2S44CYR 90 = 1990 91 = 1991 92 = 1992 93 = 1993 97 =Refused AP Test NELS:88: F2S44DYR 90 = 1990 91 = 1991 92 = 1992 93 = 1993 97 =Refused On-time graduation NELS:88: F2RTROUT 1 = Spring 1992 Graduation 2 = Other 1992 Graduation 3 = Pre-1992 Graduation 4 = Diploma/Special Education 5 = Certificate of attendance 6 = Still Enrolled 7 = Dropped Out 8 = Transferred 9 =Aged Out 11 = Left/Health Reason 12 = Received GED 138 90, 91, 92, & 93 recoded to 1= did take NCV:ACT 97 recoded to 1 = did not take 90, 91, 92, & 93 recoded to 1= did take NCV:AP 97 recoded to 1 = did not take 90, 91, 92, & 93 recoded to 1= did take NCV: On time graduation 6, 7, 9, & 11 recoded to 1 =Did Not Graduate in 1992 4, 5, & 12 recoded to 2 = Graduated with alternative diploma 1, 2, & 3 Graduated with a diploma in 1992