Keeping at-risk students in school

Material Information

Keeping at-risk students in school the effects of social capital, family structure, peer influence and self concept
Amundson, Erin Lynn
Publication Date:
Physical Description:
vii, 72 leaves : ; 28 cm


Subjects / Keywords:
Dropouts -- Prevention ( lcsh )
Social capital (Sociology) ( lcsh )
Family demography ( lcsh )
Peer pressure in children ( lcsh )
Self-perception in children ( lcsh )
Dropouts -- Prevention ( fast )
Family demography ( fast )
Peer pressure in children ( fast )
Self-perception in children ( fast )
Social capital (Sociology) ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 65-72).
General Note:
Department of Sociology
Statement of Responsibility:
by Erin Lynn Amundson.

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:
55201014 ( OCLC )
LD1190.L66 2003m A58 ( lcc )

Full Text
Erin Lynn Amundson
B.A., University of Colorado, Boulder
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts

This thesis for the Master of Arts
degree by
Erin Lynn Amundson
has been approved
/2 // Jo 3
Virginia Fink

Amundson, Erin Lynn (M.A., Sociology)
Keeping At-Risk Students in School: The Effects of Social Capital, Family Structure,
Peer Influence and Self Concept
Thesis directed by Associate Professor Candan Duran Aydintug
A number of factors influence students academic attainment. This study
applies a filter to the National Education Longitudinal Survey, 1988, Second Follow
up to include only those students defined as at-risk. ANOVA analysis and logistic
regression coefficients show that positive peer influence, participation in social
capital-forming activities, positive self-concept and high internal locus of control are
all factors that help prevent at-risk students from dropping out of high school.
Social capital in the form of human interaction was not significant to the model,
suggesting that involvement in group activities such as sports or youth groups is a
more important determinant of educational success.
Positive self concept and high internal locus of control can be influenced by
many factors. The results of this study suggest that students who have a negative
self concept or a low internal locus of control need more positive reinforcement that
could be provided through schools.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
Candan Duran Aydintug

I would like to thank my family for the constant encouragement to pursue my
educational dreams. I also want to thank my advisor Candan Duran Aydintug for
believing in and supporting my research vision. I would also like thank Yili Xu for
his tremendous help with the statistical component and Virginia Fink for her
inspiration and encouragement.

1. INTRODUCTION..............................................1
2. THEORETICAL PERSPECTIVES..................................4
3. LITERATURE REVIEW.......................................14
Action Research Perspectives/School Responsibility....14
Social Capital........................................16
Familial, Teacher, School And Peer Influences.........17
Differences In Class Values...........................28
Reproducing Stereotypes............................. 28
Parental Education Effects............................30
Student Locus of Control..............................31
4. AIM OF THE STUDY/HYPOTHESIS..............................33
5. METHODS..................................................35
Procedure........................................... 37
Measurement of Variables/Concepts.....................42

ANOVA Results....................................50
Logistic Regression Results......................55
7. DISCUSSION..........................................58

7.1.1 ANOVA for Family Structure, Peer
Influence, Social Capital and Self Concept...............................50
7.2.1 Logistic Regression for Family Structure, Peer
Influence, Social Capital and Self Concept...............................55

Throughout history, several researchers and educators have made attempts at
improving educational standards. Our childrens education is a top concern of
parents, politicians, and everyday citizens of the country. Most recently, research
has focused actively and theoretically on possibilities for increasing the numbers of
high school graduates (Loveless, 1997), improving the quality of education
(Loveless, 1997) and providing group programs (Loveless, 1997; Alexander,
Entwisle and Horsey, 1997) that best suit individual needs of students at risk of
dropping out of school short of a high school diploma. According to the National
Center for Education Statistics (NCES), 10.9 percent of the 34.6 million 16- through
24-year olds were classified as dropouts with no supplemental high school
credentials in 2000 (U.S. Department of Education, 2000). The NCES report also
found that Blacks continue to have a higher dropout rate than Whites (13.1 percent
compared to 6.9 percent), and that Hispanics have higher dropout rates than any
other race (27.8 percent) (U.S. Department of Education, 2000). Lewis, Ross and
Mirowsky (1999) found that dropping out of high school reduces the perceived
internal locus of control of the student once he or she reaches adulthood. Dropping
out of high school reduces the amount of cognitive skills a person has which

negatively affects the internal locus of control. Losing this sense of control prevents
a person from having an attitude that promotes success, possibly preventing a person
from going back to school or pursuing a successful career. Researchers have also
found that high school dropouts commit both violent and petty crimes at a
significantly higher rate than non-dropouts (Felson, Liska, South and McNulty,
1994; Jenkins, 1995) and at the least, students who drop out of high school are much
more likely to be unemployed or to live in poverty later in life (Kerckhoff and Bell,
1998). Several researchers and policy makers (Teachman, Carver and Paasch, 1995)
have used NCES data to attempt to determine the reasons some students dropout and
others succeed. Others (Coleman and Hoffer, 1987; Kalmijn and Kraaykamp, 1996;
White and Glick, 2000) have examined the gap in dropout rates between White
students and Black or Hispanic students of native or immigrant status. All of these
projects find that students from negligible backgrounds (at-risk students) are more
likely to dropout than advantaged students. Despite the answer that being at-risk is
what causes students to drop out, despite the efforts of these researchers and policy
makers, and despite the endless number of research projects revealing patterns of
high school dropouts and community education needs, one thing is clear: Some at-
risk students are more likely to dropout than others. Determining the why some at-
risk students did not drop out of high school may help other potential at-risk students
choose instead to graduate. The question that remains asks: what are the
contributing factors that lead some at-risk students to graduate and other at-risk

students to dropout? The most recent literature attempts to answer this question
employing a wide range of theoretical perspectives.

Ability grouping theory suggests that each student is bom at a different
ability level, and should be placed in school accordingly. Ability grouping claims to
put students of similar intelligence levels in the same classes with the idea that the
more intelligent students will not be held back academically by the less intelligent
students. Hoffer (1992) claims that grouped students, whether placed in higher
ability classes or lower ability classes, perform better academically than students
who are not grouped according to ability.
On the other hand, social stratification theory suggests that students are
grouped not according to ability but rather according to some other social trait such
as race, class or gender. Therefore, students who are placed in the wrong ability
group will suffer. Social stratification theory also claims that in ability grouped
situations, the lower-grouped students will inevitably suffer for not being exposed to
the same level of education as higher grouped students. Neither ability-grouping nor
social stratification theories take into account personal background characteristics of
students (such as abuse or other psychological damage), nor do they determine
whether or not specific students actually belong in the groups they are assigned to.
In action research, (Zinck and Littrell, 2000) solution-focused theory is used

to suggest that one positive change in a persons life is sufficient to initiate a series
of changes that will ultimately improve that persons overall quality of life.
Specifically, Zinck and Littrell (2000) suggest that students who are at risk of
dropping out of school due to their involvement with a large number of dangerous
behaviors or because they are being raised in negligible surroundings will benefit
from the introduction of one positive and lasting experience. The understanding is
that these students often feel they have nothing to gain and nothing to lose, but that a
significant positive experience may convince them otherwise. Prevention programs
such as the Big Brother or D.A.R.E. programs were based on this theory.
Unfortunately, who will and will not benefit from such programs is a gamble, at best.
Social network theory claims that social relationships represent distribution
points for social resources and opportunities. This theory is often used to test the
effects of the upper class exposure to advantageous social resources. Alienation
theory, on the other hand, focuses on those who lack access to social resources.
Because it is nearly impossible to take social resources away from students that come
from privileged backgrounds, Croninger and Lee (2001) suggest testing the effects of
giving access to social resources to students who have not previously had such
access. Croninger and Lee (2001) imply that such an experiment will determine
whether the disadvantaged student can achieve as much success on average as an
advantaged student when given the same social opportunities. However, Croninger
and Lee (2001) fail to take into account the psychological effects of never having

had access to resources and then suddenly having resources. For example, it cannot
be assumed that students with no prior knowledge of computers would benefit from
simply being provided with a computer. They also fail to address the impact of
parents attitudes about school and education on students motivation for education,
with or without resources. Both of these factors impede resource advantages (Ross
& Wu, 1997; Ross & Van Willigen, 1997).
In addition to theories about social structures, economic theory is often used
to explain inequalities in education. As applied to education, economic theory
claims that childrens success depends on the combination of sufficient economic
resources and sufficient care-giving in the home. For example, Whitbeck, Simons,
Conger, Wickrama, Ackley & Elder (1997) found that economic strain in families
negatively affects parenting behaviors which contributes to adolescents self-efficacy
as it relates to school. Alexander et. al. (1997) suggest that the socioeconomic status
of a childs family causes the child to drop out regardless of the family structure.
Teachman (1987) and Bianchi and Robinson (1997) suggest that single-parent
households in which the parent works full-time constitute a disadvantage to the
academic success of young children because there is no adult nurturer available.
However, these studies do not account for two-parent families in which both parents
work. In this case, according to economic theory, the availability of a nurturing
parent is also in question. Economic theory as applied to education needs to be
revised to include the now common exceptions to these social rules.

Other studies (Oakes, 1995 and Oakes and Wells 1998) suggest that the
theory of cultural capital provides a clear explanation for the differences in success
levels of students. As Pierre Bourdieu and Jean Claude Passerson (1977) claim in
their classic work Reproduction in Education, Society and Culture, cultural capital is
defined when schools bring together students of all different cultures and family
backgrounds, and teachers teach the same curriculum to everyone. What is often
missed is that each student obtains different types of knowledge at home based on
the economic and cultural situation of their family. In simpler terms, each childs
level of success in school is dependent on what the child has already been taught.
Bourdieu and Passeron (1977) claim that family social class, defined by family
income and inheritance, is an extremely significant determinant of how complex a
students language skills are. Language provides a system of classification for
complex logical and aesthetic systems, which comprise the basis of academic
material used to determine a students academic performance value. Specifically,
Bourdieu and Passeron (1997) claim that the higher the class status, the more likely it
is that the student has been previously exposed to pedagogic work similar to what
they will be expected to know in school. In addition, they claim that a measurement
of linguistic ability (or such other test of knowledge such as the SAT) cannot
possibly take into account the social origin, sex or race of the student, and thus is
unfairly biased towards the dominant class.

In addition, Bourdieu and Passeron (1977) point out that the school systems
supply an insufficient study population to look at in terms of determining linguistic
capital or educational success. This is because those students who do not make it
through the system into higher education are left out and thus there is no comparison
group to study along side the dominant group. Therefore, as Bourdieu and Passeron
(1977) claim, by looking at only the survivors of the system, it is culturally
ignored that school systems are set up to reproduce long-standing social inequalities.
Bourdieu and Passeron (1977) suggest that tracking systems place lower class
students in remedial classes based on the exposure to cultural knowledge in the
home. Cultural knowledge exists in the form of exposure to historical museums,
theatrical productions or art, to name a few, and this exposure is extremely varied
according to class, race and gender. When lower class students are placed in lower
achieving tracks, they no longer have access to the courses that teach materials that
are covered on standardized placement exams, and are therefore disadvantaged
towards falling through the cracks of the educational system. This lack of success
is attributed to laziness or a lack of academic prowess, rather than the fact that there
was never a chance for these students to learn what they need to be able to pursue
higher education.
Bourdieu and Passeron (1977) found economic background to be the largest
contributing factor to educational disparity. They considered religion as another

factor and found that the religiousity of a respondent in any social class has no real
bearing on educational attainment (Bourdieu and Passeron, 1977). Bourdieu and
Passeron note that while the upper class students tend to devote their time to
extracurricular activities that enrich their cultural knowledge (art, music, theater),
working class students are much more likely to have to spend leisure time working
or helping to care for their families (Bourdieu and Passeron, 1977). As a result, the
lower class students are further separated from the upper class society, not having the
know-how to perform culturally in a social setting, which is an inherent part of being
a member of the upper class. Recent studies (Bianchi and Robinson, 1997; Buchel
and Duncan 1998) have found a positive correlation between parental involvement in
cultural and social activities (such as participation in sports or attending museums or
libraries regularly) and childrens educational aspirations. In addition, parents who
report engaging in community volunteer work are more likely to have children who
are academically successful (Buchel and Duncan 1998).
James Coleman (1988) has a more recent and somewhat more specific take
on social capital. He suggests that three types of capital: financial, human, and
social, are present in every childs home in different amounts. Financial capital
defines parents income level as well as the provisions of the income. When related
to childrens success in school, financial capital specifically refers to materials in the
home designed to aid study (a library, computer, etc.) that parents are able or unable
to purchase because of their level of income. Human capital, according to Coleman

(1988), is the parents level of education. Well-educated parents in general, are
better able to provide their children with an environment that will foster the childs
own success in school. Coleman (1988) defines social capital as the relationship
exchange between parent and child. Children whose parents have relationships with
people of power are included in these relationships by default and therefore have an
advantage over other children from the beginning of their lives. The interesting twist
to this separation of types of capital is that a parent with a high amount of human
capital does a child no good if there is a lack of social capital. On the other hand, a
parent with no human capital (education) can provide advantages of human capital to
their children through social capital (having relationships with well-educated adults).
Coleman (1988) refers to an example of an under-educated Asian family purchasing
two textbooks so that an adult may study with the child and assist in the childs
learning of the given material.
Coleman (1988) refers to connections outside the family as social structure.
He claims there is an important difference between an open and a closed social
structure. Closure is important for developing norms and tends to occur among a
group of people who are all familiar with one another. In addition, closed structures
provide a sense of community trust, and Coleman (1998) argues that it is easier to be
successful in a trusting community. An example of a closed social structure would
be a small town in which there is limited choice of schools, stores and places of
worship. Almost everyone in the town would attend the same schools, stores and

places of worship, thereby becoming familiar with one another. Open structures, on
the other hand, may have more connections involved, but these connections are
loose. Open structures may allow an individual a loose connection to an important
figure, but the individual has no chance of developing a reputation in an open
structure, which may be important for acceptance. An open structure would more
likely occur in a larger city in which a family does not associate itself with one
particular community, but rather chooses to take advantage of a wide range of social
connections. However, since none of the social connections are connected to each
other, the family does not necessarily have a chance to develop a reputation.
In more recent work, Coleman (1992) suggests that the rise in single parent
families and maternal employment led to fewer adults at home to supervise
childrens activities, less energy spent reading to or with younger children or helping
with homework, less time spent volunteering in the schools and an overall decrease
in the time spent building social networks to provide resources for children.
Following Coleman (1988), McNeal (1999) notes that there are generally
three distinct elements of social capital that researchers must focus on in order to
capture the essence of the theory. He claims that social capital should be
operationalized according to structure of social ties, including depth of social
network, intensity of relationships, and existence of structural holes. He also
believes that the definition of social capital should include the notion that there is
some sort of investment going on through social ties and that a return is expected

(McNeal, 1999). That is to say that connections are considered to be two-way, and
thus it is difficult to develop a helpful social tie for a person who has little or nothing
to offer in return. The third central concept McNeal includes in his definition of
social capital is that it should account for both resources inside the particular
persons social network and potential resources that may exist outside the network
(1999). Thus, in order for a student to have a lot of social capital, that student would
have to visualize the type of success he or she wants, access ties that will help his or
her situation immediately, gain knowledge of resources that are not readily available,
and figure out a way to connect with these resources. Based on this notion, an adult
with an extensive knowledge of helpful resources, and the knowledge and
willingness to point a student in the right direction when something beyond the
particular adults scope of ability comes into the picture would be highly beneficial
to a student at risk. However, at-risk students do not often have anything to offer in
return, and thus, according to McNeal, would be unable to find lasting useful
connections (1999).
Sandefur and Laumann (1988) discuss social capital in terms of three benefits
(as opposed to forms) including information, influence and control, and social
solidarity. Information, in terms of students social capital, is a benefit when parents
form a closed structure around their children with other adults that can give them
information about their children (school progress, behavior) and the parents can then
act accordingly (Sandefur and Laumann, 1988). Influence and control exists as both

the power to influence others (i.e., to convince a school that Spanish-speaking
children are worth the extra effort of bi-lingual education) or to remain free of the
influence of others (i.e., a student that is not negatively influenced by friends
decisions to drop out of school) (Sandefur and Laumann, 1988). Social solidarity
exists when there is a degree of mutual trust and commitment among members of
any community. According to Stanton-Salazar (2001), social solidarity becomes a
problem for Mexican-American youth in schools because they are raised in
communities in which the only people they can trust are family members. Without
easily establishing trust in looser ties, these youth are limited to the networks and
information of their families. In most, cases, these parents do not speak English and
have no ties outside family, severely limiting many Mexican-American youth from
Within this framework, it is easy to see where parental involvement becomes
social capital. However, what other kinds of social interaction might provide capital
for at-risk students?

Action Research Perspectives/School Responsibility
Action researchers, such as Joy Dryfoos (1998), suggest the idea of
intervention programs designed to aid at-risk students. Dryfoos refers to three
separate types of intervention used in programs designed to prevent youths from
entering a life of delinquency. Primary intervention focuses on the entire population
and uses general preventative measures, such as the education found in the D.A.R.E.
Program. Secondary intervention programs are more specific and targeted toward a
direct need of a specific at-risk group. An example would be a program created to
help children of abusive parents with school. Tertiary intervention programs are
employed to help reverse delinquent behavior and often include counseling or
incarceration (Dreyfoos, 1988).
Dreyfoos (1998) has counted over 85,000 public and 25,000 private non-
profit organizations designed to aid at-risk students, as well as 400 nationally based
programs with local chapters (i.e., Girl Scouts, 4-H) and more than 17,000
community-based organizations all working to help youth. These are just among
those easily countable, and yet with all these programs in place, something is still

amiss as the number of delinquent and at-risk students continues to grow yearly.
Some students benefit from these programs and move on to higher education.
(Dreyfoos, 1988). Others do not, and it is still unclear what the differences
are between the at-risk students who succeed and the at-risk students who do not.
Other action researchers have developed programs of strategic learning, in
which particular books, entertaining to students of a specific age group, are selected
and discussed in the contexts of gender roles, mathematical thinking, and critical
analysis. (Hubbard, Barbieri, and Power, et. al., 1998) These programs cover the
span of learning development and encourage students at a young age to think
critically for themselves, regardless of race, gender, family income or any of the
other factors that have deterred students from accelerating in academic achievement.
One such project encouraged fourth grade students to embark on a high school level
research project (Hubbard, Barbieri, and Power, et. al. 1998). This project,
measuring students heights, weights and shoe sizes of students in their age group
snowballed as the students decided to expand the project to include measurements
from other countries. Encouraged to think critically about what might come of such
a project, students figured that shoe and clothing companies and television producers
would be interested in this marketable information and wrote letters to several of the
companies offering their services. Such programs are unfortunately rare and
unrealistic for every school system. However, it is evident that given the mix of
students from different races, socioeconomic status and gender, and given the

success at critical thinking each student showed, that it is possible for at-risk students
to excel under the right circumstances.
Social Capital
More recently, Cronginer and Lee (2001) suggested that social capital in the
form of knowledge, resources and job marketability, is the key to providing students
with the necessary means to market themselves in our economy. Researchers refer
to two types of risk for dropping out of school, social (Luthar & Zigler, 1991) and
academic (Catterall, 1998). A lack of social capital is an obvious pair to social risk.
Students who do not possess a high level of social capital often fall behind in school
and eventually drop out. Several studies link high school dropouts to low paying
wages and criminal behavior (Coleman, 1990, Stanton-Salazar, 1997). Social capital
has also been linked to preventing dropouts based on academic risk (LeCompte &
Dworkin, 1991, Wellman, 1983). These studies all use theories of alienation to
describe how a lack of social capital can negatively affect an at-risk student.
Problems with these studies are that the students are already at-risk due to such a
multitude of other factors that it becomes difficult to accurately measure what kind
of damage is really done by a lack of social capital. In addition, these studies
provide, at best, vague definitions of social capital, so the links between amount of
social capital and social risk or dropping out of school are unclear.

Familial. Teacher. School And Peer Influences
Brooks-Gunn, Duncan, Klebanov & Sealand. (1993) found that certain
neighborhood characteristics have an effect on behavioral patterns of adolescents.
For example, they found significant associations between having more affluent
neighbors and having better developmental outcomes, including IQ scores and non-
pregnancy in teen years. However, they found that increasing the ratio of affluent
families to non-affluent families in low-income neighborhoods did not change
developmental outcomes of the adolescents living there, suggesting a possible causal
error. Their findings suggest that family income, rather than neighborhood
influence, affects adolescent behavior and education outcomes. In addition, their
study reinforces previous findings that suggest family income and mothers
education are powerful and significant predictors of adolescent success in school.
To double check their findings, the authors also examined the effect of moving from
a less affluent neighborhood to a more affluent neighborhood on IQ scores and found
that in the case of significant increase (17.2% or more) in affluent neighbors, IQ
scores improved significantly for the adolescents living there. Overall, their
findings suggest that the absence of low-income neighbors is less important to
affecting behavioral outcomes than the presence of higher-income neighbors. This
suggests that while children from low-income families have a higher tendency to
exhibit delinquent behaviors, these behaviors can be off-set if the low-income family

lives in a neighborhood surrounded by higher income families (Brooks-Gunn, et. al.,
1993). This study suggests overall that peer influence has an effect on childrens
behavior, sometimes overpowering the effect of socioeconomic status.
In addition, researchers are now defining social capital in terms of familial
sources and non-familial sources. Cronginger and Lee (2001), look at the positive
effects social capital can have on an at-risk student who might not have otherwise
made it through school. However, Cronginger and Lees (2001) definition of social
capital shifts from the traditional definitions provided by Bourdieu and Passeron
(1977) and Coleman (1988) to suggest that social capital can be provided by
educators. Croninger and Lees (2001) study seeks to measure the effect that
student-teacher contact has on the success of at-risk students. However, with a
skewed definition of social capital and no control for such variables as race,
socioeconomic status and base academic ability, the results of this study are
somewhat misleading.
Dika and Singh (2002) provide an extensive review of social capital in
education literature and found that socioeconomic status is positively related to
childs success in school and also that parents knowledge of childrens friends is
strongly associated with childrens future socioeconomic outcomes, suggesting that
parental monitoring of childrens activities has a positive influence. In addition, they
found that the childs friends education expectations were a strong influence on the
childs own expectations. However, as Biddle (2001) suggests, studies such as these

that measure childrens educational success based solely on measures of parents
SES, which include occupation and income level, are misleading in that occupation
and income level often have very different and distinct effects on childrens
educational success.
In research studying the effects of family-based and school-based social
capital, Parcel and Dufur (2001) found that in families where there was smaller
amounts of human capital available (undereducated parents), students who attended
schools with higher amounts of human capital had better behavioral outcomes,
including measures of academic success. In addition, Parcel and Dufur (2001) found
that attending a private catholic school provided an advantage to students with lower
social capital in the home, suggesting that the closed community provided by the
private schools mix with religion is a beneficial form of social capital. Parcel and
Dufur (2001) claim that parental financial capital is a significant predictor of school
social capital, most likely because higher amounts of financial capital provide a
means of ensuring that their children attend better quality schools. Parcel and Dufur
(2001) found that students from homes that lacked human and financial capital could
have the same positive behavioral outcomes as those students who did not lack
human and financial capital if that lack of capital was offset by another form of
social capital, such as church attendance (involvement in the neighborhood). The
question their study fails to address is whether or not the effects of a lack of human
or financial capital at home can be offset by attending a school with higher amounts

of capital, though not necessarily a private catholic school. Their research suggests
that it is possible.
Teachman, Paasch & Carver (1996) found that students who attend Catholic
schools have more access to social capital, most likely because of Colemans notion
of the closed structure. These children are much more likely to have parents that
know the parents of their friends and the friends themselves. In addition, they are
less likely to have changed schools, have parents who are more connected with their
schools and experience more interaction with their parents. Living a with divorced
father or mother or with a step-parent are all also related to less social capital.
Positive parent child interaction decreased the chances that students dropped out of
school. Again, the more human and financial capital parents have, the less likely
students are to drop out of school.
Further, Sewell and Shah (1968) found that the higher the parents education,
the greater the students perception of parental encouragement to go to college was.
They also found that the more intelligent a student was, the higher perception of
parental encouragement they sensed, regardless of social class and sex differences
among respondents. Also, those that perceived parental encouragement were
significantly more likely to attend and graduate from college than those who had a
low perception of parental encouragement.
Goyette and Conchas (2002) specifically explore the effects of familial and
non-familial social capital on Vietnamese and Mexican American youth with

NELS:88 data. Their study suggests that base differences in cultural background
may contribute to differences in amounts of family social capital. On average,
Vietnamese American families have higher socioeconomic status than Mexican
American families (Goyette and Xie, 1999), which suggests a higher level of social
capital. However, Goyette and Conchas analysis of Vietnamese American and
Mexican American students suggests that Vietnamese American students actually
have less familial social capital in general than Mexican American students.
Because Vietnamese American students perform better academically than Mexican
American students, familial social capital is ruled out as the explanatory factor.
Goyette and Conchas (2002) suggest that teacher and peer support networks are
more favorable for Vietnamese Americans than for Mexican Americans, explaining
the difference in academic performance level. In other words, peer or teacher
influences may be more beneficial conduits of social capital than family influences.
As Mark Granovetter (1973) suggests, weaker ties are often more valuable
sources of social capital because stronger ties are likely to run into the same network
of information the individual in question already has contact with. In addition,
many researchers (Steinberg, Dombusch and Brown, 1992; Mehan et al., 1996
Valenzuela, 1999;) recognize that peer groups have a large amount of influence on
students educational success or failure. Some at-risk students, particularly
members of immigrant or racial minority groups, depend more on peers for network
support than on parents, who may not understand how school systems work.

Focusing on family socioeconomic status, Whitbeck, et. al. (1997) ran
separate models for working mothers and working fathers and found that job type
had an effect on parenting style, which in turn had an effect on adolescents self-
efficacy. More specifically, they found that fathers with more autonomous jobs
(higher level, higher income) were more likely to use inductive parenting techniques
as opposed to harsh disciplinary techniques and their adolescent children were more
likely to have a higher sense of self efficacy. They found basically the same for
women, although they found that their indicators for workplace autonomy for
working mothers was somewhat different, suggesting that flexibility was the most
important factor to women in work autonomy (rather than being own boss or having
lots of freedom).
In a study restricted to Baltimore, Alexander, et. al. (1997) found that parents
attitudes and values about education in the early years of the students lives directly
effected which students dropped out and which did not. Teachman (1996) found that
family characteristics do not bias the effects of intellectual skill on a students
academic performance in that matched sibling pairs often had variable links between
their intellectual skill and academic performances.
Hyjer Dyk and Wilson (1999) found that children of Appalachian families in
which mothers had higher levels of education had their own higher aspirations for
jobs. In addition, youth from larger families had higher educational and labor force
expectations, possibly due to increased chances of social networking that would

naturally follow. In addition, Hyjer Dyk and Wilsons (1999) results suggest that a
lack of resources may be motivation enough for families to encourage and support
higher educational aspirations as the sample of this study were all from a lower
resource area. This study also found that students who talked with parents or
family members about their education or future job prospects were more likely to
pursue higher levels of education and occupational attainment. Overall, the
researchers suggest that low-income, Appalachian parents seem to compensate for
the lack of financial capital with other forms of social capital such as Coleman
Other researchers prefer to investigate the effects of minority status by
comparing minority students to white students rather than to other minority groups.
Ricardo Stanton-Salazar (2001) refers to social capital as apart of social networks
that shape structural oppression of race and class in terms of access to career
mobility, political power and institutional privileges and resources. Stanton-Salazar
relies on attachment theories (Wallace and Vaux 1993) that determine childrens
levels of sociability, empathy, resiliency, and self-respect. Researchers indicate that
children with positive help-seeking orientation based on their healthy attachment
style are more likely to seek advice and resources from a wide range of support
systems, more likely to develop strong relationships, and more likely to feel socially
connected, important aspects determining whether or not a student will seek
assistance from the social context of the school. (Vaux, Burda & Stewart, 1986).

Stanton-Salazar (2001) suggests that school systems invariably have the power to
make or break such low-class or ethnic minority students.
School personnel, by virtue of their status as institutional
agents and their potential positioning within the social webs of
working-class youth, are structurally endowed with extraordinary
powersfirst and foremost, the capacity to protect the adolescent
from the worst effects of class reproduction, and second, the capacity
to act as a transformative influence in the lives of their low-status
students (Stanton-Salazar in press cited in Defensive Network
Orientations as Internalized Oppression How Schools Mediate the
Influence of Social Class on Adolescent Development. (2001 pg
Unfortunately, as several researchers point out, schools in general do not accept this
responsibility, and in fact, reinforce the internalized racism and classism experienced
by disadvantaged students. Salazar (2000) points out that norms dictate the
undesirability of seeming incompetent in school systems, preventing students from
seeking the assistance they may need. Salazar (2000) also argues that students have
difficulty accessing assistance from teachers and counselors due to the students fear
of rejection or due to the teachers and counselors busy schedules. In addition, large
school populations increase the competition for students to enter into select academic
programs, athletic programs and other extra-curricular activities that could be a
potential source of networks for students lacking social capital (Barker & Gump,
1964; Berliner & Biddle, 1995).
Stanton-Salazar (2001) suggests that intervention programs that are designed
to incorporate professionals who are committed to the success of academically

competent at-risk youth are needed. He claims that someone needs to teach at-risk
youth to decode academic culture in order to better understand the norms that must
be followed for social success and to provide a social support network for at-risk
students that will empower previously silenced students (see Rueveni, 1979;
Cochran, 1990). While many people in our society believe that parents are
responsible for the academic motivations of their children, the reality is that some of
these students parents do not even speak English. Stanton-Salazar, Mehan,
Villanueva, Hubbard & Lintz, (1996) and Stanton-Salazar, Vasquez & Mehan
(1996) studied the success of such a program, Advancement Via Individual
Determination, and found that immigrant students who were exposed to positive
adult role models were more successful in school than their counterparts.
In another study, Stanton-Salazar (2001) found that Mexican immigrant
parents seemed to place an emphasis on the importance of education, but their views
about the importance of college seemed to follow their childrens desires. In
general, the parents he interviewed lacked a strong personal involvement in the
process of schooling, believing instead that giving their children a strong moral base
would lead to success in school. The parents themselves spoke of how their own
limited education prevented them from assisting their children. Many of these
parents felt educators were better-equipped to handle that part of their childrens
lives, yet criticized Mexican parents in general for not being involved in the
schooling process. Stanton-Salazar (2001) found that a majority of the 55 students

he interviewed indicated that they would likely ask their parents for academic help if
they felt their parents were well-equipped to provide that help. On the other hand,
he discusses the effects of what he refers to as accumulated folk capital, which
describes the influence parents have on their childrens educational aspirations by
telling horror stories about the negative experiences theyve experienced, which they
attribute to their own personal lack of education.
Taking another perspective, some researchers study students across race,
class and gender categories to determine what contributes to success and failure.
Phillip Wexler (1992) studied schools that contain primarily working class students
and found that teachers and students have different perspectives. While many
teachers think students dont care to learn, many students think teachers dont care to
teach them. In addition, teachers believe the school system itself doesnt care about
the education of working class youth, making it difficult to create a fostering
environment in which to teach. The varying attitudes of teachers, students and
administrators in this environment created a cycle in which it became impossible to
learn or to teach (Wexler, 1992).
Lee and Bryk (1989) found that the characteristics of the schools students are
attending are often a more important determinant of academic performance than
characteristics of the families and of the students themselves. Mary Metz (1990)
found that a significant number of working class schools employed teachers with
working class backgrounds, decreasing the chance of students exposure to valuable

social resources they do not obtain in the home. She also found that in mixed-
income schools, tracking systems placed working class students in lower tracks, and
the lower track classes were assigned to younger, less experienced and less educated
teachers, reducing the chance that working class students would learn much of
anything at school.
As Grant and Murray (1999) suggest, teachers are generally primarily white,
female and middle class. In addition, the teachers education they receive often
prepares them to teach white middle or upper class children and does not generally
deal with issues of different backgrounds, cultures or class systems. Moll and
Gonzalez (1997) point out that culture is a dynamic and contextual facet, not
something that can be arbitrarily defined. Therefore, in situations in which teacher
education does address cultural differences, often the real world experience teachers
are exposed to after graduation is a shocking one.
Valerie Lee (1997) claims that restructuring schools to be smaller, more
closely knit communities would benefit students of all cultures as students and
teachers would naturally come to know one another better than in current larger
structures, and through this process, would come to understand differences in
cultures and backgrounds.

Differences In Class Values
Gerris, Dekovic and Janssens (1997) discuss the idea that lower class parents
place value on conforming to society and upper class parents tend to place value on
autonomy and self-direction. Because of this phenomenon, the authors suggest that
it is no surprise that lower class students would reproduce their own place in society
and upper class students would learn to be self-sufficient.
Reproducing Stereotypes
Other researchers (Schwalbe, Godwin, Holden, Schrock, Thompson,
Wolkomir, 2000) employ the negotiated-order perspective to explain the
reproduction of inequality, which suggests that the structures and social conditions
are reproduced by specific actors and their face to face interactions with one another
within institutions. This notion rejects the idea that institutions themselves are
responsible for reproducing inequalities and instead focuses on individuals within the
institutions. These researchers also suggest that a generic social process contributes
to the reproduction of inequality, meaning that there are generic social ways of doing
things such as building trust or forming communities. When certain actors, such as
immigrants, do not understand the acceptable way to do these things, then they
become outcast

Othering (Schwalbe & Mason-Schrock, 1996; Schwalbe, et al. 2000) is
another way of explaining the reproduction of inequality. Othering occurs when a
group of people develop categories of people and develop ideas about what makes
people part of those categories. Often, a dominant group creates categories with
negative connotations about a subordinate group, such as the fictitious notion that
all Mexicans are unintelligent. What Schwalbe, et. al. (2000) suggest is that
subordinate groups have different ways of dealing with their subordination under
othering that often lead to simultaneously occurring positive and negative
consequences. Some subordinates will accept their position, reproducing the
inequality, but at the same time drawing resources they may not otherwise have
access to from the dominant group. An example would be the minority model
student, who, by drawing attention to his/her minority status is drawing attention
towards the inequality, is at the same time enjoying affirmative action benefits.
Schwalble et al. (2000) and Schwable and Mason-Schrock (1996) claim that
another response to othering is simply dropping out, whether from school or from
actively pursuing a decent wage. When a large number of people from the same
subordinate group drop out, the very subordination that causes them to drop out in
the first place is reproduced. Often times, those who are dropping out see their doing
so as a rejection of the dominant groups norms. However, dropping out of the
structure is doing nothing to change these norms and to better the individuals
position in society.

Parental Education Effects
In a time-diary method of data collection, Bianchi and Robinson (1997)
found that children whose parents were college-educated spent significantly more
time reading or being read to. They also found that the higher a familys income
level was, the more time a child spent reading or being read to. On the reverse side,
children whose parents were not college-educated spent more time watching TV per
day than those children whose parents were. The results of this study, based on a
sample of 1200 that is not representative of race or socioeconomic status, suggest
that parents who have a college education are more likely to realize the importance
of assisting their children in their own educational pursuits (Bianchi and Robinson,
1997). The researchers found no significant difference in the amount of reading a
child experiences based on the differences between single and two parent homes nor
did they find differences based on employment status of mother. What is not taken
into account, is whether or not the parents who assist their children were themselves
assisted as children.
Others (Lewis, Ross and Mirowsky, 1999 and Gerris, Dekovic and Janssens
1997) suggest that better educated parents are more likely to encourage and reward
independence in their children in addition to providing assistance with developing
cognitive skills. These researchers also suggest that since education increases an
individuals sense of control over his or her own life, better educated parents may

pass down this sense of control to their children, giving them a higher chance of
success in society.
In in-depth qualitative studies of about 100 students from each class, race and
gender, LaReau (2001) showed that parents with more education attempted to use
reasoning skill with their children to discipline them, as opposed to lesser educated
parents, who often beat or grounded their children as a form of discipline. She also
found, as demonstrated by the above literature, that those with more education were
in higher socio-economic class segments than those without. In addition, she found
that more educated parents were spending more time encouraging their children to
read and do well in school than parents with lower levels of education.
Teachman (1987) found that parental use of educational resources in the
home had a positive affect on both men and womens educational successes,
regardless of family income level. The problem is, as others have shown (Stanton-
Salazar, 2001), low-income families are less likely to understand the importance of
providing such resources, or even when they do, they are unable to afford them.
Student Locus Of Control
While many researchers (Ross & Wu, 1995 & Lewis, Ross & Mirowsky,
1999) have explored the effects of education on locus of control, few (Fine, 1986;
Alexander, et. al 1997) have explored the idea that locus of control might in fact
effect education attainment. Alexander, et. al. (1997) found in their Baltimore

sample that students who eventually dropped out of school tended to have more of an
external locus of control, believing that their destinies are not up to them, but left to
fate. What the study does not explore is the relationship between parents attitudes
and childrens locus of control.

These findings suggest that many different things could be responsible for the
failure of at-risk students. One thing is obvious, however, and that is that at-risk
students are indeed dropping out of high school at a much higher rate than those who
are not at-risk. An at-risk student is a student who comes from a lower
socioeconomic status family, a student who faces the mixed challenge of being a
minority and coming from a mid to lower socioeconomic status family, a foster
home or a home in which the parents have little social capital. Social capital can be
obtained from secondary education, job skills, inheritance of connections to power
from family, and access to resources needed to secure a stable wage-earning job that
is open to upward mobility. The available literature suggests that a lack of social
capital creates a huge obstacle to students who are already at-risk. The questions I
address are these: Can a student who is at-risk of dropping out of high school benefit
from resources provided by a peer or peer group, a social activity, teacher, counselor,
or other adult role model enough to off-set the disadvantages that cause the student
to dropout of school? What role do student self concept and locus of control play in
influencing whether or not a student drops out of high school? Based on the

information in the NELS (National Education Longitudinal Study), I hypothesize
that an at-risk student who has extended contact with such a role model, peer group
or social activity in excess to classroom activity will have a better chance of success
than an at-risk student who does not have such contact.

The High School and Beyond Survey, the National Longitudinal Survey of
Youth and the National Education Longitudinal Survey of 1972 are all possibilities
for a project such as this. However, the National Education Longitudinal Survey,
1988 2nd Follow Up Data conducted by the National Center for Education Statistics
is the best fit for this project. The NELS study was chosen because it provides a
wide range of data about family structure, socioeconomic status, race, academic
ability, students self-perceptions and students interactions with those around them
as they specifically apply to education goals and pursuits. In addition, this data set
corresponds with my definition of dropout, and this is one of the few data sets in
which questionnaires and standardized tests were administered to students who had
dropped out. This data set contains all of the indicators needed to satisfy the
operational definitions for my variables.
The NELS sample for the second follow-up was designed to maintain the
representative sample used in the original study. The sample is a two-stage, stratified
sample design. The schools serve as the first-stage and the students as the second
stage. Each school selected 24 eighth grade students randomly with a possible extra

two Hispanic students for oversampling. If a school had less than 26 students, all
students were included in the sample. 1,734 schools were randomly selected from a
national frame of approximately 39,000. 1,052 of these schools participated and
provided usable data. The sample was stratified according to whether a school was
urban, suburban or rural and also according to racial composition, and according to
whether the school was public or private. The sample objectives were to include
about 21,500 of the same students that were in the 1988 eighth grade sample in order
to provide a longitudinal cohort. Students were traced from their original school of
enrollment if they transferred between the eighth and twelfth grades, and the sample
was refreshed for randomness.
In addition, the sample was a valid probability sample of all current tenth
grade enrolled students. The follow up sample was entirely student driven, in that
the most important component of the sample was including students who were a part
of the initial sample. The researchers found 99 percent of the original sample by
tracing. Response rates for student questionnaires, tests, dropout questionnaires, and
school administrator questionnaires were all above 90%. The response rate for the
dropout tests was approximately 48%, largely due to the difficulty in locating
students once theyve dropped out. However, the dropout test scores were weighted
in order to be used for analysis.

The researchers obtained approval for the study and questionnaire designs
from the Educational Information Advisory Council (ELAC), the Council of Chief
State School Officers, the National Catholic Education Association (NCEA), and the
National Association of Independent Schools (NAIS). In addition, endorsements
from the American Association of School Administrators (AASA), the National
Association of Secondary School Principals (NASSP), and the National School
Boards Association (NSBA). Permission was obtained from each Chief State School
Officer and from the District of Columbia. A state coordinator was appointed as
project liaison for each state in order to field questions from schools and districts
regarding the research. School superintendents and principals also gave approval
for the study to a NORC representative who contacted each of them. Ethical
considerations were taken into account in the approval process.
Prior to administering the questionnaires, researchers visited each school and
asked the principals to appoint a liaison to assist with interviewing activities such as
sample freshening, distribution and collection of questionnaires, and verification of
student enrollment. Principals were asked to schedule a day for completion of the
student questionnaires and a make-up day for anyone absent on the appointed day.

Data were collected in two phases for students, administrators, dropout
students, teachers and school administrators to ensure that those students identified
as dropouts were actually dropouts.
Two NORC field representatives and a team leader supervised the
completion of the student questionnaires and took attendance. Each questionnaire
was checked for completeness. After a ten minute break, a cognitive test battery was
administered covering math, reading, science and social studies to measure each
students cognitive abilities. After the test was completed, representatives attempted
to have students complete any unfinished questionnaires.
Off-campus questionnaire sessions were held for dropouts or students who
had participated in the original study but had moved to a school that was not part of
the sample. In addition, students who missed the original and make-up
questionnaires were asked to participate. The tests were held at public libraries or
community meeting rooms and procedures followed those of the on-campus testing
sessions as closely as possible. Dropout students who participated in the off-campus
testing were reimbursed up to $20.00 for travel expenses. In order to raise the
response rate of the dropout students, a final telephone questionnaire was
administered to dropouts who could not attend an off-campus test. Nonrespondents
from this group were screened and interviewed in person, using an abbreviated
version of the original questionnaire. The procedures for these interviews are not

described in the data set. Telephone and in-person questionnaires were not
supplemented with a cognitive test due to time and procedural constraints.
The school administrator questionnaires were self-administered. School
officials were allowed to designate another knowledgeable school official to
complete the bulk of the questionnaire if they desired. The last section of the
questionnaire contained items on school climate that were completed only by the
schools chief administrator. The school administrator questionnaire was
distributed approximately two weeks before the set day for in-school questionnaires
by the school coordinator. A cover letter was included which provided instructions
for completing the questionnaire and asked that the questionnaire be completed
before the in-school questionnaire day. Administrators were instructed to use the
1989-90 academic year and the students who were in the tenth grade at that time as a
point of reference for the questions.
One or two teachers from the follow up sample were asked to complete self-
administered teacher questionnaires. Teachers taught combinations of classes that
spanned representatively across math, English, science and social studies. Data
collection was completed in two phases. First, approximately two weeks prior to the
schools in-school questionnaire day, school coordinators distributed a teacher
packet with the questionnaire, cover letter, and a study brochure. Teachers were
instructed to complete the questionnaire and return it to the coordinator before the
schools in-school questionnaire day. Self-addressed stamped envelopes were

included for those teachers who were unable to turn in a finished questionnaire
within the given time constraints.
In data preparation and processing, the questionnaires were monitored for
completeness, edited and coded for analysis and the documents were prepared for
archival storage. Questionnaires were checked for errors, inconsistencies and
missing data. If the information in error could not be resolved by taking into account
other portions of the questionnaire, the proper coding for missing data was given to
the inconsistency. Questionnaires were takes to NORCs Lake Park facility for data
entry. Each questionnaire was 100 percent key verified including all skip patterns
and zero-filling of numeric fields. Final data from the first batch of keyed
questionnaires was checked against the original document to ensure accuracy.
For my analysis, I applied a filter to the sample of dropout and in school students in
the second follow-up data to include only those students who fit my definition of at-
risk. Students will be classified as at-risk or not at-risk according to a scale that
measures family socioeconomic status and race. Socioeconomic status is measured
according to fathers education level, mothers education level, fathers occupation,
mothers occupation and family income. The selection criterion to create the filter is
as follows: If (family socioeconomic status is in the bottom quartile) or (family
socioeconomic status is in the 2nd quartile and race is black) or (family

socioeconomic status is in the 2nd quartile and race is Hispanic) or (family
socioeconomic status is in the 2nd quartile and race is American Indian/Alaskan) then
student is at-risk. Asian is not included as an at-risk race because a preliminary
analysis of the test scores showed that Asian students are more likely to excel in
academics than other races. In addition, because of the literature is inconclusive as
to whether or not single families cause students to be at-risk (see Chapter 3), famiy
status was not included for the filter in this project. The sample size is 4,395. 2,125
are male and 2,270 are female. There are 242 Asians, 1,248 Hispanics, 800 Blacks,
1,996 whites and 84 American Indian/Alaskans in the sample. 73 of the students in
the sample attend private schools and 4,056 were from public schools. This disparity
in numbers is expected because the low socioeconomic status of all the students in
the sample would suggest that most families could not afford private schooling
For this project, two statistical models are used. I use an ANOVA model
because ANOVA is a powerful statistical procedure that shows how variables differ
in effects of two different large groups, in this case, at-risk dropout students and at-
risk non-dropout students. ANOVAs show how the mean scores for each evenly
distributed variable differs in a clear and simple manner.
In the ANOVA model, I measure the mean differences between dropout and
non-dropout students with regards to peer influence, parental influence, including

parents education and their involvement with their childrens education, family
background effects, students self concept and locus of control, academic measures
(GPA and standardized tests), social capital, and the effects of teacher influence.
The ANOVA model measures all of these effects as separate variables.
A logistic regression model is used in order to calculate the strength of the
linear relationship between two or more independent variables that have an effect on
the dependent variable. In other words, some of the independent variables in the
ANOVA model may affect other independent variables, but the ANOVA
calculations are unable to take that into account. Logistic regression shows these
effects. In addition, because my independent variable is dichotomous (dropout or
non dropout), logistic regression is the appropriate form of regression analysis to use.
This model measures the influence of social capital as two different factors,
human interaction and social involvement, the effects of students educational
aspirations, the effects of peer influences, parental influences, respondents sex, type
of school and academic measures (standardized tests) on dropout status.
Measurement of Variables/Concepts
The dependent variable in the model is dropout status. Dropout status is
measured through school records indicating whether or not a student has dropped out
in the last four years. Sample members who were no longer enrolled in a diploma-

granting high school program in the spring of 1992 and had not earned an
equivalency certificate were classified as dropout. Sample members who were not
enrolled in a diploma-granting institution in the spring of 1992 but who had
completed an equivalency program were not classified as dropouts.
Respondents educational aspirations are measured with four items.
Respondents were asked what they thought the chances of graduating high school
and going to college were. The scale for these items ranged from 1 being very low
to 5 being very high. Respondents were also asked how far they eventually expected
to take their education with a scale ranging from 1 being less than high school
graduation to 10 being PhD or Masters Degree. In addition* respondents
determined how important it is in life to get a good education. This item is coded in
a scale that ranges from 1 being not important to 3 being very important.
Parents aspirations for their childrens education was measured by asking
how far they would like to see their children take their education and was coded with
a scale ranging from 1 (less than high school graduation) to 10 (PhD or Masters
Degree). To determine the parents involvement with their childrens education,
respondents were asked who decides whether or not the respondent should go to
college or technical school. The item is coded as 1 = parents decide by themselves, 2
= parents decide after discussing with me, 3 = we decide together, 4 = 1 decide after
discussing with my parents and 5 = 1 decide myself. Another item measures the
amount of time respondents spend talking or doing things with their parents. This

item is coded as 1 = never/rarely, 2 = less than once per week, 3 = one or two times
per week, 4 = every day or almost every day. Parents education is measured with an
item asking the parental respondent how far he or she got in school and another item
asking how far his or her spouse got in school (1 = less than high school graduation,
10 = PhD or Masters Degree)
Respondents took standardized tests measuring their overall proficiency in
math, reading and science. Item Response Theory Scoring (IRT) variables are
employed in this project because they allow the comparison of students who have
taken different levels of achievement tests. A core of items shared among the
different test forms made it possible to establish a common scale that all students
could be rated on regardless of their track in school. IRT uses the pattern of right,
wrong and omitted responses to the items actually administered in a test form, and
the difficulty, discriminating ability and guess-ability of each item, to place each
student on a continuous ability scale. It is then possible to estimate the score the
student would have achieved for any arbitrary subset of test items calibrated on this
In addition, by using the overall pattern of right and wrong responses to
estimate ability, IRT can compensate for the possibility of a low-ability student
guessing several hard items correctly. If answers on several easy items are wrong,
a correct difficult item is assumed to have been guessed. Omitted items are also

less likely to cause distortion of scores, as long as enough items have been answered
right and wrong to establish a clear pattern.
The IRT scale was calibrated using PARSCALE software. The test
responses of the longitudinal sample members were used for calibration. Item
parameters were computed for test items that had appeared in any of the test forms at
any time. Holding these parameters fixed, Bayesian estimates of placement on the
continuous ability sale were obtained for all test takers at all points in time. The
procedure used takes into account group memberships (year and test form) to
minimize floor and ceiling effects. The ability estimates were used in conjunction
with the item parameters to compute IRT scores in the database.
For this model, students were measured by the overall proficiency levels in
reading, math and science. These scores were assigned only for students who had
complete and consistent response patterns for the item clusters within each subject
area. The presence of reversal patterns, or of too many critical items omitted,
resulted in second follow-up proficiency scores not being assigned for about 4
percent of the students who took the reading test, 11 percent of mathematics takers
and 10 percent of those with science test scores. The proficiency scores provide a
means of distinguishing total scores and score gains as measured by overall IRT-
Estimated Number Right scores. At several points along the score scale of the
reading, mathematics and science tests, four-item clusters of the test questions
having similar content and difficulty were identified. A student was assumed to have

mastered a particular level of proficiency if at least three of the four items in the
cluster were answered correctly, and to have failed at this level if two or more items
were wrong. Clusters of items provide a more reliable test of proficiency than do
single items because of the possibility of guessing in a multiple choice test: it is very
unlikely that a student who has not mastered a particular skill would be able to guess
enough answers correctly in a four item cluster. (For some of the students who had
not answered critical items, an IRT-based procedure was undertaken to resolve
proficiency score assignments.) The proficiency levels were assumed to follow a
Guttman model, that is, a student passing a particular skill level was expected to
have mastered all lower levels; a failure should have indicated non-niaster at higher
Respondents locus of control was measured in a composite variable with
three items. The three items are In my life, good luck is more important than hard
work for success, Every time I try to get ahead, something or somebody stops me,
and Planning only makes a person unhappy, since plans hardly ever work out
anyway. Each of the above items was standardized separately to a mean of zero
and a standard deviation of 1 and recoded into four categories: 1 = low and 4 = high.
Respondents self-concept is measured with four items hsking the student to rate how
true each item is. The items are I feel good about myself, I feel I am a person of
worth, the equal of other people, I feel I am able to do things as well as most other
people, and On the whole, I am satisfied with myself. These items were

combined into a composite variable in the same manner as the locus of control
variable with a scale ranging from 1 (low) to 4 (high).
Respondents were asked how important they feel strong friendships are in
life. This item was coded with a scale where 1 is not important, 2 is somewhat
important and 3 is very important. The model also tests a number of positive and
negative peer influences. Positive peer influences include the importance of
spending time with friends, the importance of finishing high school, the importance
of studying, the importance of getting an education beyond high school, the
importance of attending class regularly and the importance of continuing education
beyond high school. Negative peer influences include the importance of having sex,
the importance of partying, the importance of drinking and the importance of doing
drugs. These items were framed as a multi-part question that asked among close
friends, how important is to..Each item is coded on a scale with 1 being not at
all, 2 being somewhat important and 3 being very important. Respondents were also
asked how many of their friends dropped out of school, how many friends had no
plans to go to college, how many friends had plans to go to a two-year college and
how many friends had plans to go to a four-year college. These items are coded with
1 being none, 2 being a few, 3 being most, 4 being all. Respondents were asked how
much time they spend talking to or doing things with friends. This item is coded as 1
= never/rarely, 2 = less than once per week, 3 = one or two times per week, 4 = every
day or almost every day.

In addition, respondents were asked questions about recent disruptions
causing stress in their family life. A multi-part item asks have any of the following
happened to your family in the last two years? The items were all answered yes or
no. The events included, family moved to a new home, parents divorced, parents
remarried, parent lost job, parent started work, parent got better job, you became ill,
parent died, relative died, sister got pregnant, brother dropped out of school, family
was on welfare, family went off welfare, family member got ill, or family member
did drugs.
In this model, social capital is measured as social capital gained from
participation in social activities and social capital as provided by interactions with
people. A multi-part item asked respondents how often do you spend time on the
following activities outside of school? The responses were coded from 1 being
never/rarely to 4 being every day. The activities include using personal computers,
doing volunteer or community service, talking or doing things with adults other than
ones parents, taking classes, and taking sports lessons.
Three items measure respondents perceptions of school environment.
Respondents were asked if they agree that disruptions were a big problem at school,
if they agree that students make friends with other racial groups at school, and how
strongly they agree fights occur between racial groups at school (1 = strongly agree,
4 = strongly disagree).

Teacher influence on respondents is measured by the students and dropouts
perceptions of the teachers. Two items were used asking respondents how strongly
they agree that teaching was good in the school and that teachers at school were
interested in the students. These items were coded on a scale with 1 being strongly
agree and 4 being strongly disagree.
The type of school respondent attends or attended is recoded into a
dichotomous variable where 0 = public school and 1 = private. In all research
projects of this type, it is standard practice to control for public school attendance
(NCES, 1988).

ANOVA Results*
Table 7.1.1
ANOVA for Family Structure,
Peer Influence, Social Capital
and Self Concept
Concept Variables Student Mean Dropout Mean Mean Diff. (P)
Student perception of Disruptions at school a problem 2.61 2.52 .002.
school environment Racial fighting a Problem 2.91 2.71 .000
Friendly with other races 1.98 2.17 .000
Academic measures Math test scores 1.88 1.11 .000
Reading test scores 1.37 1.05 .000
Science test scores 1.07 .74 .000
GPA 14.5383 5.787 .000
Parental Influence Parents Education 1.86 1.65 .000
Parents involved in college plan 3.99 4.07 .000
Educational Aspirations Plans to attend a 2-year school 2.66 2.33 .000
Plans to attend a 4-year school 2.93 2.08 .000
Likelihood for college 3.79 2.39 ;000
Likelihood of graduating HS 4.68 2.75 .000
Negative Peer Influences Friend dropped out of HS 1.77 2.57 .000
Friends not going to college 2.43 2.83 .000
Importance of drinking 1.48 1.46 .568
Importance of having sex 1.82 1.91 .001
Importance of doing drugs 1.17 1.21 .047
Positive Peer Influence Importance of going to class 2.43 2.19 .000
Importance of good education 2.42 2.15 .000
Importance of studying 2.24 2.17 .002
Importance of good grades 2.427 2.314 .000
Importance of finishing HS 2.77 2.45 .000

Table 7.1.1 Cont.
Importance of Peers Importance of being with friends Importance of good friends Time spent with friends 2.48 2.68 3.22 2.40 2.62 3.07 .008 .006 .000
Social capital Taking classes outside of school 1.20 1.14 .006
Community service 1.30 1.19 .000
Computer use outside school 1.57 1.38 .000
Playing sports 1.16 1.10 .001
Adult interaction (not parents) 2.42 2.72 .000
Family stressor events Family member did drugs 1.86 1.84 .084
Family member got ill 1.80 1.79 .416
Family member moved 1.79 1.72 .000
Family went off welfare 1.94 1.94 .692
Family went on welfare 1.87 1.80 .000
Respondent became ill 1.93 L93 .807
Family member got better job 1.84 1.81 .013
Parent died 1.96 1.94 .002
Parents got divorced 1.92 1.89 .002
Parent lost job 1.84 1.81 .077
Parents remarried 1.93 1.89 .000
Parent started work 1.82 1.81 .471
Relative died 1.64 1.65 .472
Sister got pregnant 1.90 1.83 .000
Psychological variables Student self perception 2.43 2.03 .000
Student locus of control 2.37 2.03 .000
Student perception of Teaching is good 2.14 2.03 .000
teachers Teachers care about students 2.25 2.07 .000
Unless otherwise noted in the text, all variables are significant at the .000 level.
Whether or not class disruptions are seen as a problem is significant at the
.002 level, and the perception that fighting between racial groups and that students
were friendly with other races are significant at the .000 level. The dropout mean for
the perception that disruptions were a problem in class was 2.52, and the student
mean was 2.61, showing that students who did not dropout perceived disruptions in
class to be a bigger problem. The student mean for the perception that fights

between racial groups is a problem is 2.91 and the dropout mean is 2.71. The student
mean for the perception that racial groups are more friendly is 1.98 and the dropout
mean is 2.17. This suggests that dropouts perceived racial relations to be less
volatile at school.
On average, students who did not drop out have a higher self concept than
students who do (2.43 to 2.03). Locus of control was also higher for students who
did not drop out (2.37 to 2.03).
Parents education level is also a significant predictor of dropout status.
Dropouts parents mean education is 1.65, while students parents mean education
is 1.86. Test scores and GPA predicted dropout status, math score means Were 1.11
for dropouts and 1.88 for students, reading scores averaged 1.05 for dropouts and
1.37 for students and science scores averaged .74 for dropouts and 1.07 for students.
Mean GPA for dropout students was 5.787 and 14.5358 for students.
School aspirations for dropouts and students also differ significantly. For
aspirations to attend a two-year college, the dropout mean is 2.33 and the student
mean is 2.66. For aspirations to attend a four-year college, the dropout mean is 2.33
and the student mean is 2.66. For students self reported likelihood of going to
college, the mean score for dropouts was 2.39 and for students 3.79. For the
probability of graduating high school, the dropout mean was 2.75 and students mean
was 4.68, suggesting that students who believe they will graduate from high school

Students who did not drop out also had far fewer friends who had dropped
out of high school. The mean for dropout students is 2.57, and the mean for students
is 1.77. Dropouts were more likely to have friends who did not expect to go to
college (2.83 to 2.43). Students were much more likely to have friends who placed a
high level of importance on going to class regularly (2.43 to 2.19), getting an
education beyond high school (2.15 to 2-42), studying (2.17 to 2.24), getting good
grades (2.427 to 2.314) and finishing high school (2.77 to 2.45). The importance of
drinking is not significant to this model (p = .568) The importance of having sex and
of doing drugs were also significant, and dropout students tended to do more of both.
The importance of getting together with friends and the importance of having good
friendships in life are significant (.006), but the means for both groups did not differ
dramatically. For dropouts, the mean score for the importance of having good
friends in life is 2.62 compared to students mean score of 2.68 The mean of the
importance of getting together with friends among the group of friends was 2.40 for
dropouts and 2.48 for students. However, while both groups believed that having
good friends and getting together with friends is important, dropouts reported
spending less time on average with friends than students (3.07 to 3.22).
Social capital indicators are all significant to the model. On average, students
who did not drop out took more classes outside of school than dropouts (1.20 to
1.14). Students who used computers outside of school (1.57 to 1.38), who engaged
in community service (1.30 to 1.19) and who played sports (1.16 to 1.10) more often

were also less likely to dropout. However, dropout students reported that they spent
more time talking to or doing things with other adults more often than non-dropouts
(2.72 to 2.42).
The respondents perception of teaching was significant, but the mean score
for dropouts was higher for good teaching (2.14 to 2.03) and for the perception that
teachers care about students (2.25 to 2.07).
The amount of time spent doing things with or talking to parents is not
significant (p = .696), however, whether or not parents helped decide if the student or
dropout should continue education beyond high school is significant. The mean
score for students is 3,99, leaning towards the parents and I decide together and the
mean score for dropouts is 4.35, which is closer to I decide myself, suggesting that
parents are more involved with academic decisions for students who graduate than
for dropouts.
Dropout students were somewhat more likely to experience stressful family
events within the past two years than non-dropout students, however, on average, the
mean scores for these events were not that much different for dropouts than for

Logistic Regression Results
Table 7.2.1
Logistic Regression for Family Structure,
Peer Influence, Social Capital
and Self Concept*
Concept Variables Regression Coefficient Standard Error Wald d f Sig. (p)
Academic Measures Standardized test scores -.113 .049 5.274 1 .022
Social Capital activities Social Capital (includes taking classes, community service, sports and participation in after school program) -.125 .065 3.715 1 .054
Social Capital human interaction Social Human Capital (includes spending time with parents, spending time with other adults, spending time with friends) .070 .044 2.551 1 .110
Education Aspiration Likelihood of graduating -1.446 .097 221.455 1 .000
Aspiration for 4-year college -.486 .098 24.639 1 .000
Importance of continuing education -.144 .169 .726 1 .000
Negative Peer Inf. Friends dropped out .549 .115 22.579 1 .000
Positive Peer Inf. Importance of going to class .204 .180 1.293 1 .256
Self Concept Student self concept .318 .096 10.892 1 .001
Respondents sex Respondents sex .292 .205 2.037 1 .154
Parental factors Parents education -.019 .124 .023 1 .880
Parents recently divorces -.434 .296 2.149 1 .143
Public school Student attended public school -.692 .733 .890 1 .345
* This model has a -2 Log likelihood is 745.312, and the Cox & Snell R Square is
.351. This index shows that the model has a good overall fit.

Standardized test scores are significant at the .002 level. As standardized test
scores go down, the likelihood of dropping out increases, suggesting that students
with lower academic ability are more likely to dropout than students with higher
academic ability.
Social capital in the form of activities is very close to being significant at the
.05 alpha level (.054). The more a student is involved in activities outside of school,
the lower are the chances that the student will dropout. Students who take extra
classes, participate in community service, sports or other after school programs will
be less likely to drop out of school. However, contrary to the literature, social
capital in the form of human interaction is not significant to this model. Spending
more time with parents, spending time with other adults, and spending time with
friends do not affect whether or not an at-risk student will drop out of school. It
should be noted that one of the variables used in this factor reads, On average, how
much time do you spend talking to or doing things with other adults (not your
parents)? This indicator could be problematic in that dropout students may spend
more time doing negative activities with other adults, such as buying drugs. In
addition, delinquent students may be spending time with parole officers or therapists
which would also qualify under this question. For future analysis, this indicator will
not be included. In addition, for further data collection, this question should be re-
worded to reflect the specific kinds of interaction between students and adults to
reduce any future causal errors in analysis.

Respondents educational aspirations are significant. As the respondents
self reported likelihood that they would graduate from high school decreases, the
chance that they will drop out increases. Along the same lines, as a respondents self
reported likelihood that he/she will go to a four-year college decreases, the chance of
dropping out increases. Students who drop out are also less likely to strongly
believe that continuing education is important.
Negative influence from peers is significant to dropping out of high school.
Students with more friends who have dropped out of school are more likely to drop
out themselves. On the other hand, positive peer influence is not significant to this
Respondents self concept is significant at .001. The higher a respondent
thinks of themselves, the more likely they are to graduate high school, suggesting
that strong self image has a positive effect on education.
Respondents sex, parents education, and attending a public school serve as
control variables in this model.

Given the importance that has been attached to the parental component of
social capital, both the ANOVA results and the logistic regression results suggest
that considerably more attention should be placed on social capital within the larger
social structure. Because involvement in social activities significantly reduced the
chances of dropping out, schools that provide more extra curricular activities and that
encourage students to engage in them would potentially have fewer drop out
students. As Coleman (1988) suggests, financial capital (having books, computers,
etc. in the home) is an important determinant of students success in school. Students
that have access to financial capital are better able to participate in activities outside
of school, and as Coleman (1988) suggests, these group activities provide a closed
social structure in which the group develops norms and socially advantageous
connections tend to occur.
In general, the literature linking locus of control and self concept to education
(Lewis, Ross and Mirowsky, 1999, Ross & Wu, 1995) suggests that dropping out of
high school causes a low internal locus of control. However, the results of the
ANOVA and logistic regression suggest that a low internal locus of control and low
self concept actually cause students to drop out of high school. Given this

information, it is important for school policy makers to look into better ways to
support students self esteem. Developing support groups, monitoring student
behaviors, giving more achievement awards and encouraging teachers to use positive
feedback on a regular basis would all help to increase student confidence, self image
and locus of control, which in turn prevents dropping out.
In the ANOVA analysis, parents education is an important predictor of
dropout status. This finding is consistent with Coleman (1988) and Goyette and
Conchas (2002). However, this is a condition that is difficult to change. Instead,
policy makers must look to programs that will be successful in encouraging students
who have undereducated parents to see the value in their own education (Stanton-
Salazar, 2001). The ANOVA results also show that whether or not parents are
involved in the students decision to continue education beyond high school is a
significant factor determining drop out status. Given that parents who are not
educated rarely encourage their children to continue education (Sewell & Shah,
1968), schools should work harder to encourage students to pursue education beyond
high school.
Both the ANOVA results and the logistic regression results suggest a very
strong negative peer influence on drop out status. Dropouts tend to have friends that
engage in negative behaviors such as doing drugs or dropping out of school, and
according to the ANOVA results, non-dropouts tend to have friends that place a high
level of importance on education. Given that the ANOVA results show a high level

of significance on the importance of having good friends and spending time with
friends, it is important that students have friends who will be a good influence.
Providing after school programs that encourage students to take an interest in
positive activities will allow students to be introduced to potential new friends that
may have a positive influence on them, preventing them from taking on attitudes that
may lead them to drop out.
In addition, both ANOVA and logistic regression results suggest that students
who have higher educational aspirations are less likely to drop out of school. Again,
policy makers need to provide encouragement and motivation for students to set their
educational goals higher If students do not understand the potential life
consequences of being uneducated, students educational goals will not be as high.
Clearly, students who do not have high educational aspirations place less importance
on finishing high school.
The ANOVA suggests that dropout students perceive that teaching is good
and that teachers care about the students more so than non dropouts. While this is
not what my hypothesis expected, there are several explanations. First of all, it is
possible that the teaching did not have an effect on dropout status. It is also possible
that non-dropouts did not report that they liked their teachers because of peer
My hypotheses ask whether a student who is at-risk of dropping out of high
school can benefit from resources provided by a peer or peer group, a social activity,

teacher, counselor, or other adult role model enough to off-set the disadvantages that
cause the student to dropout of school and what role do student self concept and
locus of control play in influencing whether or not a student drops out of high
school? Because teacher and other adult roles were not significant in the logistic
regression model, I will not argue that teachers or other adult role models have a
strong influence on dropout status. However, according to my results, peer influence
has a tremendous effect on dropout status, and therefore, a peer or peer group can
provide benefits to an at-risk student arid possibly prevent them from dropping out of
high school. In addition, since at-risk students who Were involved in activities
outside of school were less likely to drop out, social activities are also a source of
benefit for students at risk.

This research project is one of the first to discuss the effects of locus of
control and self concept on dropping out of high school. In the past, research has
suggested that dropping out of high school causes a negative self concept and low
internal locus of control, when in fact, it may be the other way around. The
implications of these results are very important for psychologists and sociologists
alike. If a person has maintained a negative self concept and a low internal locus of
control since childhood, psychological treatment would need to be different than
what is used to treat a person who has developed a negative self concept and low
internal locus of control as a result of dropping out of high school or other specific
life events.
This is also the first research project of its kind to apply a filter that studies
at-risk students only and determines factors that prevent them from dropping out of
high school. Previous studies focus only on what causes students to drop out of high
school. Because all of the respondents in this sample are classified as at-risk
according to standard definitions in the literature, the model provides a comparison
that shows what factors prevent some at-risk students from graduating while others

see success. This comparison is key to understanding how policies can help keep
students in school.
Because the sample is nationally representative, the results of this project
may be generalized to all at-risk high school students in the United States. In
addition, the NELS data set provides indicators that are not available in other data
sets that allow this study to examine the effects of social capital, peer influence,
parental influence, and self concept in the same model, which has not previously
been done.
The use of both ANOVAs and logistic regression for analysis allow for a
close examination of all the factors that are involved in influencing students to drop
out or to remain in school. The ANOVA shows group differences on individual
variables and the logistic regression shows the causal assumptions.
I was not able to find a more recent data set, which would give more strength
to the policy applications of this research, as educational systems and youth culture
have changed so much over the last ten years. However, the process of collecting
and preparing data is so extensive that it would be very difficult to obtain a solid data
set that was less than a few years old.
The effects of the variables in this project could be better understood with the
use of factor analysis or with a structural equation model. Further research on how
negative self concept and internal locus of control should focus on combining the
variables in such a model to determine their effects on one another, as well as the

effects on school retention or drop out status. In addition, further research focusing
on the longitudinal data would reveal patterns of family structure, peer influence,
locus of control and how they pave the path to graduation or dropping out.
Another interesting study would be comparing at-risk dropout and non-
dropout students to non-at-risk dropout and non dropout students to determine if the
significant predictors are the same or similar for both groups. In addition, although
this data set does not measure it, examining the influence of parenting styles on
students academic attainment would explain a lot more about the influence parents
have on their childrens education.
In addition, running separate models for males and females would determine
whether or not there are significantly different predictors of drop out status for males
and females. Given that the literature suggests that males and females experience
adolescence differently, the results of such a study would be an interesting addition
to education literature.
Finally, evaluation research would useful to determine whether or not some
of the suggested programs would work to provide a positive influence on education

Alexander, Karl L, Entwisle, Doris R. & Horsey, Carrie. 1997. From First
Grade Forward: Early Foundations of High School Dropout. Sociology
of Education, Vol. 70, No. 2:87-107.
Barker, Roger G. & Gump, Paul V. 1964. Big School-Small School: High
School Size and Student Behavior. Palo Alto, CA: Stanford University Press.
Berliner, David C. & Biddle, Bruce J. 1995. The Manufactured Crisis: Myths,
Fraud, and the Attack on Americas Public Schools. New York:
Addison- Wesley-Longman.
Bianchi, Suzanne M. & Robinson, John. 1997. What Did You Do Today?
Childrens Use of Time, Family Composition, and the Acquisition of
Social Capital. Journal of Marriage and the Family, Vol. 59: 332-344.
Biddle, Bruce J. 2001. Ethnicity and Achievement in American Schools,
Social Class, Poverty and Education Policy and Practice. New York;
London: RoutlegeFalmer.
Boulter, Lyn. 2002. Self-Concept as a Predictor of College Freshman
Academic Adjustment. College Student Journal, Vol. 36, No. 2:234-46.
Bourdieu, Pierre & Passerson, Jean Claude. 1977. Reproduction in Education,
Society and Culture. London; Beverly Hills, CA: Sage Publications.
Brooks-Gunn, Jeanne, Duncan, Greg J., Kato Klebanov, Pamela, Sealand,
Naomi. 1993. Do Neighborhoods Influence Child and Adolescent
Development? American Journal of Sociology, Vol. 99, No. 2:353-95.
Brooks-Gunn, Jeanne & Duncan, Greg J. 1997. The Effects of Poverty on
Children. The Future of Children, 7(2): 55-71.

Buchel, Felix & Duncan, Greg J. 1998. Do Parents Social Activities Promote
Childrens School Attainments? Evidence from the German
Socioeconomic Panel. Journal of Marriage and the Family, Vol. 60,
Issue 1 (Feb. 1998),:95-108.
Catterall, J. S. 1998. Risk and Resilience in Student Transition to High
Schools, American Journal of Education, Vol. 106:302-333.
Coleman, James S. 1988. Social Capital in the Creation of Human Capital.
American Journal of Sociology. Vol. 94, Issue Supplement:
Organizations and Institutions: Sociological and Economic Approaches
to the Analysis of Social Structure, S95-S120.
Coleman, James S. 1990. Commentary: Social Institutions and Social Theory.
American Sociological RevieWi Vol. 55, No. 3:333-339.
Coleman, James S. 1992. The Rational Reconstruction of Society. American
Sociological Review, 58:1-15.
Coleman, James S. & Hoffer, Thomas P. 1987. Public and Private Schools. New
York: Basic Books.
Croninger, Robert G. & Lee, Valerie, E. 2001. Social Capital and Dropping
Out of High School: Benefits to At-Risk Students Of Teachers Support
and Guidance. Teachers College Record, Vol. 103, No. 4: 548-581.
Department of Education National Center for Education Statistics. Office of
Educational Research and Improvement. 2000. Dropout Rates in the
United States.
Dika, Sandra L. & Singh, Kusum. (2002). Applications of Social Capital in
Educational Literature: A Critical Synthesis. Review of Educational
Research, Volume 72, No. 1: 31-60.
Dreyfoos, Joy. 1988 Making It Through Adolescence In a Risky Society: What
Parents, Schools and Communities Can Do. New York:Oxford
University Press, Inc.
DuBois, David L. 2001. Family Disadvantage, the Self, and Academic
Achievement, Social Class, Poverty and Education Policy and
Practice. New York; London: RoutlegeFalmer.

Felson, Richard B., Liska, Allen E., South, Scott J. & McNulty, Thomas. 1994.
The Subculture of Violence and Delinquency: Individual vs. School
Context. Social Forces, Vol. 73, No. 1:155-173
Fine, Michelle. 1986. Why Urban Adolescents Drop Into and Out of Public
High School. Teachers College Record, 87:393-409.
Gerris, Jan R.M., Dekovic, Maja & Janssens, Jan M.A.M. 1997. The
Relationships between Social Class and Childrearing Behaviors:
Parents Perspective Taking and Value Orientations. Journal of
Marriage and the Family, Vol. 59:834-47.
Goyette, Kimberly & Xie, Yu. 1999. Educational Expectations of Asian
American Youths: Determinants and Ethnic Differences. Sociology of
Education, 72(l):22-36.
Goyette, Kimberly & Conchas, Gilberto Q.. 2002. Family and Non-Family
Roots of Social Capital among Vietnamese and Mexican American
Children. In Bruce Fuller and Emily Hannum (Eds.) Schooling and
Social Capital in Divers Cultures: Research in Sociology of Education,
Vol. 13. Oxford: Elsevier Science.
Granovetter, Mark 1973. The Strength of Weak Ties. American Journal of
Sociology, 78: 1360-80.
Grant, Gerald & Murray, Christine E. 1999. Teaching in America: The Slow
Revolution. Cambridge, MA: Harvard University Press.
Hall, Peter M. 2001. Social Class, Poverty and Schooling. In Biddle, Bruce J.
(Ed.), Social Class, Poverty and Education. Policy and Practice. New
York, London: RoutelegeFalmer.
Hoffer, Thomas B. 1992. Middle School Ability Grouping and Student
Achievement in Science and Mathematics. Educational Evaluation and
Policy Analysis, Vol. 14, No. 3:205-227.
Hyjer Dyk, Patricia & Wilson, Stephan M. 1999. Family-Based Social Capital
Considerations as Predictors of Attainments Among Appalachian
Youth. Sociological Inquiry, Vol. 69, No. 3:477-503.

Jenkins, Patricia H. 1995. School Delinquency and School Commitment.
Sociology of Education,. Vol. 68, No.3:221-229.
Kalmijn, Matthijs & Kraaykamp, Gerbert. 1996. Race, Cultural Capital, and
Schooling: An Analysis of Trends in the United States. Sociology of
Education, Vol. 69, Issue 1:22-34.
Kerckhoff, Alan & Bell, Lorraine. 1998. Early Adult Outcomes of Students at
Risk. Social Psychology of Education, 2: 81-102.
LaReau, Annette. 2001. Linking Bourdieus Concept of Social Capital to the
Broader Field. In Biddle, Bruce J. (Ed.), Social Class, Poverty and
Education Policy and Practice. New York, London: RoutelegeFalmer.
LeCompte, M.D., & Dworkin, A.G. 1991. Giving Up On School. Student
Dropouts and Teacher Burnouts. Newbury Park, CA: Corwin Press.
Lee, Valerie. 1997. Gender, Equity, and the Organization of Schools. In BJ.
Bank & P.M. Hall (Eds.), Gender, Equity, and Schooling: Policy and
Practice (135-158). New York: Garland Publishing.
Lee, Valerie E. & Bryk, Anthony S. 1989. A Multilevel Model of the Social
Distribution of High School Achievement. Sociology of Education,
Lewis, Susan K, Ross, Catherine E. & Mirowsky, John. 1999. Establishing a
Sense of Personal Control in the Transition to Adulthood. Social Forces, Vol.
Loveless, Tom 1997. The Uses and Misuses of Research in Educational
Reform. Paper presented at The State of Student Performance in
American Schools, a conference at the Brookings Institution, May 29-30,
1997, Washington, D.C.
Luthar S.S. & Ziglar, E. 1991. Vulnerability and Competence: A Review of
Research on Resilience in Children. American Journal of
Orthopsychiatry, Vol. 61:6-22.
McNeal, Ralph B. 1999. Parental Involvement as Social Capital: Differential
Effectiveness on Science Achievement, Truancy, and Dropping Out.
Social Forces, Vol. 78, No. 1:117-144.

Mehan, Hugh, Villanueva, Irene, Hubbard, Lea, & Lintz, Angela. 1996.
Constructing School Success: The Consequences of Untracking Low
Achieving Students. Cambridge: Cambridge University Press.
Metz, Mary Haywood. 1990. How Social Class Differences Shape Teachers
Work. In M.W McLaughlin, J. Talbert, & N. Bascia (Eds.), The
Contexts of Teaching in Secondary Schools: Teachers Realities. (40-
107). New York: Teachers College Press.
Moll, Luis & Gonzales, Norma. 1997. Teachers as Social Scientists: Learning
About Culture From Household Research. In P.M. Hall (Ed.), Race,
Ethnicity, and Multiculturalism: Policy and Practice (89-114). New
York: Garland Publishing.
Oakes, Jeannie. 1995. Two Cities Tracking and Within School Segregation.
Teachers College Record (summer 1995), Vol. 96, No. 4: 681-690.
Oakes, Jeannie & Wells, Amy Stuart. 1998. Detracking For High Student
Achievement. Educational Leadership, Vol. 55, No. 6:38-41.
Oakes, Jeannie & Weiner, Kevin. 1996. (Li)ability Grouping: The New
Susceptibility of School Tracking Systems to Legal Challenges.
Harvard Educational Review, Vol. 66, No. 3:451-470.
Parcel, Toby L & Dufur, Mikaela J. 2001. Capital at Home and at School:
Effects on Child Social Adjustment. Journal of Marriage and the
Family, Vol. 63: 32-47
Ross, Catherine & Van Willigen, Marieke. 1997. Education and the Subjective
Quality of Life. Journal of Health and Social Behavior, Vol. 38, No.
Ross, Catherine & Wu, Chia-Ling. 1995. Education, Age, and the Cumulative
Advantage in Health. Journal of Health and Social Behavior, Vol. 37,
No. 1:104-120.
Sandefur, Rebecca L. & Laumann, Edward O. 1998. A Paradigm for Social
Capital. International Political Science Review, Vol 10(4): 481-501.

Schwalbe, Michael, Godwin, Sandra, Holden, Daphne, Schrock, Douglas,
Thompson, Shealy & Wolkomir, Michele 2000. Generic Processes in
the Reproduction of Inequality: An Interactionist Analysis. Social
Forces, Vol. 79(2): 419-52.
Schwalbe, Micheal & Mason-Schrock, Douglas. 1996. Identity Work as Group
Process. Advances in Group Processes, Vol. 13:13-47.
Sewell, William H. & Shah, Vimal P. 1968. Parents Education and Childrens
Educational Aspirations and Achievements. American Sociological
Review, Volume 33, Issue 2:191-209.
Stanton-Salazar, Ricardo D. 2000. The Development of Coping Strategies
Among Urban Youth: A Focus on Help-Seeking Orientation and
Network-Related Behavior. In M. Montero-Sieburth & F.A. Villaruel
(Eds.), Making Invisible Latino Adolescents Visible: A Critical Approach
to Latino Diversity. New York: Falmer.
Stanton-Salazar, Ricardo. 2001. Defensive Network Orientations as
Internalized Oppression How Schools Mediate the Influence of Social
Class on Adolescent Development. In Biddle, Bruce J. (Ed.), Social
Class, Poverty and Education Policy and Practice. New York,
London: RoutelegeFalmer.
Stanton-Salazar, Ricardo. 2001. Manufacturing Hope and Despair The School
and Kin Support Networks of U.S.-Mexican American Youth. New York,
London: Teachers College Press.
Stanton-Salazar, Ricardo D., Vasquez, Olga A., & Mehan, Hugh. 1996.
Engineering Success Through Institutional Support. In A. Hurtado, R.
Figueroa, & E.E. Garcia (Eds.), Strategic Interventions in Education:
Expanding the Latina/Latino Pipeline. (100-136). Santa Cruz: Regents of
the University of California.
Steinberg, Laurence, Sanford M. Dombusch & B. Bradford Brown. 1992.
Ethnic Differences in Adolescent Achievement: An Ecological
Perspective. American Psychologist, 47(6): 723-39.
Teachman, Jay. 1987. Family Background, Educational Resources, and
Educational Attainment. American Sociological Review, Vol.

Teachman, Jay D. 1996. Intellectual Skill and Academic Performance: Do
Families Bias the Relationship? Sociology of Education, Vol. 69
Teachman, Jay, Carver, K., & Paasch, K. 1997. Social Capital and the
Generation of Human Capital. Social Forces, Vol. 75, No. 4: 1343-
Teachman, Jay D., Paasch, Kathleen and Carver, Karen. 1996. Social Capital
and Dropping Out of School Early. Journal of Marriage and the
Family., Volume 58:773-783.
Valenzuela, Angela. 1999. Subtractive Schooling: U.S.-Mexican Youth and the
Policies of Caring. New York: State University of New York Press.
Vaux, Alan & Wood, J. 1987. Social Support Resources, Behavior, and
Appraisals: A Path Analysis. Social Behavior and Personality: An
International Journal, 15:107-111.
Vaux, Alan, Burda, Philip & Stewart, Doreen. 1986. Orientation Toward
Utilization of Support Resources. Journal of Community Psychology,
14 (April):159-170.
Wallace, John L & Vaux, A. 1993. Social Support Network Orientation: The
Role of Adult Attachment Style. Journal of Social and Clinical
Psychology, Vol. 12, No. 3:354-365.
Wellman, B. 1983. Network Analysis: Some Basic Principles. In Bert Adams
and R.A. Sydie (Eds.) Sociological Theory, (pp.155-200). San
Francisco: Jossey-Bass.
Wexler, Phillip 1992. Becoming Somebody: Toward a Social Psychology of
Schoo., London: Falmer Press.
Whitbeck, Les B., Simons, Ronald L., Conger, Rand D., Wickrama, K. A. S.
Ackley, Kevin A.& Elder, Glen H. Jr. 1997. The Effects of Parents
Working Conditions and Family Economic Hardship on Parenting
Behaviors and Childrens Self-Efficacy. Social Psychology Quarterly,
Vol. 60, No. 4:291-303.

White, Micheal J. & Glick, Jennifer E. 2000. Generation Status, Social Capital
and the Routes out of High School. Sociological Forum, Vol. 15, No.
Zinck, Kirk & Littrell, John. 2000. Action Research Shows Group Counseling
Effective with At-Risk Adolescent Girls. Professional School
Counseling, Vol. 4, No. 1:50-59.