Current Approaches In Investigating Inequalities In Access To Post Secondary Education

Current Approaches in Investigating Inequalities in Access to Postsecondary Education Presented by Iria Puyosa

November 10, 2005

Table of Contents Introduction ....................................................................................................................3 I. The Problem of Inequalities in Access to PSE ..........................................................8 1) Access to PSE in the United States.......................................................................8 2) Financial Barriers and Academic Deficits............................................................10 3) Importance of Equal Opportunities in PSE Access .............................................14 II. Brief Overview of Major Theories Addressing PSE Access...................................16 1) Rational Choice Theory: No Information Asymmetries? .....................................16 2) Social Reproduction Theory: No Room for Upward Mobility?.............................18 3) Status Attainment Theory: No Extra-familial Resources?....................................20 III. Fundamental Concepts in Social Capital Theory ...................................................21 1) Major Approaches to Social Capital in the Sociology Literature..........................21 2) Toward a Social Capital Explanation of Inequalities in Access to PSE...............28 IV. Research on PSE Access Using Social Capital Theory ........................................34 1) Conceptualizing Social Capital in Postsecondary Education Literature..............36 2) Modeling Approaches in Researching Social Capital Effects on PSE Access....42 3) Findings on social capital effects on postsecondary education access ..............52 V. Limitations in Using Social Capital Theory in PSE Access Research ....................57 1) Avoiding Social Capital Conceptual Tautologies.................................................57 2) Developing Robust Testable Social Capital Postulates ......................................58 3) Developing Reliable Social Capital Measures.....................................................59 VI. Directions for Further Research ............................................................................62 1) Framework for Analyzing Social Capital Effects in PSE Access.........................62 2) Directions for Future Research in Social Capital Effects on PSE Access ..........64

2

Introduction Inequality in access to postsecondary education is arguably the most critical issue in educational public policy (Heller 2001; St. John 2004). In the United States, the overall gap in postsecondary education participation rates between the highest and the lowest income quintiles is roughly 40% for the high school class of 1992 (Baum and Payea 2004). Additionally, gaps in college participation between young White adults and marginalized ethnic groups persist (Measuring Up, 2004). Postsecondary education access inequalities carry economic, political, and social consequences. Foremost among the economic consequences is the income gap between high school graduates and college graduates has increased significantly over time. Postsecondary education graduates higher levels of civic participation and volunteering is one the most important political consequences (Langelett 2002; Kingston, Hubbard et al. 2003; Baum and Payea 2004). Additionally, research suggests that postsecondary education changes peoples attitudes from accepting the status quo toward taking initiative to build a more prosperous and equitable society (Pascarella and Terenzini, 1991; Langelett 2002). Given the importance of the subject, is not surprising that several theories have been used to explain why there are differences in postsecondary education access. Most researchers have built their conceptual frameworks on one of the following theories: rational choice, social reproduction, or status attainment. An emerging theoretical approach is social capital theory, which provides a conceptual framework for

integrating claims and findings from the above-mentioned theories, and addressing unanswered questions. This examination presents a synthesis of the social capital theoretical concepts found in general sociology and the sociology of education literatures. From the variety of notions of social capital, I adopt the one provided by Lin (2001a), which can be summarized as follow: social capital consists of the accumulation of resources embedded in a social structure or network that is accessed and/or mobilized by an individual in purposive actions with the expectation of obtaining some utility or return on investment. A growing body of research has shown significant relationships between social capital and inequalities in PSE access, although the interpretation of such findings has often been theoretically feeble (Baron, Field et al. 2000; Dika and Singh 2002; Field 2003). This review analyzes a corpus of empirical research studying postsecondary education access, limited to those in which the outcome variable is one of the following: enrollment in any post-secondary institution, enrollment in a 4-year institution, enrollment in a 2-year institution, and number of years of schooling. The review intends to answer the following research questions: 1) How do social capital factors contribute to explain inequalities in access to PSE? 1.1)

How has access to PSE been defined in the higher education literature? What inequalities in access to PSE are currently foremost in the United States? Why are equal opportunities of access to PSE important for individuals?

1.2)

How have PSE access inequalities been explained by rational choice, social reproduction and status attainment theories?

1.3)

Which are the most influential approaches to social capital in the sociology of education literature? Which are the major concepts in social capital theory? What process in the formation of social capital seems to be more relevant for explaining inequalities in access to PSE? What social capital factors seem to be more relevant for explaining inequalities in access to PSE? Which processes are critical for an individual to mobilize social capital in order to access PSE?

1.4)

How has social capital theory been used in the access to PSE literature? How have social capital variables been measured and entered into multivariate models when studying access to PSE? Which are the relevant findings in social capital effects on PSE access?

1.5)

What are the limitations in the use of social capital theory in studying inequalities in PSE access? How should a conceptual framework for studying social capital effects in PSE access be depicted? How may common pitfalls in social capital effects on PSE access research be solved?

The response to above questions is structured as follows. In the first section, I present a brief overview of the problem of access to postsecondary education in the United States, focusing on financial and academic barriers. Additionally, I portray the importance of attaining a postsecondary education by pointing out individual benefits from it. In the second section, I present an overview of the major theories addressing postsecondary education access: rational choice, social reproduction, and status attainment.

In the third section, I review approaches to social capital often used in the sociology of education literature, focusing on the contributions of Bourdieu, Coleman, and Lin. Then, I synthesize the major concepts in social capital theory that are relevant for explaining inequalities in access to postsecondary education. I conclude this section by identifying three ways in which social capital may generate returns related to an individual education. In the fourth section, I review empirical research on postsecondary education access in which social capital constructs have been incorporated into explanatory models. I analyze and critique these empirical works’ conceptual frameworks, measures of social capital included, and the modeling approaches applied. Finally, I summarize relevant findings on social capital effects on postsecondary education access and point out some of their limitations. In the fifth section, I dig more into the limitations found in the postsecondary education access literature. I focus on three major limitations: 1) conceptual vagueness; 2) weaknesses in testable postulates; and 3) lack of reliable measures. Then, I propose some guiding ideas on how to avoid conceptual tautologies, how to develop robust testable social capital postulates, and how to develop reliable measures of social capital constructs. In the sixth, closing section, I draw a framework for studying social capital effects in postsecondary education access that should help to fill conceptual gaps in the literature, and propose directions for measuring social capital variables.

Overall, I expect that this review will portray a view of the current body of knowledge regarding social capital effects on postsecondary education access inequalities, and help to advance a promising research agenda for understanding the problem, predicting students’ outcomes, and informing public policy in this regard.

I. The Problem of Inequalities in Access to PSE The problem of inequality in access to postsecondary education is the most crucial issue in educational public policy at present. Consensus exists about the importance of addressing this problem and the need for progressing toward equitable access opportunities for students from all socioeconomic backgrounds. Nonetheless, what framework should guide public policy toward such end is vigorously disputed. Indeed, postsecondary education access inequalities have been researched through several approaches, focusing on different aspects of the problem. Mainstream research has framed postsecondary education access inequalities as either academic or financial challenges, often neglecting competing explanations (Fitzgerald & Delaney, 2002; Heller, 2001; St. John, 2002; St. John, 2004). Academic access usually refers to students’ qualifications and ability to meet standards for admission to a 4-year postsecondary institution, while financial access refers to the ability to afford enrollment in any postsecondary institution (St. John, 2002).

1) Access to PSE in the United States In the United States, inequalities in access to postsecondary education have been evident for several decades. There was some improvement in the late 1960s and the 1970s, due to progressive policies such as need-based financial aid and affirmative action (Heller, 2001; St. John, 2004). However, at the beginning of the 21st century, trends toward equalizing access are reversing (Baum & Payea, 2004; St. John, 2002, 2004).

According to the National Education Longitudinal Study 1988-2000, more than three fourths of the U.S. 8th graders aspire to some sort of postsecondary education, and there is not significant difference in aspirations among income levels. Nonetheless, participation rate in higher education across different socio-economic segments is unequal. The overall gap in participation rates between the highest and the lowest income quintiles is roughly 40% for the high school class of 1992. Among the 1992 high school graduates, 97% of the individuals in the highest income quartile whose parents held a terminal degree enrolled in post-secondary education within 20 months of high school graduation. Conversely, only 52% of individuals in the lowest income quartile whose parents were high school dropouts enrolled in some sort of postsecondary education (Baum and Payea 2004). Gaps in college participation between high and low income students have widened over the last ten years (Measuring Up, 2004). Low income students are almost twice as likely to begin postsecondary education in a public 2-year institution (40%) as in private 4-year colleges (21%), while high income students are almost as likely to begin postsecondary education in a private 4-year colleges (38%) as in public 4-year colleges (41%), and are less likely to enroll in 2-year public colleges (21%) (Baum and Payea 2004). Additionally, gaps in college participation between White young adults and youth from marginalized ethnic groups persist (Measuring Up, 2004). Both income and racial gaps are interrelated, since 40% of Latino families and 34% of African American families are below the poverty line, while only 15% of White families are in that socioeconomic level (Carnevale and Fry 2002).

2) Financial Barriers and Academic Deficits Both financial and academic issues affecting postsecondary education access have been widely addressed by researchers. Mounting evidence supports claims on the individual and financial situation effects on postsecondary education access inequalities. Similarly, evidence of effects from differential academic preparation exists. Lack of resources to pay for tuition and other expenses has been identified as the biggest obstacle to postsecondary education enrollment (Hossler, Schmit, & Vesper, 1999; Hu & Hossler, 2000; Hu, 2003). Affording costs of attendance is obviously more challenging for low-income students. By 1999, the cost of attending a 4-year postsecondary education institution represented 5% of the income of an upper class family, 17% of the income of a middle class family, and 62% of the income of a lower class family (Gladieux 2002). The effect of family income on ability to pay is unquestionable, though policy researchers argue about the effectiveness of financial aid policies for mitigating the unequal starting point and allowing for access equity (Becker, 2004; DesJardins, McCall, Ahlburg, & Moye, 2002; Heller, 2004; Lee, 2002; Paulsen & St. John, 2002; St. John, Chung, Musoba, & Simmons, 2004; Terenzini, Cabrera, & Bernal, 2001). In 2003-04, the average Pell Grant covered 23% of the total charges at the average 4-year public institution, down from 35% in 1980-81 (Baum and Payea 2004). The declining trend in the purchasing power of Pell Grants is mirrored by college participation gaps both for marginalized ethnic groups compared to Whites and for low income students compared to high-income students during the same period (St. John, 2002). These trends indicate that inadequate financial aid yielding to growing unmet

need is increasingly constraining postsecondary education access enrollment for lowincome students (Fitzgerald & Delaney, 2002; Lee, 2002; St. John, 2002). Middle income students also have unmet need for covering postsecondary education costs. The expected family contribution (EFC) is calculated by using a formula that includes not only income, but family size, savings, students’ earnings, and other related variables. In general, EFC provides a more accurate estimation of ability to afford postsecondary education expenses than a rough measure of income would provide. However, the EFC still overestimates middle class families’ discretionary income to pay for their children’s postsecondary education expenses (Lee, 2002). Middle income parents complain that it does not take in account payments for house mortgages, insurance, and other loans. Differences between the EFC and middle income families actual disposable income have been voiced for these interest groups representing middle class constituencies. Their demands influenced politicians’ and policymakers’ decisions to switch financial aid provisions from need-based grants toward subsidized loans and merit-based grants (Gladieux 2002; Lee 2002). Substituting merit-based grants for need-based financial aid is also a growing trend at the state level, since the introduction of the Georgia Hope program in 1993. Most states’ merit aid programs appear to do nothing for equalizing opportunities for access to postsecondary education, since such grants tend to be awarded to uppermiddle and high income students rather than low and low-middle income students (Heller 2002). Overall, financial aid policies in the late 1990s and early 2000s have been ineffective with respect to reaching the goal of social equity in postsecondary education access (McPherson and Schapiro 2002).

Research has shown that low income high schools students are more concerned about postsecondary education costs than they are about any other factor affecting college choice. Indeed, low income students show greater price-sensitivity than others income groups (Terenzini, Cabrera et al. 2001; Fitzgerald and Delaney 2002). More to the point, data collected by the American Council of Education indicates that low income families tend to overestimate postsecondary education costs (Gladieux 2002). Since expectations about college costs have even greater effects on enrollment decisions than actual costs have, such overestimation extensively affects postsecondary education access for low income students (Hu and Hossler 2000; Fitzgerald and Delaney 2002; Gladieux 2002; Paulsen and St. John 2002). Paradoxically, low income students receive less (Fitzgerald and Delaney 2002) about availability of financial aid than middle and upper income students (Fitzgerald and Delaney 2002). Many low and middle income students eligible for need-based financial fail to apply for college admission and for financial aid, apparently because they did not receive adequate school counseling (Horn and Chen 1998; Fitzgerald and Delaney 2002). Such lack of information may also discourage low income students from taking college preparatory courses and preparing for college entrance examinations (Fitzgerald & Delaney, 2002; St. John, 2002). Investing in academic preparation may be considered futile by low income students, who are not aware of financial aid opportunities that could make postsecondary education affordable for them. Alternatively, the U.S. Department of Education sponsored research claims that differences academic resources—as opposed to financial issues—explain differences in postsecondary education access. Most of the research under this approach has used

either the College Qualification Index (Berkner and Chávez ,1997)—which includes high school GPA, senior class school rank, ACT or SAT scores, NELS 1992 aptitude test, and high school coursework—or the Adelman’s Academic Resources Index (Adelman, 1999)—which includes the highest level of math taken in high school, non remedial math and English courses, AP courses, core science labs, foreign language courses, computer courses, high school GPA quintiles, high school rank, and NCES test scores. Not surprisingly, most of these academic variables are positive predictors of enrollment in postsecondary institutions (Adelman 1999; Cabrera, La Nasa et al. 2001). Such a finding should be expected since high school GPA, high school rank, standardized test scores, and high school coursework are generally considered for granting admission. Moreover, students who score high in these measures are encouraged to pursue postsecondary education, while those who score low are discouraged. Nonetheless, requiring taking standardized tests, making high quality and high intensity curriculum mandatory, or increasing high school graduation requirements seem to be counterproductive. St. John et al. (2004) found that states’ policies in this regard do not have significant effect on college enrollment rates. However, such policies do negatively affect high school graduation rates (St. John, Musoba, & Chung, 2004), diminishing chances of entering in postsecondary education pipeline for low income students and marginalized ethnic groups (Paulsen & St. John, 1997, 2002; St. John, Paulsen, & Starkey, 1996; St. John, 2002; St. John, Hu, Simmons, Carter, & Weber, 2004). Research indicates that differences in academic preparation may be due to schools’ lack of capacity for offering college preparatory courses, quality teaching, and

adequate counseling. Since college enrollment rates have been found to be higher for students who participate in academic or college preparatory curricular tracks in high school (Hossler, Braxton, & Coopersmith, 1989), lack of school capacity to offer such coursework

to most students

clearly undermine their postsecondary access

opportunities. All the same, low income students in urban and rural schools—especially if they are also from marginalized ethnic groups—are also more likely to select nonacademic courses even though academic ones may be available. Additionally, they are less likely to prepare for college entrance examinations (Horn and Chen 1998; Terenzini, Cabrera et al. 2001). These students may be self-selecting out the postsecondary education pipeline due to family and personal perception of inability to afford college costs (Fitzgerald & Delaney, 2002; St. John, 2002; Terenzini et al., 2001).

3) Importance of Equal Opportunities in PSE Access Equity in postsecondary education access is highly consequential for individuals because such inequalities carry economic, political, and social effects. Economic consequences include increasing earning capability and greater employability. For instance, the income gap between high school graduates and college graduates has increased significantly over time. Bachelor degree recipients earn on average 75% more annually than high school graduates (Bureau of Labor Statistics 2004). According U.S. Census data, median annual earnings for college graduates in 2004 is nearly $50,000, while for high school graduates is roughly $30,000 (Baum and Payea 2004). Some scholars argue that this earnings premium is due to greater productivity after having taken advantage of opportunities for developing skills and acquiring knowledge (Becker 1993; Langelett 2002; Kingston, Hubbard et al. 2003). Others pose that a

postsecondary education credential itself is an advantage in the labor market (Kingston, Hubbard et al. 2003). A third explanation is that attending postsecondary education enhances personal social networks, which will also boost one’s ability to mobilize resources in order to get a better paid job (Lin 1999). At any rate, the greater earning capacity associated with postsecondary education is clearly a source of advantage for individuals. Another clear advantage comes from the fact that unemployment rates are consistently lower among more educated groups of individuals (Baum and Payea 2004). Additionally, there are social benefits associated with postsecondary education. Research suggests that postsecondary education changes people’s attitudes from accepting the status quo toward taking initiatives to build a more prosperous and equitable society (Langelett 2002; Pascarella and Terenzini, 1991). Such attitude changes may result in postsecondary education graduates generally displaying higher levels of civic participation and volunteering (Langelett 2002; Kingston, Hubbard et al. 2003; Baum and Payea 2004). Crime rates also decline when educational attainment increases (Hossler, Braxton, & Coopersmith, 1989). In summary, individuals who attend postsecondary education institutions accrue a series of economic, social, and political benefits, while those who do not attend are left behind in terms of employment opportunities, quality of life, and ability to take part in the public life.

II. Brief Overview of Major Theories Addressing PSE Access Scholars studying inequalities in postsecondary education access have employed a wide array of theoretical frameworks and methodological approaches. Mainstream research in postsecondary education access is commonly grounded in the disciplines of economics, sociology, or psychology. Most researchers bring into play one of the following theories: rational choice, social reproduction, or status attainment. In this section, I will summarize the conceptual assumptions of these three lines of research, and point out their limitations in explaining postsecondary education access inequalities.

1) Rational Choice Theory: No Information Asymmetries? Rational choice theory states that individuals make their choices upon the analysis of both the expected utility and the expected costs of the alternatives available to them (Becker, 1993; Cohn and Geske, 1990; Schultz, 1961). The basic assumption is that individuals seek the maximization of their utility (Schultz, 1961; Coleman, 1990; Becker, 1993). Accordingly, a student would assess the expected utility and the expected costs of enrolling in a postsecondary education institution, and then, s/he would decide whether to apply and/or to enroll if the expected utility outweighs the expected costs (Becker, 1993; Beekhoven, De Jong, & Van Hout, 2002; DesJardins & Toutkoushian, 2005; Hossler, Braxton, & Coopersmith, 1989). Most of the research applying rational choice theory to postsecondary education enrollment actually focuses on the analysis of how economic factors—such as tuition, financial aid, and family income—affect student enrollment choices (Beekhoven, De

Jong, & Van Hout, 2002; Cohn and Geske, 1990; DesJardins & Toutkoushian, 2005; Hossler, Braxton, & Coopersmith, 1989; Hossler, Schmit, & Vesper, 1999). There is also a stream of research analyzing return on educational investment focused on life-long earnings and other economic payoffs from a postsecondary education (Becker, 1993; Cohn and Geske, 1990; Langelett, 2002; Manski & Wise, 1983; Schultz, 1961). Available alternatives are limited by individual preferences, and, more importantly, the constraints faced by each person (Beekhoven, De Jong, & Van Hout, 2002; Cohn and Geske, 1990). Nonetheless, fewer works attempt to analyze students’ preferences and the actual alternatives available to them (Cohn and Geske, 1990), partly because such data is not readily available. Thus, many researchers are limited to examining the choices made by individual students, controlling for their socioeconomic background. Based upon patterns found among individuals grouped by demographic or socioeconomic characteristics, scholars pose broad inferences about why group differences (e.g. differences by income, race, or parental education) are found consistently (Beekhoven, De Jong, & Van Hout, 2002; Cabrera, La Nasa, & Burkum, 2001; Hossler, Schmit, & Vesper, 1999; Perna, 2000; Terenzini, Cabrera, La Nasa, & al., 2000). Moreover, there are few studies on postsecondary education choice that consider non-pecuniary issues affecting students’ preferences (e.g. available majors, closeness to family and friends), evaluation of expected utility (e.g. college life, academic reputation), and self-assessment of probability of success (e.g. institution selectivity, self-assessment of academic ability) (Hossler, Braxton, & Coopersmith, 1989; Cohn and Geske, 1990; Perna, 2000; DesJardins & Toutkoushian, 2005).

Measures of constructs related to those issues should be added to traditional econometrics models for increasing their explanatory power. The most important limitation of most of the empirical research grounded in rational choice theory is the failure to control for information asymmetries among students from different socio-economic groups. Many works overlook the fact that most students may not have accurate information about either cost of attending a postsecondary education institution or potential future earnings from a postsecondary degree (DesJardins & Toutkoushian, 2005). Information asymmetries among students from different socioeconomic backgrounds may be critical for explaining inequalities in access to postsecondary education. Some researchers control for sources information available to students during their choice process (Hossler, Schmit et al. 1999; Cabrera, La Nasa et al. 2000; Cabrera, La Nasa et al. 2001); however, measures of actual usage of such information are seldom (if at all) found in empirical research on postsecondary education access inequalities. Theoretical frameworks and analytical models that incorporate this sort of measures are required to explain inequalities in access to postsecondary education.

2) Social Reproduction Theory: No Room for Upward Mobility? Social reproduction theory states that educational institutions and policies are instrumental in maintaining and reinforcing class-based social and economic stratification (Paulsen and St. John 2002). Within this framework, educational alternatives are constrained by social class, and, at the same time, individual’s preferences are shaped by ascribed status (St. John, Hu, Simmons, Carter, & Weber, 2004). Scholars using this theoretical lens claim that students’ postsecondary education

choices are highly context dependent (Paulsen and St. John 2002). Accordingly, students’ predispositions toward educational attainment may also depend on social class. Some researchers emphasize that the stratified structure of the United States educational system as another device supporting social class reproduction (McDonough 1997; Paulsen and St. John 2002). Some recent research works within this approach have incorporated the notions of cultural capital and habitus posed by Bourdieu (1977a, 1977b; Bourdieu & Passeron, 1990) in order to explain differences in educational attainment norms or aspirations that seem to be related to social class (McDonough 1997; Karen 2002; Paulsen and St. John 2002). There are two major limitations to using social reproduction theory as theoretical lens. First, it portrays individuals as powerless for making choices based upon their preferences and interests. Second, but not least important, this theory does not provide much latitude for explaining upward social mobility, although evidence of such mobility based on educational attainment obviously exists. Thus, using social reproduction theory may render analyses at the individual level meaningless, since the theoretical premise is that educational outcomes are determined by the social structure. Social reproduction theory brings to the scholarly discussion the key issue of the social structure influence on actual postsecondary education alternatives and preferences held by individuals. That is an important contribution to the field. Nonetheless, this theory has little heuristic value for understanding why some individuals are able to overcome their socioeconomic status origin while others are not

able to do so. This theory also leaves unanswered an even more puzzling question: why some individuals from privileged socioeconomic status origin fail to achieve the postsecondary education success that the social structure allegedly granted to them?

3) Status Attainment Theory: No Extra-familial Resources? The influential status attainment theory—developed by Blau and Duncan (1967) —states that family socioeconomic background and student academic ability interact to configure educational aspirations, and eventually to determine postsecondary education enrollment. According to status attainment theory, parental education and occupational status influence the next generation’s educational outcomes by shaping aspirations and expectations (Hossler, Braxton, & Coopersmith, 1989; St. John et al., 2004). Sociopsychological variables measuring individual motivations and encouragement from significant others have often been added to status attainment models (Hossler, Braxton, & Coopersmith, 1989; Xie & Goyette, 2003). Additionally, some status attainment models include measures related to contextual variables such as high school norms, and peer culture effects (Hossler, Braxton, & Coopersmith, 1989). The main limitation found in works grounded in status attainment theory is the tendency to overlook the effects of financial factors—such as tuition, financial aid, and unmet need. A second important limitation is status attainment theory narrow focus on familial resources, while disregarding effects from community and school resources.

III. Fundamental Concepts in Social Capital Theory Social capital theory provides an appropriate conceptual framework for understanding how contextual differences affect students’ academic and financial decisions in the path to postsecondary, while pondering both effects from individuals’ actions and effects from the social structure This theory presents avenues for filling the gaps identified in most mainstream research: little analysis of information asymmetries, weak explanations for both upward and downward mobility in the social structure, and overlook of the effects of extrafamilial resources. Differences in social capital may account for differences in students’ understanding of alternatives available to them, their preferences when making choices, their subjective valuation of the utility derived from each step in the path to college, and their self-evaluation of probability of success.

1) Major Approaches to Social Capital in the Sociology Literature The notion of social capital has been traced back to ideas in the works of late 19 th and early 20th centuries sociologists, such as Durkheim (1893) (e.g. organic solidarity), Tönnies (1897) (e.g, purposive associations), and Weber (1922) (e.g. shared style of life) (Adam and Roncevic 2003; Field 2003). Nevertheless, the concept only starts to mature by the 1960s and early 1980s, driven by Bourdieu’s (1977; Bourdieu and Passeron, 1977) theoretical propositions regarding social structure reproduction, Granovetter’s doctoral dissertation on information, mobility and job market (1970; published 1973; article version, 2003), and Lin’s (1982) study about the use of social resources for occupational attainment.

In 1990, Coleman placed the concept of social capital in the core of his rational action theory,

opening the path for disciplinary development. Although the

popularization of social capital owed much to Putnam (Putnam, Leonardi & Nanetti, 1993; 2000; Putnam, Felsten & Cohen, 2003), application of the concept to educational research is better understood within sociology. For that reason, this review will focus on Bourdieu’s, Coleman’s, and Lin’s contributions to the development of social capital theory. 1) Bourdieu: Social Capital and Institutionalized Access to Collective Resources The French sociologist Pierre Bourdieu has often been credited as a the first who attempted to systematize the concept of social capital (Baron, Field et al. 2000; Dika and Singh 2002; Adam and Roncevic 2003). Along with his intuitive definition of cultural capital, the concept of social capital is an important piece in Bourdieu’s theory of the social structure reproduction. In his influential article The Forms of Capital, Bourdieu (1986) states that: Social capital is the aggregate of the actual or potential resources which are linked to possession of a durable network of more or less institutionalized relationships of mutual acquaintance and recognition–or in other words, to membership in a group--which provides each of its members with the backing of the collectivity-owned capital, a "credential" which entitles them to credit, in the various senses of the word. These relationships may exist only in the practical state, in material and/or symbolic exchanges which help to maintain them (pp. 248-249). In simpler terms, the social capital possessed by a given agent depends on the economic or cultural capital possessed by those to whom s/he is connected and her/his actual ability to mobilize such capital effectively (Bourdieu 1986). Hence, Bourdieu's

notion of social capital entails two components: first, the relationships that allows an individual to claim resources possessed by her/his social group, and, second, the quantity and quality of such resources (Siisiäinen 2000; Dika and Singh 2002). Bourdieu (1986) argues that the ability of the dominant class to preserve the capital accumulated in previous generations and to pass it to the next generation assures the reproduction of the social structure. That appropriation process occurs by means of institutionalization devices. Economic capital is institutionalized in the form of property rights, whereas cultural capital is institutionalized in the form of educational qualifications. The institutionalization of social capital is less clear since it is made up of social obligations whose transferability is more limited. Nonetheless, Bourdieu indicates several ways to institutionalize social capital, including titles of nobility, elite schools alumni associations, selective club memberships, and so on. Such connections would allow privileged individuals to obtain greater profits from economic or cultural capital virtually equivalent to what is held by other individuals who lack similar connections (Siisiäinen 2000). One caveat is that Bourdieu's concept of social capital should operationalized including both field and habitus, the two central concepts in his work (Siisiäinen 2000; Dika and Singh 2002). Field defined as the space in the social structure in which individuals strive to achieve their ends while adjusting to tacit rules of action. Bourdieu uses the term habitus to amalgamate a set of dispositions, responses, and behavior patterns that people have acquired through acting in the different positions they hold in social settings (e.g. family, schools, neighborhoods, clubs, voluntary groups, churches, political parties) (Bourdieu, 1977; Bourdieu and Passeron, 1977). The habitus creates a

stable generative principle that guides the individual in making choices between alternatives (Siisiäinen 2000). From Bourdieu’s conceptualization, research on social capital effects on postsecondary education access should adopt the notion of institutionalized devices for of appropriating resources belonging to their social relationships. Researchers should identify measures of devices that help individuals in the process of mobilizing social capital for achieving educational goals. 2) Coleman: Social Capital and Norms to Guide Rational Action Coleman’s overarching theoretical interest was how social structure enables action, and how rational choice operates within different social settings (Coleman 1990). Coleman (1988; 1990) argues that an individual’s rational action is enabled by the social capital available to her/him within a given social structure (e.g. a community, an industry, the educational system). According to Coleman (1990), social capital is an intangible sort of capital—alike human capital— embedded into relationships, and having the forms of trust (obligations and expectations), norms, and information. In Coleman’s conceptualization, trust results from two or more parties’ engaging in actions that provide a benefit to each other on the basis of expectations and obligations of reciprocity that are usually not formally sanctioned. Norms are somewhat more formalized aids for maintaining a purposeful exchange with other actors. Norms are enforced by sanctions, which are either rewards for performing proper actions or punishments for performing improper actions. Finally,

information that facilitates action will be available through exchange between and among actors. Focusing on how social capital relates to educational attainment, Coleman (1988; 1990) defines it as a set of resources embedded in family and community that are useful for the cognitive and/or social development of children and adolescents. According to Coleman, the most important of the intervening factors in the formation of such social capital is closure among individuals belonging to the same community. Closure allows establishing and reinforcing norms, help overcome power imbalance, and help build trust (Coleman 1990). Moreover, intergenerational closure helps to pass norms, information and trust from adults to children (Coleman, 1988). Thus, stability among members of the collectivity (e.g. few residential moves) and a common ideology (e.g. same religion) are regarded as contributing factors for the maintenance of social capital (Coleman 1990). The idea that intergenerational closure helps to institutionalize positive norms has been influential in educational research, and has bolstered policies promoting parental involvement and cohesive communities (Dika and Singh 2002). The emphases on community closure have distracted researchers from pursuing studies related to Coleman´s central hypothesis of social capital as instrumental for rational choice. Examining the interplay among social settings, social capital, and rational choice should be the path for a next generation of studies on social capital and postsecondary education access. 3) Lin: Social Capital and Resources Mobilized in Purposive Action

Establishing an analogy with other forms of capital, Lin (2001c) states that individuals generate social capital by making an “investment in social relations with expected returns” (pg. 6). Lin (2001a) defines social capital “as resources embedded in a social structure that are accessed and/or mobilized in purposive actions” (pg. 58). There are three main points in his conceptualization: 1) Social capital is structurally embedded in society, community, and groups; 2) it can be mobilized by individuals; and 3) individuals’ goals drive such mobilization. Furthermore, Lin (2001c) argues that social capital enhances individual’s action outcomes because it 1) facilitates the flow of information, 2) helps to exert influence on decision makers, 3) certifies an individual’s social credentials, and 4) reinforces identity and group recognition. Figure 1 depicts Lin’s (2001a) model of social capital theory. The social structure and the individual's position are pre-conditions and precursors, which facilitate or constrain the investment in social capital. Positional elements and structural elements (e.g. ideology, culture, level of industrialization, technological development, aggregated educational level, physical resources, productivity, and so on) affect opportunities to build and maintain social capital. Thus, inequalities shaping the distribution of collective assets embedded in an individual’s network lead to inequalities in resources mobilization, and, consequently, to inequalities on returns. However, individuals are not powerless; a person’s ability to mobilize embedded resources (e.g. money, power, and prestige) in order to achieve her/his goals (e.g. enrolling in PSE) increases or diminishes as a result of individual’s choices and actions. Lin (2001c) poses three major hypotheses about how social capital works: 1) the social position of origin has a positive effect on accessing and mobilizing social capital

(structural effect); 2) the use of strong ties1 will positively affect the success of expressive action, supporting the maintenance of already possessed resources; and, 3) the use of weak ties or bridges will increase access to heterogeneous resources and ability to mobilize social capital (action effect) for accomplishing instrumental actions. Figure 1 Lin’s model of social capital theory

Source: Lin, 2001a.

1

Each individual is supposed to be connected to a social network made by strong ties and weak ties.

Strong ties are those relationships showing a high degree of intimacy, frequent contacts, reciprocity, and acknowledgement of obligations Lin, N. (2001). Social Capital. A Theory of Social Structure and Action. New York, Cambridge University Press. , Granovetter, M. (2003). The Strength of Weak Ties. Networks in the Knowledge Economy. R. Cross, A. Parker and L. Sasson. New York, NY, Oxford University Press. . Strong ties yield to sharing resources and reinforcing lifestyles (Lin, 2001c). Weak ties are those that link or bridge to other social circles with dissimilar characteristics and different lifestyles, as well as others assortments of resources (Lin, 2001c).

Lin’s (1978; 1999) studies suggest that access to hierarchical positions is a critical factor in the process of status attainment. Provision of access to higher hierarchical positions may be the reason why weak ties yield greater return than strong ties (Granovetter, 1973; Burt, 1992). For instance, college recruiters visiting a high school may have greater influence than parents on institutional choice, because they hold a higher position in the educational system. Contrary to Coleman’s (1988) closure hypothesis, Lin (1978; 1999) indicates that extensive networks also have greater return than constrained networks when trying to obtain new resources. Lin (1999, 2001c) hypothesizes that closure may be important for maintaining resources already possessed by the individual, but an open network, having numerous bridges, may be more suitable for obtaining additional resources. To recapitulate, Lin (1999, 2001a, 2001b, 2001c) offers a comprehensive theory of social capital. His works advance explanations on the major issues: 1) how social capital is defined; 2) how social settings interact with availability of resources; 3) how individual accesses those resources; 4) which type of effects can be expected from social capital mobilization. Those explanations provide grounds for a developing a set of testable postulates on social capital effects on postsecondary education access.

2) Toward a Social Capital Explanation of Inequalities in Access to PSE Building upon the fundamental concepts in this literature, I will suggest three ways in which social capital may generate returns for individuals. The analysis of these three return paths should be useful for examining how social capital affects postsecondary education access opportunities and inequalities.

2.1) Social Capital Effects on Access to Relevant Information The most important way in which social capital affects postsecondary education access is through inequalities in accessibility of relevant information. As stated by Lin (2001c), social capital facilitates the flow of information, and consequently it enhances an individual’s ability to attain her/his goals. Indeed, a student needs to get relevant information in order to make opportune decisions that increase her/his likelihood of enrolling in postsecondary education (e.g. take algebra by 9 th grade or earlier, fill applications for financial aid, and so on). Social capital helps individuals to deal with problems of asymmetric information and to enhance efficiency in decision-making by allowing the use of social interaction mechanisms such as copying and pooling. Accordance to Collier (2002), both copying and pooling facilitate the gathering of information about alternative courses of action. Copying is a somewhat simple mechanism based upon observation of those that are higher in a given hierarchy. For example, sophomore high school student Paul may learn what courses should he takes in order to prepare for college by looking at what his basketball teammate Barry—who is a senior—took. Pooling is a more interactive process requiring a purposeful exchange of information. For instance, Norma may identify the opportunity of obtaining a scholarship for attending a small liberal arts college after discussing her goals with Susan, who is not interested in such scholarship because she wants to attend a large institution. Nonetheless, opportunities for productive copying and pooling may not available for low income and ghetto students. If high school junior Robert does not have any friend or acquaintance who has applied for college admission, copying is not available

to him. Indeed, he will need to put more effort in gathering information on how to apply for college than other students having college going acquaintances. Likewise, pooling resources on postsecondary education opportunities may be a straightforward process for middle and high income students with college educated parents, but it is more demanding for poor kids whose parents have not attained any postsecondary education. Thus, access to relevant information may depend on a student’s social networks. According to Granovetter (1973) ties to individuals from different social circles may facilitate access to useful information that otherwise would not be available. Social networks linking an individual to others who possesses better resources is required in order to bring in a progressive flow of information. The effect of social networks in giving access to information may explain why extracurricular activities and community involvement are correlated with educational attainment. Indeed, those activities provide an extensive network from which an individual could draw information about resources that help her/him to successfully complete college preparatory coursework in high school and obtain guidance on how to apply for postsecondary education. On the contrary, constrained networks—having few links to other circles—may limit the access to opportunities, resources, and valuable information (Burt 2001). Such limitation in access to resources and information constitute a tough barrier for teenagers living in ghettos and attending segregated schools. 2.2) Social Capital Effects on Attainment Norms The second most important way in which social capital affects postsecondary education access is through differences in norms that are supportive of educational

attainment. General consensus about the desirability of educational attainment may exist. However, different social groups differently value behaviors and attitudes that are instrumental for achieving educational success. To begin with, norms are externally defined and usually imposed by authority figures (e.g. parents, teachers, priests) (Coleman, 1988; 1990). Hence, an individual’s exposure to norms supportive of educational attainment will depend on her/his relationships with persons holding such norm-enforcing roles. Youngsters who live in somewhat dysfunctional households, attend schools staffed by uncaring teachers, and/or lack other links to nurturing norms-enforcing groups may be greatly disadvantaged. Secondly, norms enforcement largely depends on the existence of a trustworthy system of rewards and sanctions. If perception of inequity or unfairness—either a real issue or a misperception—exists, there is little incentive to follow the norms. Indeed, youngsters living in high-poverty neighborhoods, who perceive inequity and unfairness ruling their lives, have few incentives to behave in ways conducive to educational attainment. Furthermore, the confluence of several reinforcing rules may facilitate individuals’ adoption of educational attainment norms. Conversely, conflicting norms may hamper the adoption of such norms. For instance, conflicting norms may affect educational outcomes of a teenager who aspires to a college education, but has several friends who belong to gangs. Failing to account for conflicting norms may explain why different

groups of students experience unequal outcomes even though they are exposed to the same formally enacted norms. Finally, Coleman (1988; 1990) argues that positive norms are better enforced when an individual lives in a community with networks closure—that is where people know each other well, share values and religious beliefs, and so on. Lin (2001a) also refers to strong social networks’ effects on reinforcing identity and enhancing group recognition, both factors that may boost positive educational attainment norms. However, both Lin (2001c) and Burt (2001) show that closure seems detrimental for accessing new information. Since limited access to new information is damaging for those in the bottom of the social hierarchy, policies that emphasize networks closure— e.g. high parental control, and zero-tolerance schools—may not be beneficial for students coming from low socioeconomic status families. As Lin (2001a) claims, mobility opportunities are better achieved by gaining access to other groups’ resources, and moving away from one’s original group. Thus, for low income students to gain access to the resources required to stay within the educational system, they need to use contacts beyond their strong ties (Burt 1998; Lin, 1999; Portes, 1998; Stanton-Salazar & Dornbusch, 1995; Stanton-Salazar, 1997; Stanton-Salazar and Urso Spina, 2000). 2.3) Social Capital Effects on Support for Navigating the System The third way in which social capital affects postsecondary education access is by the provision of support for navigating the educational system. The pipeline to postsecondary education enrollment involves a series of interwoven decisions that may not be obvious for teenagers who lack beneficial social capital. Those sequential decisions include taking college preparatory courses in high school, cultivating

relationships that will provide recommendations letters later on, preparing for standardized tests, assessing one’s attitudes in order to chose a future career, taking college entrance examinations, searching for appropriate institutions, writing personal statements, filling admission applications, applying for financial aid, etcetera. Students from different socioeconomic backgrounds have very unequal support networks for navigating this process, and taking all the required steps in a timely way. Differences in availability of guidance along all those steps may a factor that influences inequalities in access to postsecondary education. Social networks connecting low income students with either institutional counselors or informal mentors are instrumental for guiding these students in navigating such processes.

IV. Research on PSE Access Using Social Capital Theory There is a sizeable body of research that explores social capital effects on inequalities in postsecondary education access (Baron, Field et al. 2000; Dika and Singh 2002; Field 2003). In this section, I review a selection of the empirical research on postsecondary education access using social capital constructs. The main aspects I analyze are conceptualization, modeling approaches, and findings related to social capital effects. I will focus on the extent to which this empirical research postulates testable hypothesis for explaining how social networks provide relevant information, supportive academic attainment norms, and encouragement and support in navigating the educational system. Review Criteria The present review is grounded in the social capital theoretical framework synthesized in the previous section. This review does not attempt to be exhaustive, but to provide a critical panoramic of the research on the subject published in the United States within the last ten years. Selected articles meet the following criteria: 1) Focus on access to postsecondary education. 2) Published in the United States, within the last 10 years (An exception to this criterion is Coleman’s foundational article). 3) Reporting quantitative non-experimental studies based upon survey data. 4) Presenting original multivariate statistical analyses. Delimiting postsecondary education access outcome variable

For the purpose of this review, studies in postsecondary education access were delimited to those in which the outcome variable is one of the following: enrollment in any postsecondary institution, enrollment in a 4-year institution, enrollment in a 2-year institution, and number of years of schooling. Most commonly used outcome variable in postsecondary education access research are enrollment in any postsecondary institution and college attendance (Ainsworth, 2002; Duncan, 1994; Hofferth, Boisjoly & Duncan, 1998; Horn & Chen, 1998; Musoba, 2004; Perna, 2000; Plotnick & Hoffman, 1999; St. John, 2004; St. John, Hu, Simmons, Carter, & Weber, 2004; Terenzini et al., 2001; Zaff, Moore, Romano & Williams, 2003). Selecting these outcome variables implies that the researcher considers that the commonalities between different kinds of post-secondary education exceed their differences. However, several researchers maintain that different postsecondary education paths constitute in fact different outcomes. To account for institutional types, other outcome variables have been used. Enrollment in a 4-year college is widely used because most researchers tend to place a premium in this long-established type of postsecondary education (Berkner & Chávez, 1997; Horn & Chen, 1998; Horn & Núñez, 2000; Zaff, Moore, Romano & Williams, 2003). Research focused on community college enrollment is less common, but it is growing; hence, enrollment in 2-year institutions is becoming another typical outcome variable for studying access (Berkner and Chávez 1997). All outcome variables mentioned above are dichotomous. However, some researchers prefer to use continuous variables as an outcome. In such case, years of schooling is the preferred choice (Duncan, 1994; Hofferth, Boisjoly & Duncan, 1998). Search Approach

The starting point of this review search process was examining the NELS:88 Annotated Bibliography published by NCES (2005). That annotated bibliography was derived from computer searches of online bibliographic databases such as Dissertation Abstracts, ERIC, Psychological Abstracts, and Sociological Abstracts. From this annotated bibliography, 21 journal articles, conference papers, and policy research reports published between 1994 and 2004 were selected. No dissertation was selected, but a subsequent search in online databases by the dissertations’ authors’ names led to two more works (a journal article, and a conference paper). An additional work was identified by tracking down the list of primary studies in social capital and educational attainment reviewed by Dika and Singh (2002). A final online databases search (ERIC, JSTOR, ProQuest, Sociological Abstracts, and Psychological Abstracts) using the keywords “social capital” and enrollment did not generate additional relevant articles. Ten other works were found through references in the initially selected articles.

1) Conceptualizing Social Capital in Postsecondary Education Literature Most of the empirical research linking social capital to educational outcomes is grounded in Coleman’s (1988) work. The majority of the reviewed works refer to Coleman’s definition of social capital as resources embedded in the family social networks that facilitate children’s educational attainment (Carbonaro, 1999; Hofferth, Boisjoly, & Duncan, 1998; Stanton-Salazar & Urso Spina, 2000; Yan, 1999). Although Coleman (1990) postulates that social capital is one among several resources available in order to pursue rational goals, most of the reviewed studies fail to encompass social capital effects within a comprehensive theory of rational action.

In his 1988 study, Coleman tested hypotheses about effects on educational outcomes from social capital variables, such as networks closure and intergenerational closure. He defined networks closure as the ability of several actors to combine their resources to monitor and either sanction or reward others’ behaviors. Intergenerational closure was operationalized as a measure of whether parents befriend their children’s friends’ parents. Coleman hypothesizes that such relationship would make it easier for them all to establish norms and to control behavior adjusted to such norms. Coleman also tested hypotheses about “intra-familial social capital”, which refers to quantity and quality of family interactions related to educational activities. Although Coleman acknowledges that “intra-familial social capital” should interact with parental education as well as with family income, his study does not control for such interactions. This pitfall weakens his claims that family structure and parental involvement affect educational outcomes. Nonetheless, such claims have been the most influential among Coleman’s (1988) reported findings. Indeed, many other researchers have frequently included similar variables in their models with mixed findings (Ainsworth, 2002; Cabrera, La Nasa, & Burkum, 2000; Horn & Chen, 1998; Horn & Núñez, 2000; Perna, 2000). Mullis, Rathge and Mullis (2003) assert that social capital can be measured by the quality and quantity of the networks that connect children and adolescents with the resources of their parents. According to them, a similar process allows children to gain access to schools’ and communities’ resources. Mullis and associates add parental expectations as a mediator between familial social capital and adolescents’ academic achievement. They also hypothesize that familial capital and community social capital

interact to shape children and adolescents’ behaviors and positive valuation of educational attainment. Sandra Hofferth—who worked with Coleman researching of Catholic school effects (Coleman and Hofferth, 1987)—has continued to advance empirical research in extra-familial social capital and educational attainment. Scholars in this stream of research claim that families holding tight social networks are better able to encourage children’s educational attainment (Hofferth, Boisjoly, & Duncan, 1998). They pose two reasons for explaining why that is the case: 1) linkages among families create a functional community with clear norms and effective sanctions (and rewards) that shape children’s achievement norms for supporting school success; and 2) parental networks may provide information on postsecondary education opportunities. Hofferth and associates also emphasize that family social networks interact with family structural characteristics, such as income, parental education and household composition. They hypothesize that being connected to people who are not able to reciprocate is an additional hindrance to low income students’ educational attainment. According to their analyses, a network linking mostly to low-resource contacts may be a burden rather than an asset. Such network would generate obligations but not retributions, taking away resources that would otherwise be invested in children’s education. Laura Perna (2000) presents a more complete conceptual framework, drawing from both Coleman’s (1988) and Bourdieu’s (1986; and Bourdieu and Passeron, 1977). According to Perna (2000), social and cultural capitals are both sources of resources that may be invested to facilitate upward mobility. She gives emphasis to information-

sharing as the way in which social capital is invested and its profits are accrued. Perna also builds upon McDonough’s (1997) in-depth qualitative study on Californian female high school students’ college choice processes. McDonough (1997) attempted to expand the understanding of the college choice process by adding together economic and psychological aspects of these girls’ decision-making processes, as well as the cultural facets shaped by colleges, high schools, parents, friends, and mass media. In her extended econometric model of college choice, Perna (2000) includes constructs that reflect differences in expectations, preferences, tastes, and certainty about higher education investment decisions based upon habitus and social capital. However, Perna does not discuss in detail how the field and the habitus developed by her African American sample interact to bind their educational choices. Karen (2002) also incorporates the notion of habitus into his conceptual framework, which for the most part is grounded in status attainment theory. Karen examines how high school graduates become linked to colleges with particular selectivity levels. Following Bourdieu and Passeron (1977), Karen points out that institutionalized processes match an individual educational destination with her/his social origin. The development of realistic expectations of educational success and the mapping of potentially reachable destinations is shaped by the habitus. Thus, students beginning to think about their postsecondary education opportunities will adjust their aspirations to those usually ascribed to their class. In spite of the sociological roots of the notion of habitus as the link between social class and rules of action, Karen portrays this process in a rather individualistic way, without undertaking a thorough analysis of how institutional and economic barriers interact with students’ dispositions.

Despite their limitations, both Perna (2000) and Karen (2002) make contributions in integrating cultural and social capital notions into rational choice and status attainment conceptual frameworks. Perna (2000) hypothesizes that individuals who lack the required cultural capital may lower their educational aspirations or self-select out of particular situations (e.g., do not apply to a 4-year institution) because they do not know the particular cultural norms, or they may receive fewer rewards for their educational investment (e.g. lower grades, failure to obtain recommendation letters from influential acquaintances). Likewise, Karen (2002) hypothesizes that students whose parents do not have a college education would be more likely to apply for and to enroll in less selective or open admissions institutions, whereas students whose parents have advanced degrees would be more likely to aspire, apply to and attend to highly selective institutions. Paulsen and St. John (2002) examine the influence of school and family environments

on

postsecondary

enrollment

choice

and

the

importance

of

acknowledging differences in financial means, mobility, opportunities and choice sets. Inserting their research approach within the social reproduction theory tradition, Paulsen and St. John propose a nexus model that links financial factors to students' choices across social classes. The authors stress the notion of habitus, but they only barely refer to the notion of field (Bourdieu, 1977a, 1977b). A greater focus on the notion of field would have been appropriate since their research focuses on inequalities in resources and opportunities due to social class. Including the notion of field, as well as specific measures of social capital would have enhanced the nexus model’s capability

for investigating structural features that foster class inequalities regarding resources and opportunities. Neighborhood effects models provide another conceptual approach for analyzing social capital effects on educational attainment (Ainsworth, 2002; Duncan, 1994; Sampson, Morenoff, & Earls, 1999; Sampson & Raudenbush, 1999). Wilson (1987; 1996)

describes

five

interrelated

mechanisms

through

which

neighborhood

characteristics affect educational achievement; these are collective socialization, social control, social capital, differential occupational opportunity, and institutional (i.e., school) characteristics. Wilson (1996) argues that individuals’ involvement in social networks may either facilitate or prevent access to resources that are instrumental for successful transition into adulthood through educational opportunities and links to occupations. According to Wilson, the quality and amount of resources embedded in social networks within a given community influence educational outcomes and occupational opportunities. Wilson claims that children who live in affluent neighborhoods are more likely to be connected to valuable social networks and to have access to adults who can provide resources, information, and opportunities that may be educationally advantageous, while children living in impoverished neighborhoods are less likely to connect to valuable networks and resources. Ainsworth (2002) further develops this conceptualization by analyzing effects of neighborhood structural and social capital variables such as intergenerational closure and peer networks in academic achievement. Ainsworth (2002) hypothesizes that ties to individuals with few resources are detrimental because those ties represent obligations rather than resources.

Duncan (1994) also examines neighborhood effects using the lenses of epidemic, collective socialization, institutional resources, competition for resources, and relative deprivation theories. Among the five, collective socialization and epidemic theories are relevant for this review, because of their relatedness to social capital theory. Collective socialization emphasizes the beneficial effects of positive role models and well-monitored neighborhoods, whereas epidemic theory calls attention to peerinduced disruptive behavior. Collective socialization and epidemic theories may be illuminating only one side of the picture each. Alternatively, social capital illuminates both simultaneously by assisting a scholar in differentiating between relationships that provide resources for maintaining current status or elevating it, and relationships that subtract resources from a person without possibility of recompense.

2) Modeling Approaches in Researching Social Capital Effects on PSE Access This sub-section examines the measures and modeling approaches that have been used for assessing social capital effects on postsecondary education access. I use a selection of the reviewed empirical works as illustrations for discussing methodological developments and shortcomings. In his seminal work, Coleman (1988) uses the High School and Beyond 1980 data. HS&B 1980 collects detailed socioeconomic status and personal characteristics data from a stratified random sample of 14,799 sophomores in United States high schools. Follow-up interviews were conducted in 1982, 1984, 1986, and 1992 in order to obtain information about students’ cognitive test scores, college enrollment and course taking, the type of institution attended, and labor force participation (Source: http://nces.ed.gov). There are not direct social capital measures in that dataset, so

Coleman drew on some proximal variables for bringing to light the theoretical constructs in his framework, especially the measures for closure, which is hypothesized to enhance the social capital available to the family. Number of residential moves between 5th grade and 10th grade was first used by Coleman (1988) as a proxy for lack of intergenerational closure. His rationale was that families that move often would not have had opportunity for building relationships with other people in the community. Residential moves during school years (Coleman, 1988; Furstenberg & Hughes, 1995; Hofferth et al., 1998) or residential stability (Ainsworth, 2002) have continued to be used as proxies for community closure. Another proxy for intergenerational closure used by Coleman (1988) is enrollment in a Catholic high school2. Coleman makes a case for considering religiously based high schools as a basis for multiplex relationships—individuals are linked in several ways according to their different social roles (e.g. neighbors, colleagues, friends, fellows in faith, and so on). The existence of such multiplex relationship among most students attending religious schools—not only Catholic ones—contributes to increasing available social capital for all students regardless their individual religious observance. Intergenerational closure indicators as key measures of a student’s networks have continued to be used in postsecondary education access research. Among the most commonly included of these indicators in the reviewed literature are frequency of student contact with adults other than parents and time a student spend with adults other than parents (Ainsworth, 2002; Carbonaro, 1999; Furstenberg & Hughes, 1995). 2

Coleman (1988) alleges that his analysis focus on Catholic schools due to sample size restrictions, but

his hypothesis held for any denomination including non-Christian religious schools.

Perna (2000) includes high school control as a proxy for achievement expectations, which relates somewhat to Coleman’s (1988) conceptualization of school closure as a mechanism for infusing academic achievement norms. Hagy and Staniec (2002) include the percentage of the student’s high school class enrolled in collegepreparatory courses as a proxy for environment fostering academic attainment. A wider range of closure measures are available in the National Education Longitudinal Study – 8th grade 1988 cohort (NELS:88) and its follow-ups. NELS students surveys comprise questions about an extensive range of topics: school, work, and home experiences; educational resources and support; relationships with parents and peers; neighborhood characteristics; educational and occupational aspirations; other student perceptions; risky behaviors; and extracurricular activities (Source: http://nces.ed.gov). NELS:88 presents several advantages for studying social capital effects on postsecondary education access. First, it is nationally representative sample. Second, it collects data longitudinally from 8th grade to 8 years after programmed high school graduation; therefore, it makes feasible the examination of temporal and causal relationships among variables. Third, it includes students, parents, teachers, and school administrator data, allowing triangulation of data sources. Fourth, it can be merged with other datasets, such as the Integrated Postsecondary Education Data System (IPEDS), Federal Student Aid application records (FAFSA), and the U.S. Census Data, allowing inclusion into the analytical models of variables accounting for financial aid and community characteristics. Nonetheless, using NELS:88 also has shortcomings. The more important one of those shortcomings is that some variables conceptually required when studying social

capital effects are not included in the dataset. Thus, using NELS entails adjusting theoretical constructs to available variables and measurement scales. Extensive data transformation is necessary for creating variables that correspond to the conceptual framework. Furthermore, proxy variables may be required in order to substitute variables suggested by the theory that are not readily available in NELS or other public datasets. There is also a problem with a high proportion (20% to 30%) of missing data for some relevant variables, such as typical students’ networks variables, which encompass those measuring interactions with and influences from peers, high statusindividuals, adult models and mentors (Ainsworth, 2002; Horn & Chen, 1998; Perna, 2000; Zaff et al., 2003). The number or proportion of the student’s close friends who either dropped out of high school (Ainsworth, 2002) or plan to attend college (Horn and Chen 1998), which is hypothesized to affect a student’s educational outcomes. Perna’s (2000) model incorporates a composite of importance of education among friends that attempt to measure effects of students’ peer networks. Measures of extracurricular activities (Mullis, et al., 2003; Zaff, et al., 2003) and participation in community organizations (Mullis, Rathge et al. 2003) are also among this group of variables. Reliability tests indicate that participation in community organizations is a stronger indicator of students’ social networks than participation in school based extracurricular activities, probably because the latter interact with school structural features and climate. Other datasets provide better-fitting measures for closure, and family support networks. For instance, the Panel Study of Income Dynamics includes variables that capture the family support networks construct by measuring whether a person has

offered help in the way of time or money to a friend or neighbor and whether a person has received help in the way of time or money from a friend or neighbor (Duncan, 1994; Hofferth et al., 1998; Plotnick & Hoffman, 1999). Measures of support network to which the family can turn for assistance in case of disruptive life events (e.g. unemployment, divorce, severe illness, family member’ death, childbirth, etc.) seem to be reliable indicators of family social capital (Hofferth, Boisjoly et al. 1998; Mullis, Rathge et al. 2003). In contrast, a study using structural equation modeling indicates that parental involvement in school activities is a weak indicator of family networks (Mullis, Rathge et al. 2003). Nonetheless, parental involvement indices are pervasive in the postsecondary education access literature (Ainsworth, 2002; Cabrera et al., 2000; Cabrera et al., 2001; Furstenberg & Hughes, 1995; Glick & White, 2004; Horn & Chen, 1998; Mullis et al., 2003; Perna, 2000; Tierney & Jun, 2001; Yan, 1999; Zaff et al., 2003). Indicators of availability of school counseling or college choice advising services, as well as information about financial aid (Cabrera, La Nasa et al. 2000; Cabrera, La Nasa et al. 2001) are widely used. A commonly used measure of this construct is assistance from school personnel with college admissions requirements (e.g recommendation letters and help preparing personal statements) (Cabrera et al., 2000; Cabrera et al., 2001; Perna, 2000). Perna (2000) also includes encouraging interactions with a favorite teacher or school counselor who suggests that the student should enroll in college. Postsecondary

education

researchers

examining

social

capital

effects

occasionally include in their models variables related to institutional networks, such as outreach and/or college preparation programs (Horn & Chen, 1998; Musoba, 2004;

Tierney & Jun, 2001; Trent, Gong, & Owens-Nicholson, 2004). Other school resources, as well as high school structural features and climate, are also often included in models for examining social capital effects in postsecondary education access (Perna, 2000). School structural features include percent minority students (Musoba, 2004; St. John et al., 2004) and percent poverty—usually measured by the proxy percentage of students entitled to free lunch (Ainsworth, 2002; Musoba, 2004; St. John et al., 2004). School climate is usually a factor or index composed by indicators such as student and teacher morale, students’ priorities on learning and other activities, school behavior management and discipline, classroom control, teachers’ ability to motivate students and so on (Ainsworth, 2002). In contrast, community resources are seldom included in explanatory models for postsecondary education access, probably because of lack of proper measures. Exceptions are those studies that test hypotheses on how collective socialization and differentials in institutional resources availability affect educational attainment (Ainsworth, 2002; Duncan, 1994). Community demographics variables are typically included in models of neighborhood effects; such variables include percentage of low income families, percentage of high income families, percentage of persons from racially disadvantaged groups, percentage of single mother households, unemployment rate, and high status adult residents, usually a standardized composite of proportion of college graduates among adults over 24 years old and the proportion of employed persons with professional or managerial occupations (Ainsworth, 2002; Duncan, 1994; Hofferth et al., 1998; Perna, 2000).

Lack of proper measures sometimes attenuate the explanatory power of wellconceptualized models for examining social capital effects in postsecondary education access. For instance, Perna (2000) hypothesizes that the most important way in which social capital influences expectations, preferences, and evaluation of the return from postsecondary education is through the provision of relevant information. However, Perna’s model does not include direct measures of information, but it incorporates proxies such as percentage of the high school graduates enrolling in college and high school desegregation. These may be appropriate proxies for students´ peer networks, but they are rather indirect measures of available information. The study also includes region and location as proxies for information availability, but these measures are even more indirect and less plausible 2.1) Analytical methods for examining social capital effects on PSE education access Researchers in social capital effects on postsecondary education access do not yet agreed about analytical methods best suited for this stream of research. Relational and spatial statistics (e.g. cluster analysis) that have being used in other fields have not entered into postsecondary education access research (Breiger, Carley, and Pattison, 2001). Several types of regression models are found in the reviewed postsecondary education access literature. Logistic regression is the most widely used among all available analytical methods (Adelman, 1999; Ainsworth, 2002; Coleman, 1988; Zaff, Moore, Romano & Williams, 2003). Indeed, most studies deal with dichotomous outcomes, and logistic regression is the most appropriate statistical technique for analyzing such variables (Menard, 1995). A number of researchers claim that

enrollment in a 4-year institution is the most appropriate outcome variable for examining postsecondary education access, since the predominant social expectation for educational attainment is a bachelor degree, (Horn & Chen, 1998; Horn & Núñez, 2000; Perna, 2002; Zaff at al., 2003). For instance, Perna (2002) justifies focusing on enrollment in a 4-year institution because the long-term economic benefits for those who complete a bachelor’s degree are greater than for those who complete an associate degree. Perna’s model of college enrollment decisions expands the traditional econometric model by including measures of social and cultural capital as proxies for expectations, preferences and tastes, and uncertainty. The model stipulates the decision to enroll in a 4-year institution as a function of direct costs, labor market opportunities, future benefits, financial resources, academic ability, and social and cultural capital. Duncan’s (1994) study includes both logistic regression models and OLS regression models. Using a similar set of family background and neighborhood predictor variables, Duncan examines two dichotomous outcomes, high school completion and attending postsecondary education, as well as a continuous outcome, years of schooling at mid-20s. The major limitation of Duncan’s study is that his selected dataset (Panel Study of Income Dynamics), while rich in neighborhood and family measures, lacks information about schools and academic experiences, besides the attainment milestones used as outcomes. Hoffert, Boisjoly and Duncan (1998) also use the Panel Study of Income Dynamics to examine social capital effects on years of schooling measured at mid-20s, high school completion, and attending postsecondary education.

All the neighborhood and family measures included in Duncan (1994) are included here also, but a set of direct measures of familial social networks were added. For those studies in which the outcome variable is continuous (e.g. years of schooling), OLS regression is an appropriate analytical model (Duncan, 1994; Hofferth, Boisjoly, & Duncan, 1998; Trent, Gong, & Owens-Nicholson, 2004). However, OLS is inappropriate for dichotomous outcomes such as enrollment at any post-secondary institution, enrollment in 2-year institutions, and enrollment in 4-year institutions. The inappropriate use of OLS regression compromises the validity of the findings reported by Berkner and Chávez (1997) and Horn and Núñez (2000). These works have also been criticized for serious methodological pitfalls related to selection bias, endogeneity, omitted variables bias, and collinearity (Becker, 2004; D. Heller, 2004). In his study of differences among African Americans and other ethnic groups in enrollment by college selectivity, Karen (2002) also uses an OLS model with variables entered in two blocks, broadly delimited as personal background and academics. Variables somewhat related to social capital are linked to school experiences. In this study, selectivity is measured by the average SAT score of students entering in a given institution. Researchers interested in examining both enrollment in 2-year institutions and enrollment in 4-year institutions usually specify separate models for each outcome and run two sets of logistic regressions (Cabrera, La Nasa et al. 2000; Cabrera, La Nasa et al. 2001). Nonetheless, it can be argued that each one of these outcomes competes with the other, and they must be in the same equation in order to allow for proper estimation. Therefore, selecting an analytical technique that allows for distinct outcome

categories in the dependent variable would be appropriated. Multinomial and nested logit models are proper analytical techniques for estimating choice models in which the outcome variable is categorical or only partially ordered. In the reviewed literature, there are not studies using a nested logit model for analyzing postsecondary education enrollment, but there are a few using multinomial models (MNLM). The MNLM is a nonlinear regression, and it allows effects of the independent variables to differ for each outcome, by estimating simultaneously binary logits for all possible comparisons among the outcome categories (Long, 1997; Raundenbush and Bryk, 2002). Glick and White (2004) use a multinomial model to predict the likelihood of participating in postsecondary education versus other activities two years later. Their model contrasts postsecondary attendance (i.e., enrolled in a 2- or 4-year institution in 1994), with finishing high school (i.e., a high school diploma attained by 1994), and not completing secondary education (i.e., less than high school education attained by 1994). Similarly, Hagy and Staniec (2002) use a multinomial model to estimate choices from among five postsecondary enrollment alternatives: 4-year public institutions, 4-year private institutions, less-than-four-year public schools (2-year community/junior colleges), less-than-four-year private institutions (mostly proprietary vocational/technical schools), or non-enrollment in any postsecondary institution. Given that students are nested within institutions (e.g. high schools and colleges) and enduring groups (e.g. families and communities), and their postsecondary enrollment decisions are not independent of their relationship with such groups, analytical methods that account for multilevel structure in the data are advisable (Raundenbush and Birk, 2002). However, only one work within the reviewed literature

uses logistic multilevel regression (Musoba 2004). Two other works address the issue of multilevel data by using fixed-effects OLS regression (St. John, Chung, Musoba, & Simmons, 2004) and fixed-effects logit (Plotnick and Hoffman 1999). However, the majority of the remaining works do not even acknowledge the multilevel structure of their data. Another methodological issue in postsecondary education access is that some predictor variables may have time-varying effects (DesJardins 2003); however, there were no studies using event history models within the reviewed literature.

3) Findings on social capital effects on postsecondary education access In this sub-section, I will examine empirical research findings related to social capital influence on the ability to take advantage of resources and the effects of group norms on individuals’ educational decisions. In addition, I will point out conceptual or measurement issues that may undermine or obscure the interpretation of these findings. In general, intergenerational closure—opportunities for interacting with adults other than close relatives—has positive effect on educational attainment (Carbonaro, 1999; Coleman, 1988; Furstenberg & Hughes, 1995). Peer behavior and plans have been found to be positively related to enrollment in a 4-year institution (Ainsworth, 2002; Cabrera et al., 2000; Horn & Chen, 1998; Perna, 2000; Zaff et al., 2003). Several researchers have found that residential moves during middle school or high school reduces the likelihood of pursuing postsecondary education, specially for low income students (Furstenberg and Hughes 1995; Hofferth, Boisjoly et al. 1998; Glick and White 2004). This finding remains consistent in studies controlling for variables that may be confounding, such as parents occupations, immigration status, and language spoken at home (Glick and White 2004). This finding is also consistent

with Coleman’s (1988) hypothesis of residential moving being disruptive of family networks, having a negative effect on social capital available to children, and hampering educational attainment. By and large, a strong support network positively affects postsecondary education enrollment (Furstenberg and Hughes 1995), but the effects of giving and receiving help appear to be moderated by income (Hofferth, Boisjoly et al. 1998). High income children show greater likelihood of attending college, as a benefit from their parents’ links to social networks. On the contrary, low income children obtain fewer years of schooling as a result of their parents having many social networks links. These findings may be revealing of networks constraints, such as low hierarchy contacts or excessive density, which according to Burt (1997) weaken the positive effects of belonging to a network. Results regarding effects of parental involvement are also different for different groups of students (Yan 1999). Perna (2000) finds positive effects of parental involvement for White students, but not for Latinos and African-Americans. Different results may be due to variability of the measures included in composites and factors that attempt to capture this construct, which in some cases may act in different directions and cancel each other out. For instance, Glick and White (2004) found that when parental involvement was motivated by behavioral concerns the likelihood of college enrollment dropped, while involvement unrelated to behavioral concerns significantly

increased

the

likelihood

of

postsecondary

education

enrollment.

Nonetheless, the interaction between parental involvement and inequality of resources due to income or race cannot be ruled out.

Findings regarding effects of extracurricular activities are mixed, too. Inconsistent results may be due to differences in the procedures for aggregating data. Apparently, some activities that foster leadership and intellectual development, such as participation in student government and school newspaper, have greater positive influence in postsecondary education outcomes than less intellectually demanding activities (Karen 2002). Indications of curvilinear effects from some activities, such as participation in varsity sports, also appear in empirical research (Zaff at al., 2003). Participation in outreach or encouragement programs have also been found positively related to postsecondary education enrollment (Horn & Chen, 1998; Musoba, 2004; Tierney & Jun, 2001; Trent, Gong, & Owens-Nicholson, 2004), but specific program features that make the difference have not been clearly identified. In order to do so, it would be necessary to compare effects of programs having different features regarding support for academic preparation, encouragement, life skills development, access to information, availability of financial aid, and timing. A great deal of evidence on positive effects of access to information has been found. Lack of school help with college application reduces the likelihood of postsecondary education for Latinos and Whites (Perna, 2002). Gathering information about financial aid and talking to individuals about aid increase the odds of enrolling in any postsecondary education (Horn and Chen 1998). Getting help with preparing for entrance exams and the college application process increased the odds of enrolling in a 4-year institution for all students (Horn and Chen 1998). The use of test preparation services increases the likelihood of postsecondary education for Whites and African-

Americans (Perna, 2002) but not for Latinos. However, interactions between information and family income should not be ruled out. Encouragement to pursue a college education from teachers and school counselors increases the likelihood of enrolling in a 4-year institution for Whites, but there are no significant effects for Latinos and African-Americans (Perna 2000). Percentage of previous cohorts of high school graduates attending a 4-year institution increases the likelihood of college enrollment for students of all ethnic groups, which may be because of role models effects, although it certainly can be a result of differences in academic quality. In any case, students who reported that most of their high school friends have plans for enrolling in a 4-year institution were far more likely to enroll in a 4-year institution themselves (Horn and Chen 1998). Overall, high school climate has been found to have effects on post-secondary education enrollment (Ainsworth, 2002; Zaff, et al., 2003). However, interactions between school resources and student personal background apparently exist. In general, attending a school with a high poverty level decrease the odds of college enrollment for all students (St. John et al., 2004), but effects are substantially larger for White than for Black students (Musoba 2004). Neighborhood characteristics also have different effects depending on gender, race, and family income. Living in a low income neighborhood has negative effect on years of schooling for White females, while living in a high income neighborhood has a positive effect on years of schooling for all groups but Black males, which only benefit when there are also a high proportion of Black high status individuals in the community

(Duncan 1994). Percentage of women headed households and percentage of working women negatively affect Black females (Duncan 1994). Overall, the accumulated evidence supports hypotheses of social capital effects on inequalities in postsecondary education access. Nevertheless, addressing measurement problems and developing proper scales still remain as unsolved issues. Another ongoing task is the refinement of analytical approaches in order to model properly causal relationships among social capital variables and postsecondary education access inequalities.

V. Limitations in Using Social Capital Theory in PSE Access Research After reviewing both the theoretical literature on social capital and the empirical research literature that uses social capital to examine access to postsecondary education, I identified several issues that need to be addressed in order to move this stream of research forward. In this section, I will summarize three major conceptual and methodological limitations found in the reviewed postsecondary education access literature. These three major limitations found are the following: 1) conceptual vagueness; 2) weaknesses in testable postulates; and 3) lack of reliable measures. In addition to summarizing them, I will propose avenues for overcoming such limitations.

1) Avoiding Social Capital Conceptual Tautologies In the empirical research literature using social capital for examining access to postsecondary education, the foremost limitation is reliance on vague definitions. Such vagueness yields to quasi-tautological theoretical reasoning, and circular interpretations of social capital causes and effects. The problem of circular definitions originates by a widespread misunderstanding of the following Coleman’s (1990) statement: Social capital is defined by its function. It is not a single entity, but a variety of different entities having two characteristics in common: They all consist of some aspect of a social structure and they facilitate certain actions of individuals who are within the structure. Like other forms of capital, social capital is productive, making possible the achievement of certain ends that would not be attainable in its absence (pg. 302).

The misunderstanding of this sentence has led some researchers to pose social capital as an immeasurable factor that can be analyzed only by observing its effects. Under such stipulation, social capital could not be operationally defined and measured; therefore, this construct could not be incorporated into predictive models, and it would be unfeasible to design related public policy interventions. Actually, Coleman postulates that social capital is a set of assets or resources that an individual reaches through her/his ties to networks with other individuals and uses instrumentally for achieving her/his rational goals. Indeed, Coleman is asserting that the essential quality of social capital is its instrumental efficacy for enabling rational action in a given context. Understanding the instrumental nature of the concept, a scholar should define the construct of social capital and represents its dimensions within a general rational action theory (Coleman 1988, 1990; Lin, 2001b; Prakash & Selle, 2004)

2) Developing Robust Testable Social Capital Postulates In the postsecondary education access research using social capital theory, the second most important limitation is the indistinctness of empirical postulates. At the current stage of the concept development, verifying or even falsifying social capital theory postulates remains problematical, because such postulates are often vague or illdefined in terms of constitutive dimensions as well as regarding causality or directionality. Researchers often pose a very general hypothesis about social capital having positive effects on increasing access for low income and racially disadvantaged students, but they often fail to explain how social capital variables produce such effects.

Determining the missing links between social capital variables and their observable effects will set grounds for further advancing this stream of research without jeopardizing its heuristic value. Researchers need to break cause-effect circularity by separating social capital variables from their hypothesized effects. Doing that requires establishing controls for directionality and controlling for the occurrence of other factors that may account for both social capital and its hypothesized effects (Portes 1998; Lin 2001; Prakash and Selle 2004). The theoretical literature already offers avenues for postulating cause-effects relationships between social capital and postsecondary education access. My review identifies three important paths: 1) access to relevant information, 2) achievement and educational attainment norms; and, 3) support for navigating the system. Nonetheless, a great deal of empirical research fails to draw such paths from the general notion of social capital to individual’s postsecondary education outcomes. Hence, researchers should make the social capital causal chain clearly identifiable. Accomplishing that step would help to move social capital theory from a set of heuristic assumptions to a series of testable and falsifiable postulates (Baron, Field et al. 2000). A well developed set of testable postulates would facilitate empirical analyses for diagnosing incorrect assumptions, modifying postulates according to empirical findings, and building further extensions on what is not rejected (Coombs 1983).

3) Developing Reliable Social Capital Measures The third critical limitation found is the lack of valid and reliable measures of social capital (Baron, Field et al. 2000). Even the leading researchers in this stream— those who have overcome conceptual ambiguity and advanced toward testable cause-

effect postulates—stop short in their research endeavor because they have to run analyses on models that lack key variables or have to rely on inappropriate proxy measures. Lin (2001) explains that social capital operates within a three-phase process: investment, access and mobilization, and obtaining returns. Measures accounting for the investment and access phases should capture an individual’s social networks and the resources embedded in such networks. At present, national datasets for studying postsecondary education access do not map students’ social networks nor include specific measures of resources embedded in such networks. As I explain in the previous section, researchers have to rely on proxy variables such as parents’ dependability on support networks, family moves, school resources, school and/or neighborhood demographics, as well as students’ peer relationships. Mapping a student’s network location by identifying her/his strong and weak ties should be the initial step. Then, a researcher should assess the social position of the persons to which a student is tied and the relevant resources held by those persons. According to Lin (1999; 2001), proper measures of social capital embedded in a student’s network should include the range of resources, the upper hierarchy resource, the heterogeneity of resources, and the average or typical resource available in her/his networks. Relevant resources for postsecondary education access should include information on academic preparation, information on financial planning and aid, mentorship, support navigating the system, bridges to postsecondary education institutions, recommendations and certification of positive personal attributes. Currently available datasets only offer rough measures of available sources of information, weak

proxies for support navigating the system, and bridges to postsecondary education institutions. There is also little data on how a student uses available social capital resources for making choices related to her/his postsecondary education goals. It is necessary to develop data collection instruments for measuring the usage of social capital resources in the sequential choices toward postsecondary education enrollment. That sequence includes a students’ choice of high school coursework, preparation or lack of it for taking college entrance examinations, search for postsecondary education institutions, application for admission if required, application for financial aid, and the decision of whether or not to enroll (Hossler, Braxton, & Coopersmith, 1989). Specific measures for each step in the postsecondary education access process should be obtained.

VI. Directions for Further Research In this closing section, I will draw a framework for studying social capital effects in postsecondary access that should help to fill conceptual gaps in the literature. Finally, I will outline directions for measuring social capital variables and building comprehensive model for explaining postsecondary access.

1) Framework for Analyzing Social Capital Effects in PSE Access Following Coleman (1988, 1990) and Lin (2001a, 2001b), I will frame social capital within the wide-ranging rational action theory. I will borrow from Lin (2001a) to define social capital as the accumulation of resources embedded in an individual’s social network that s/he can access and mobilize in purposive actions with the expectation of obtaining some utility or return on investment. In this case, the return expected is increasing opportunities for obtaining access to postsecondary education. My proposed framework for studying postsecondary education access will place social capital as a crucial intervening variable affecting both academic preparation and financial resources for affording enrollment. Figure 2 shows a Comprehensive Model for Studying Educational Outcomes. The model controls for personal background characteristics such as gender, ethnicity, parental education, and family income. It also includes controlling variables for high school structural features, as well as academic offerings, teaching quality, and counseling services. Grounded in previous research, the model incorporates three sets of explanatory variables: academic preparation, finances, and social capital variables.

Comprehensive Model for Educational Outcomes School Structural Features Instructional Expenditures Control Localization ____________ Poverty Level Minority Proportion

Academic Preparation

School Capacity

GPA Test Scores High Standards Curriculum

Admission

Course Offerings Teaching Quality ________________ Counselling Educational Outcome

Community Tax Base Value of Education

Social Capital

HS Dropout Merit Aid

Post-secondary education * None * Job Preparation * 2 years college * 4 years college

Personal Background Gender Ethnicity _________________ Parental Education _________________ Income

Financial Resources Need (PSE Costs - Family Contribution) Financial Aid (Amount and Types)

Enrollment

The framework is consistent with research findings confirming that academic preparation indicators (e.g. college preparatory curriculum, high school GPA, and scores in college entrance examinations) have positive effects on postsecondary education access. It is also consistent with counterbalancing research findings indicating that the single most influential factor affecting enrollment is unmet need, when the combination of family contribution and available financial aid is insufficient for offsetting postsecondary education costs.

The literature shows a gap regarding effects from inequalities in access to social resources that do not belong to an individual and her/his family but are owned by other influential persons in the community. To fill such gap, this framework incorporates social capital as a critical intervening variable affecting postsecondary education enrollment. Figure 3 shows a path diagram representing Social Capital Effects on PSE Access. The path diagram shows three measurable dimensions of the underlying concept of social capital; those are information, bridges, and support. These dimensions are related to the three cause-effect paths from social capital to postsecondary education access identified in the literature review: 1) access to relevant information, 2) achievement and educational attainment norms, and 3) support for navigating the system. Relevant information is hypothesized to have direct effects on PSE enrollment through institutional choice and the application process. It also has indirect effects through planning academic preparation and applying for financial aid. Support is hypothesized to have direct effects on PSE enrollment through both admission and financial aid applications. Bridges are hypothesized to have direct effects on PSE enrollment through recommendations and certification of positive personal attributes, as well as contacts with decision-makers.

2) Directions for Future Research in Social Capital Effects on PSE Access I am devising a research agenda—grounded in this review—that will start with the development of instruments for measuring social capital dimensions. The first instrument to be developed would map a student’s social network location. Such an instrument will be an adaptation of the position generator technique (Lin & Dumin, 1986). In that technique, the respondent is asked to indicate if s/he knows anyone

having a position with identified valued resources (Lin 2001). Network maps—including strong and weak ties—as well as resource indices comprising composition, heterogeneity, and upper reachability can be built from the responses. Second to be developed is a questionnaire for measuring range of valued resources, upper hierarchy resources, heterogeneity of resources, and average or typical resource (Lin, 2001b). An additional questionnaire for obtaining measures of requests for using resources, help obtained, and returns from such help should also be developed as a part of this research agenda. Responses to these questionnaires can be used for estimating factors for each social capital dimension, information, support, and bridges. Both construct validity and reliability measures for both instruments should be analyzed using confirmatory factor analysis and direct product analysis (Bagozzi, 1994; Kline, 2005; Long, 1983). The goal is to develop instruments useful for collecting valid and reliable measures of the social capital constructs that could be used for modeling postsecondary education access and estimating inequalities among groups. Having valid and reliable measures would allow researchers to test hypotheses of social capital effects on postsecondary education using well-specified models. Thus, empirical studies would add evidence for supporting our current hypotheses on social capital effects on postsecondary access inequalities or undeniably falsify such hypotheses.

Social Capital Effects on PSE Access Path Diagram

Social Capital

1 1

Bridges

Information

1

Support

Academic Preparation Finances

GPA

1 Family Contribution

1

Merit Aid

Need Aid

1

Curriculum Tests

1

1

1

PSE Enrollment

REFERENCES

Adam, F. and B. Roncevic (2003). "Social Capital: Recent Debates and Trends." Social Science Information 42(2): 115-183. Adelman, C. (1999). Answers in the Tool Box: Academic Intensity, Attendance Patterns, and Bachelor's Degree Attainment. Jessup, MD, U.S. Department of Education. Baron, S., J. Field, et al. (2000). Social Capital. Critical Perspectives. New York, NY., Oxford University Press. Baum, S. and K. Payea (2004). The Benefits of Higher Education for Individuals and Society. Education Pays / Trends in Higher Education Series. C. E. E. Board. New York, College Entrance Examination Board. Becker, G. (1993). Human Capital. Chicago, The University of Chicago Press. Berkner, L. and L. Chávez (1997). Access to Postsecondary Education for the 1992 High School Graduates. Washington D.C., U.S. Department of Education. National Center for Education Statistics. Bourdieu, P. (1986). The Forms of Capital. Handbook of Theory and Research for the Sociology of Education. J. Richardson. Westport, CT, Greenwood Press: 241-258. Bureau of Labor Statistics (2004). "Education Pays." Retrieved August 12, 2004, from http://stats.bls.gov/emp/edupays.pdf. Burt, R. (2001). Structural Holes versus Network Closure as Social Capital. Social Capital. N. Lin, K. Cook and R. Burt. New York, Aldine de Gruyter. Cabrera, A., S. La Nasa, et al. (2000). Using National Databases To Study College Choice of Low SES Students. University Park, PA, Center for the Study of Higher Education - The Pennsylvania State University. Cabrera, A., S. La Nasa, et al. (2001). On the right path: The higher education story of one generation. University Park, PA, Center for the Study of Higher Education - The Pennsylvania State University. Carnevale, A. and R. Fry (2002). The Demographic Window of Opportunity. Condition of Access. Higher Education for Lower Income Students. D. Heller, American Council on Education.

Coleman, J. (1990). Foundations of Social Theory. Cambridge, Massachusetts, Harvard University Press. Coombs, C. (1983). Psychology and Mathematics. Ann Arbor, The University of Michigan Press. DesJardins, S. (2003). Event History Methods: Conceptual Issues and an Application to Student Departure from College. Higher Education: Handbook of Theory and Research. J. C. Smart, Kluber. XVIII: 421-472. Dika, S. L. and K. Singh (2002). "Applications of Social Capital in Educational Literature: A Critical Synthesis." Review of Educational Research 72(1): 31-69. Duncan, G. (1994). "Families and Neighbors as Sources of Disadvantage in the Schooling Decisions of White and Black Adolescents." American Journal of Education 103(1): 20-53. Field, J. (2003). Social Capital. New York, NY, Routledge. Fitzgerald, B. and J. Delaney (2002). Educational Opportunity in America. Condition of Access. Higher Education for Lower Income Students. D. Heller, American Council on Education. Furstenberg, F. and M. E. Hughes (1995). "Social Capital and Successful Development among At-Risk Youth." Journal of Marriage and the Family 57(3): 580-592. Gladieux, L. (2002). Federal Student Aid in Historical Perspective. Condition of Access. Higher Education for Lower Income Students. D. Heller, American Council on Education. Glick, J. and M. White (2004). "Postsecondary School Participation of Inmigrant Youth: The Role of Familial Resources and Educational Expectations." Social Science Research 33: 272-299. Granovetter, M. (2003). The Strength of Weak Ties. Networks in the Knowledge Economy. R. Cross, A. Parker and L. Sasson. New York, NY, Oxford University Press. Heller, D., Ed. (2001). The States and Public Higher Education Policy. Baltimore, The John Hopkins University Press. Heller, D. (2002). State Aid and Student Access. Condition of Access. Higher Education for Lower Income Students. D. Heller, American Council on Education. Hofferth, S., J. Boisjoly, et al. (1998). "Parents' Extrafamilial Resources and Children's School Attainment." Sociology of Education 71(3): 246-268.

Horn, L. and X. Chen (1998). Toward Resiliency: At-Risk Students Who Make It to College. Washington, D.C., Office of Institutional Research and Improvement: 52. Hossler, D., J. Schmit, et al. (1999). Going to College: How Social, Economic, and Educational Factors Influence the Decisions Students Make. Baltimore, MD, The Johns Hopkins University Press. Hu, S. and D. Hossler (2000). "Willingness To Pay And Preference For Private Institutions." Research in Higher Education 41(6). Karen, D. (2002). "Changes in Access to Higher Education in the United States:." Sociology of Education 75(3): 191-210. Kingston, P. W., R. Hubbard, et al. (2003). "Why education matters." Sociology of Education 76(1): 53-71. Langelett, G. (2002). "Human Capital: A Summary of the 20th Century Research." Journal of Education Finance v. 28(n. 1): 1-23. Lee, J. (2002). An Issue of Equity. Condition of Access. Higher Education for Lower Income Students. D. Heller, American Council on Education. Lin, N. (1999). "Social Networks and Status Attainment." Annual Review of Sociology 25: 467-487. Lin, N. (2001). Building a Network Theory of Social Capital. Social Capital. N. Lin, K. Cook and R. Burt. New York, Aldine de Gruyter. Lin, N. (2001). Social Capital. A Theory of Social Structure and Action. New York, Cambridge University Press. McDonough, P. M. (1997). Choosing Colleges. How Social Class and Schools Structure Opportunity. Albany, State University of New York Press. McPherson, M. and M. O. Schapiro (2002). Changing Patterns of Institutional Aid. Condition of Access. Higher Education for Lower Income Students. D. Heller, American Council on Education. Mullis, R., R. Rathge, et al. (2003). "Predictors of Academic Performance during Early Adolescence: A Contextual View." International Journal for Behavioral Development 27(6): 541-548. Musoba, G. D. (2004). Postsecondary Encouragement for Diverse Students: A Reexamination of the Twenty-first Century Scholars Program. Public Policy and College Access. E. P. St. John. New York. 19.

Paulsen, M. B. and E. P. St. John (2002). "Social Class and College Costs: Examining the Financial Nexus between College Choice and Persistence." Journal of Higher Education 73(2): 189-236. Perna, L. W. (2000). "Differences in the Decision to Attend College Among African Americans, Hispanics, and Whites." The Journal of Higher Education 71(2): 117-141. Plotnick, R. and S. Hoffman (1999). "The effect of neighborhood characteristics on young adult outcomes: Alternative estimates." Social Science Quarterly. Portes, A. (1998). "Social Capital: Its Origins and Applications in Modern Sociology." Annual Review of Sociology 24: 1-24. Prakash, S. and P. Selle (2004). Why Investigate Social Capital? Investigating Social Capital. S. Prakash and P. Selle. Thousand Oaks, CA, Sage Publications. Siisiäinen, M. (2000). Two Concepts of Social Capital: Bourdieu vs. Putnam. ISTR Fourth International Conference "The Third Sector: For What and for Whom?", Trinity College, Dublin, Ireland. St. John, E. P., Ed. (2004). Public Policy and College Access. Readings on Equal Education. New York, AMS Press. Terenzini, P., A. Cabrera, et al. (2001). Swimming against the Tide: The Poor in American Higher Education. New York, College Entrance Examination Board. Yan, W. (1999). "Successful African-american Students: The Role of Parental Involvement." The Journal of Negro Education 68(1): 5-22.