Bruce Fuller, Emily Hannum Schooling and Social Capital in Diverse Cultures, Volume 13 Research in Sociology of Education 2002

INTRODUCTION: SCAFFOLDS FOR ACHIEVEMENT?
INSTITUTIONAL FOUNDATIONS OF SOCIAL CAPITAL

Bruce Fuller and Emily Hannum
Reformers around the world - eager to raise children's learning and the effectiveness of public education - are tinkering again with the mechanics of schools. Disappointing and unequal levels of achievement are attributed to dynamics inside the walls of classrooms: teachers must be unenthused or short on skills; class sizes are too big; student testing is too infrequent; the old didactic methods of phonics must be restored, as teenage students forever enjoy the wonder of inventive spelling. The counter thesis is that families and communities matter more when it comes to raising children and motivating them to engage teachers and schools. Indeed, a student's social-class background and neighborhood attributes remain the strongest predictors of achievement, not school factors, as the late James S. Coleman first detailed with such controversy 35 years ago.
GRASPING SOCIAL CAPITAL, UNDERSTANDING ITS SUDDEN RISE

But what is it about a family's class position or the dynamics of a child's immediate community that so powerfully draws the learning curves of children?
Schooling and Social Capital in Diverse Cultures, Volume 13, pages 1-12. Copyright 2002 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0817-6
BRUCE FULLER AND EMILY HANNUM
And how are we to explain the unexpectedly high levels of achievement among youth who face high barriers to schooling, such as from working-class ChineseAmerican kids to youngsters of poor Latino parents? Prior structural representations of modern societies as made up of alldetermining layer cakes have given way to more localized illumination of parenting practices, peer norms, and ethnic-rooted values that may operate within small communities, at times insulated from the overall class structure or the conserving habits of the state. Social capital has risen in meteoric fashion, offering an umbrella that shelters a bundle of local dynamics, to help explain the mechanisms behind the curtain of class reproduction, as well as individual exceptionalism when it comes to youngsters who get ahead. While its meaning connotes forms of trust, reciprocity, and social expectations about membership in a group, the construct of social capital remains defined in rudimentary ways, rife with contradictions and conceptual dead ends. Nonetheless, notable scholars and political commentators have sanctified social capital in recent years as the door through which intellectual nirvana is to be found. Certainly, it holds promise for illuminating- and revisiting - local forces that shape children's engagement with and capacity to succeed inside schools. In sketching the initial elements of social capital, Professor Coleman, along with Barbara Schneider, empirically linked the family's social structure, expectations, and press on children to achieve with their actual school performance (Coleman, 1990; Schneider & Coleman, 1993). Pushing this line forward, each chapter in the present volume demonstrates how elements within the social capital frame do illuminate the local actors, norms, and expectations that forcefully shape children's success in school - from families to peer groups, from Vietnamese and Latino communities in North America, to social networks in Kenya or formal channels in Israel that bear on youngsters' performance inside the institution of schooling. The authors and commentators in this volume express widely varying allegiance to the social capital construct. Before digging deeper into how the elements of social capital account for unequal levels of school achievement, we press to define what is meant by social capital and why the construct itself has gained so much legitimacy in the space of a decade, undergoing rapid institutionalization in its own right. Well over a century ago, Karl Marx, then later Emile Durkheim, detailed the benefits of robust group ties and the alienating effects of losing one's sense of membership. And 15 years ago Pierre Bourdieu (1985, p. 249) already was writing of "profits which accrue from membership in a group" (see Portes, 1998, for review). So why has social capital, born again, taken hold now in the postindustrial era? And if social capital is so bankable, fungible, and powerful, how
Introduction: Scaffolds for Achievement?
can children or teachers tap into it to advance youngsters' engagement with the school community? Beyond the erosion of structural accounts of individual attainment, two additional forces help to explain the sudden ascendance of the social capital idea. First is the widespread concern felt in many corners of society that civility is on the decline, as well as the worry that the pursuit of a common good is being eclipsed by tribal and parochial interests. The culture wars continue, marked by divisive issues like abortion, immigration and language, gender and ethnic-linked rights, and what children are to learn inside schools. In his comparative analysis of trust within Western societies, Francis Fukuyama (1995, p. 308) issued this sweeping claim: "The moral communities that made up American civil society at mid-century, from the family to neighborhoods to churches to work places have been under assault, and a number of indicators suggest that the degree of general sociability has declined." Robert Putnam's (1993) more scholarly study of civil society and village associations in Italy led to the same claim about decline over time, and he argued that local civic organizations struggle to sustain deeper participation and shared social commitments, relative to the atomistic North American society. He even generated maps to display U.S. states that are allegedly high on social capital and those that fall to the bottom of the pack. This has sparked an empirical debate over how scholars can validly gauge social properties like trust, reciprocity, a collective press on children to follow particular pathways (Ladd, 1999). But the fact remains that some nations in the Western sphere are becoming more fractured within, separated by gross inequalities in income, ethnic diversity, views of women's roles, even contention about what comprises life and mankind's legitimacy in interfering with it. The question of civil society's health and moral integrity holds direct implications for the cultural organization of schooling, starting with how we think about the "proper ways" of raising children. Historically, the upbringing of kids has unfolded in private spaces filled with particular norms about what children should learn, the forms of adult authority and power that surround socialization, and how young children acquire knowledge and values inside and outside classrooms. Since the eighteenth century, a public set of authorities and political mechanisms have come to organize the agenda of schools, and children now spend more years of their lives in publicly-regulated places of learning. The very reproduction and strength of civil society depends on a broad consensus about the right agenda for public schools in modern society. So, when fissures begin to appear in civil society, with trust and a feeling of common cause dissolving like sand through one's fingers, the enterprise of schooling begins to fracture as well. Witness the growing interest in home
schooling, vouchers, privatizing the management of failing schools, and publicly financed charter schools (Fuller, 2000). This brings us to the remaining force that contributes to social capital's remarkable reification: the eagerness of influential scholars to craft a theory of individual decision making that is nested in local groups, drawing on skepticism of big institutions and disaffection with the ability of liberal-rights philosophy (its social-political, not its market, features) to combat the centrifugal forces that beset many western societies. We begin this short story with Coleman's radical recasting of the original notion of social capital advanced by economist Glenn Loury.
REINVENTING SOCIAL CAPITAL

Professor Loury, in his obscure 1977 book chapter, started with the macro structure of labor in the United States, what he saw as a highly institutionalized arrangement of workers segmented along lines of race and class. This segmentation, according to Loury, held damning implications for young blacks entering the labor force: "The social structure of this economy may exhibit both racial and income stratification... (including the form) of their informal social contacts" (Loury, 1977, pp. 157-158). Loury emphasized "the role played by group processes" (p. 156), citing the marxist economist, Samuel Bowles, in determining young adults' knowledge of and access to particular layers of the job market. Under Loury's analysis education played a key, and similarly stratified, role. Rather than viewing employers as discriminatory actors, the reproduction of inequality was rooted in how particular communities experience variable access to schooling, "by which he or she acquires marketable skills." For this reason "group differences in the supply of market-valued characteristics will tend to persist" (p. 154). Loury's aim was to link the ordered destinations of individual youth to a higher level of social or labor organization. One such local mechanism briefly mentioned at the end of the chapter: "It may be useful to employ a concept of social capital to represent the consequence of social position in facilitating acquisition of the standard human capital characteristics" (p. 176). For fellow economists, this was radical stuff. "The creation of a skilled workforce is a social process." And the individual's eventual stream of earnings to some extent "are accounted for by social forces outside an individual's control" (p. 175). Pierre Bourdieu and colleagues, of course, had elaborated a theory of cultural capital in the 1970s, placing schools in the pivotal position of sanctifying certain forms of high cultural knowledge, embedded within the highly stratified
arrangements of educational organizations found in European and North American societies (Bourdieu & Passerson, 1977). But Loury's formulation simply linked racially segmented communities to differential access to schools and to the availability of adult networks that hold quite different kinds of information and marketable skills, systematically reproducing advantage and disadvantage in the labor force. This was no longer a stream of thought in critical cultural studies. Loury's primitive conception of social capital, instead, was fungible in more ways than one, fitting nicely into North American scholars' pragmatic and materialist frame. And this construct, rooted in stratified institutions and local social networks, was now being championed by an economist. Professor Coleman, after the social capital idea percolated in academic circles for a decade, seized the moment. In 1988 he published a pioneering article entitled, "Social Capital in the Creation of Human Capital." Although the direct ancestry of Coleman's rendition of the idea is abundantly clear, Coleman did not cite Loury until publishing his massive Foundations of Social Theory two years later. Coleman was intent on developing a individual-focused theory of motivation and action: "If we begin with a theory of rational action, in which each actor has control over certain resources and interests and events, then social capital constitutes a particular kind of resource available to an actor" (S98). For Coleman, the individual is still trying to maximize his or her utility. Yet social capital also is out there in a social space, a resource "defined by its function" (S98) and somewhat situationally bounded that can be accumulated and appropriated. "Unlike other forms of capital, social capital inheres in the structure of relations between actors and among actors." In a series of examples Coleman went on to illuminate aspects of "obligation, expectations, and trustworthiness of structures" (SI02) that characterize small-scale markets, political groups, and families. On the one hand, Coleman postulated that the types and intensity of this "resource" called social capital were locally bounded: "Like physical capital and human capital, social capital is not completely fungible but may be specific to certain activities. A given form of social capital that is valuable in facilitating certain actions may be useless or even harmful for others" (S98). On the other hand, Coleman assumed that the form and effects of social capital were similar across large institutions. In his 1988 paper, for instance, he tried to show that comparatively high levels of social capital were reproduced and somehow allocated to children within Catholic schools, contributing to achievement significantly more than could be uncovered in public schools. Coleman's eagerness to find a fungible form of capital, sustained in one network and exercised in another, retains a philosophical focus on the sovereign individual. He ironically fused an economist's construction of capital
with a distinctly pragmatic conception of psychology and motivation. For the likes of William James and even John Dewey, "culture is an individual acquirement; it is the name for a set of products, practices, and perspectives of which individuals can avail themselves" (Menand, 2001, p. 407, emphasis added). The person's choices across a plurality of social options can only be considered if "individual human beings (are conceived), not as partial aspects of greater metaphysical wholes, but as complete in themselves, free to enter into relations as they choose" (p. 407). But if social capital is embedded in formal organizations, such as Catholic schools, and tighter closure inside particular cults reinforces this social grist, how can we assume that the individual and his or her volition will remain so salient in this theory of human motivation, as Coleman would have it? Portes (1998) emphasizes a similar problem with Coleman's pragmatic conception of social capital. It centered on its utility and fungibility for the individual recipient. But what motivates the web of actors - be they individuals or institutions - who donate social capital to human-scale networks? In general, we have little empirical knowledge of who produces social capital and within what social networks and institutions, as Schaub and Baker emphasize in their commentary. This includes teachers, high school coaches, and members of school councils. They spend day in and day out contributing to local norms and expectations for children's achievement and mobility, often by building trust through asymmetrical, not reciprocating, relations. Certainly "the altruistic dispositions of donors (of social capital) may be bounded by the limits of their community" (Portes, 1998, p. 8). But can altruism be explained by individualistic conceptions of exchange or reciprocity? Do all those nuns in Coleman's parochial schools believe that they are making deposits of social capital in students' accounts and some day their graduates will return the favor?
BRINGING SOCIAL CAPITAL TO ISSUES OF SCHOOLING

The chapters in the present volume advance our understanding of three slippery issues related to the social capital idea. Do the webs of trust, shared commitments, and reciprocity that comprise social capital reside in culturally bounded communities - from ethnic enclaves to peer groups - or can deposits of social capital be accumulated across communities? What role do formal institutions, such as schools, play in producing and sustaining social capital? Is social capital a touchable resource that is rationally drawn down by the individual who is pursuing particular interests? Or is social capital a social
web that socializes individuals in taken-for-granted ways, built'upon trust, behavioral scripts, and normative encouragement? In addition, our contributors speak directly to an issue that sharply separate Putnam and Coleman. For Putnam social capital is an account of how group solidarity and institution building may unfold. But for Coleman the ideals of social capital express faith in neoclassical formulations of rational choice and the local group's efficacy in enforcing its conserving tendencies and sacrasanct norms. Our contributors and commentators return to this question of whether formal institutions or indigenous networks host deposits of enabling kinds of social capital. Indeed, Putnam's (2000) well known application of social capital brings into focus a pivotal issue that is so central to the future of public schooling and child development. His characterization of social capital, at first, resembles Coleman's: "familiarity, tolerance, solidarity, trust, habits of cooperation, and mutual respect" (p. 362). But rather than linking the social capital resource to a neoclassical individual, as Coleman would have it, Putnam goes macro, arguing that societies with more aggregate social capital manifest greater levels of civic engagement and a richer mix of intermediate, human-scale organizations that produce and sustain this social resource. Whereas Coleman (1990, p. 652) worried that public institutions were eroding highly localized or "primordial social capital," Putnam remains optimistic that the state and other institutions can stimulate voluntary and nonprofit organizations at the grassroots. The institution of schooling is a ripe site for illuminating the sources and uses of social capital, and for arbitrating between the contradictions that have arisen in this young literature. We have long known that differing local communities - and forms of parenting - contribute tangible scaffolds which support children's learning and socialization inside schools. The chapters that follow shine a bright light on such elements of social capital - informal and institutional - that operate in diverse ethnic settings. Schools are organizations, under the logic of modernity, that aim to build non-local forms of trust, social obligation, and human capital. And individual rights come into play in terms of equalizing simple access to formalized social networks (institutions). But the socializing and skilling function of schooling reflects a pivotal intermediary process - mediating between local norms and the modern agenda of the nation-state. In recent years, reformers who have pushed radically decentralized forms of schooling, like charter schools, are trying to recreate the village organization that was more tightly coupled to local norms, to village-level forms of social capital if you will. But will this dynamic truly advance, or simply segment and stratify, forms of opportunity for students, as commentator Patricia Fernandez Kelly asks so forcefully?
HOW DOES SOCIAL CAPITAL ADVANCE CHILDREN'S LEARNING?

Our contributors are inventively tapping into the social capital frame to understand the underlying causes of children's often mediocre and persistently unequal levels of school achievement. Each chapter offers a fresh viewpoint in advancing our understanding of how social networks outside and inside the school institution construct scaffolds for youngsters' motivation and supports for their learning. The settings and problems addressed in this volume are wonderfully varied, and the empirical investigations are revealing. Carl Bankston and Min Zhou begin by questioning Coleman's thesis that Asian children in North America do well in school because there is tight closure and family-reinforced norms inside their networks. Instead, these authors focus on the role played by formal, human-scale institutions - particularly local churches - in advancing social capital that is related to school achievement. Kimberly Goyette and Gilberto Conchas bolster this line of analysis, shedding a bright light on how peer groups and teachers, for Latino and Vietnamese-American children, contribute to learning trajectories, seemingly independent of social dynamics inside the nuclear family. Grace Kao moves the analysis back inside families, moving to broader empirical ground by using national survey data to understand one facet now placed under the social capital umbrella: educational aspirations held by ethnic-minority parents for their children. The final two chapters move us outside North America to examine the interplay between social supports felt by children and youth and the arrangement of formal institutions. Claudia Buchmann looks comprehensively at how the cultural form of schooling in Kenya variably maps onto the language and social norms that operate in particularly families. In addition, some families can invest in after-school tutorial programs ("shadow education"), which contributes further to the cultural capital that's so influential in driving up school achievement. Finally, Shavit, Ayalon, and Kurlaender examine how second-chance schools in Israel offer formal supports - and material mobility - for many offspring of working-class families. These authors do not explicitly draw from a social capital frame. But the small-scale organizations that they describe operate outside networks that provide limited social capital for these youths when it comes to education-powered mobility. This bolsters Fernandez Kelly's argument, amplified in her commentary, that most pivotal sources of social capital for many youths may spring from institutions, not from their insular networks. We asked three commentators to review these original papers. Critical reviews by Maryellen Schaub and David Baker, Patricia Fernandez Kelly, and Raymond
Wong follow chapters or pairs of chapters. They help to set this new empirical work in the wider context of the application of the social capital construct to issues of schooling and student achievement. All this fresh research prompts the crucial question: Are we making collective progress in building theory? Does the social capital construct buy us much in how we understand the role played by actors - inside and outside the school - in building stronger scaffolds for children's learning and their motivated engagement in the school itself? Figure 1 offers one framework for seeing how our contributors offer pieces for the theory-building puzzle. A well-developed account of how social capital contributes to children's learning and socialization would include a clear logic and empirical findings within each of these cells. That is, we would want to know about the life cycle of social capital in the lives of children and youth. What actors create and sustain the key elements of social capital, be it trust, strong norms, enforced expectations, or reciprocity? How is social capital drawn down inside communities and by whom? If deposits are made, who withdraws quantities of social capital? Is it expended by recipients or depleted by donors as networks break apart? At the same time, a clear account of social capital's contribution would more thickly describe where to find it, the character and content of expectations or norms, and how embedded - taken for granted or contested in a cultural sense - the stuff of social capital really is within particular settings. One way to dissect this dimension is to think about social capital's location or situational bounds. And how are the elements signaled to children? Coleman and others have talked of sanctions and clear messages in close-knit networks. But we have little ethnographic evidence to substantiate such mechanisms within the social capital framework. And another cultural issue arises; is social capital reproduced and exercised in tacit or quite self-conscious ways?
Life cycle of social network > > Creating Character and organizational embeddedness Location or situation Media and signaling Actors and intentionality Fig. 1. Social Capital Theory Building. Sustaining and Drawing down, reproducing expending
10
Social capital proponents have largely ducked this issue of actors and their level of intentionality. There is certainly a romantic attraction to human-scale communities that advance norms and positive expectations in taken-for-granted ways. We see it in Putnam's affection for civic life in Italy, and Coleman's recurring seduction with the heart of Catholic schools. But tacitness and closure are powerful elements of gangs, peer groups that blast high achieving black children for "acting white," and teacher cliques that resist innovation and change inside schools. This is why formal organizations - from school boards to neighborhood agencies - formulate goals, reforms, and policy "interventions." If the indigenous construction of social capital is so splendid, why is mobility out of poverty via school attainment still so rare? This volume's final chapter returns to the issue of whether social capital offers new tools for understanding the reproduction of inequality and stratification.
VALUE-ADDED BY SOCIAL CAPITAL? DRAWING FROM CLOSE CONCEPTUAL COUSINS

Certain chapters also examine elements now included in the social capital box that have a history all their own, including the study of parents' educational aspirations and the social networking effects of schools. This prompts the question of what related facets of social support or normative pressures within local communities are to be put into, or remain outside, the social capital framework? The construct has become so popular - a reified artifact of trust among social analysts - it's adhering to all sorts of other constructs and social forces. But what is the distinct value-added from building a theory of social capital linked to the children's development and the effects of formal schooling? It is worrisome that so many scholars - smitten by the social capital bug - either jettison neighboring theoretical frames or obsess on a particular element, like the radical idea of trust, losing sight of the broader social architecture that sustains it. Similarly, Alexander and Smelser (1999) argue that Putnam's broad claims about the erosion of civil society, evidenced by the decline in some voluntary organizations, ignore the dramatic rise of nonprofit firms and community action agencies that run all sorts of public services. More broadly, they emphasize that micro social organizations - family, church, and community groups - have long reinforced social roles and scripted norms for individuals. The explanatory problem rests at this local institutional level: how to strengthen intermediate collectives to enrich membership while sustaining a pluralistic range of organizations to ensure inclusion and civil
11
dialogue across groups (Cohen, 1999). In the absence of variety among and within neighborhoods, social capital networks could become limiting and encrusted in parochial pockets. Moving down to informal networks, we run the risk of ignoring established accounts of how individuals and small groups experience pressing norms, expectations, and rules for trust or reciprocity. From Emile Durkheim through the past three generations of cultural scholars, the reproduction of taken-forgranted routines or scripts has been a central focus of study (e.g. Goffman, 1959; Holland & Quinn, 1987; Collins, 1994). Coleman's neoclassical rendition of social capital assumes, instead, a more rational individual reflecting on and extracting situational norms. Yet in the schooling arena, family-institution conceptions of parent and child action have assumed families' express varying levels of consciousness and contestation around dominant norms (e.g. Fuller & Liang, 1999). The pivotal issue of agency and tacitness, including how institutions create norms and individuals internalize some of them, has been investigated in the cultural capital arena since the work of Antonio Gramsci, amplified by Paul DiMaggio and in the work of Annette Lareau, replete with rich descriptions of how cultural capital is drawn upon by parents and teachers alike to include or exclude certain kids (Lareau & Horvat, 1999). In our faithfilled rush to become members of the social capital camp, scholars are losing track of these earlier theoretical developments. As you read the chapters that follow, the distinct advantages of the social capital frame will certainly shine through. It's an elastic bundle of constructs - expectations, cultural bonds, expressions of trust and reciprocal support that illuminate how membership and situational norms come to vary in strength. The frame holds enormous promise for explaining group differences in achievement and how school organizations can become more resourceful hosts of social capital. The local, post-structural positioning of the framework also allows us to better understand how certain children beat the odds, responding to particular supports in particular human-scale settings. Overall, you will see in these chapters how the constituent elements of social capital operate within and outside schools, and how this new knowledge invites educators and community agencies to bolster their own stock of this social glue. Still, in reading these papers you may be struck by the fact that the idea and the ideals inherent in social capital remain works in progress, especially their applications to child development and schooling. The discrete actors and forces within networks often remain hazy, unclearly operationalized and observed. Abstract sketches of its basic architecture often hide what lays inside, especially how norms, trust, and expectations are actually felt by children who benefit from, and are starved for, social capital. And finally, the logic and
12
causal pathways by which the substance of social capital influences children's learning and motivation too often remain unspecified, empirically unexamined. The sudden excitement around social capital is reminiscent of the hoola-hoop craze of the 1960s. The novel motion is seductive. The effect is exhilarating. But explaining what's going on and how to improve these delighting effects is proving to be more challenging.
REFERENCES
Alexander, J., & Smelser, N. (1999). The Ideological Discourse of Cultural Discontent. In: N. Smelser & J. Alexander (Eds), Diversity and Its Discontents (pp. 3-18). Princeton: Princeton University Press. Bourdieu, P., & Passeron, J. (1977). Reproduction in Education, Society and Culture. Beverly Hills: Sage Publications. Cohen, J. (1999). Does Voluntary Association Make Democracy Work? In: N. Smelser & J. Alexander (Eds), Diversity and Its Discontents (pp. 263-293). Princeton: Princeton University Press. Coleman, J. (1988). Social Capital in the Creation of Human Capital. American Journal of Sociology, 94, S95-S120. Coleman, J. (1990). Foundations of Social Theory. Cambridge, Mass.: Belknap Press. Collins, R. (1994). Four Sociological Traditions. New York: Oxford University Press. Fuller, B. (2000). Inside Charter Schools: The Paradox of Radical Decentralization. Cambridge, Mass: Harvard University Press. Fuller, B., & Liang, X. (1999). Which Girls Stay in School? The Influence of Family Economy, Social Demands, and Ethnicity. In: C. Bledsoe et al. (Eds), Critical Perspectives on Schooling and Fertility in the Developing World. Washington D.C.: National Academy Press. Goffman, E. (1959). The Presentation of Self in Everyday Life. New York: Doubleday. Holland, D., & Quinn, N. (Eds) (1987). Cultural Models in Language and Thought. Cambridge, England: Cambridge University Press. Ladd, E. (1999). Tlie Ladd Report. New York: Free Press. Lareau, A., & Horvat, E. M. (1999). Moments of Social Inclusion and Exclusion: Race, Class, and Cultural Capital in Family-School Relationships. Sociology of Education, 72, 37-53. Loury, G. (1977). A Dynamic Theory of Racial Income Differences. In: P. Wallace & A. LeMund (Eds), Women, Minorities, and Employment Discrimination. Lexington, Mass.: Lexington Books. Menand, L. (2001). The Metaphysical Club: A Story of Ideas in America. New York: Farrar, Straus and Giroux. Portes, A. (1998). Social Capital: Its Origins and Applications in Modern Sociology. In: Annual Review of Sociology (pp. 1-24). Palo Alto: Annual Reviews Inc. Putnam, R. (1993). Making Democracy Work. Princeton: Princeton University Press. Putnam, R. (1995). Bowling Alone: America's Declining Social Capital. Journal of Democracy, 6, 65-78. Schneider, B., & Coleman J. (1993). Parents, Their Children, and Schools. Boulder: Westview.
SOCIAL CAPITAL AND IMMIGRANT CHILDREN'S ACHIEVEMENT

Carl L. Bankston III and Min Zhou
ABSTRACT
We suggest that the commonly used familial closure version of social capital does not provide an adequate explanation for the school achievement of children in immigrant families. Instead, we suggest that extra-familial institutions, notably immigrant religious institutions, contribute to the school performance of children. Using data from the National Longitudinal Study of Adolescent Health, we find that the average grades of immigrant children can be accountedfor by involvement in ethnic religious organizations, but not by parental involvement in social networks.
INTRODUCTION
The school adaptation of immigrant children lies at the heart of the debate on the fate of the new second generation (Gans, 1992; Portes & Zhou, 1993; Perlmann & Waldinger, 1997). Since the "new immigration" is of relatively recent vintage, with 44% of the immigrants who currently reside in the United States having come here after 1980 (a percentage still higher among those of non-European origin), the new second generation is an overwhelmingly youthful population, consisting mostly of children or adolescents, with a small but significant portion maturing into adulthood.
Schooling and Social Capital in Diverse Cultures, Volume 13, pages 13-39. Copyright 2002 by Elsevier Science Ltd. AH rights of reproduction in any form reserved. ISBN: 0-7623-0817-6
13
14
CARL L. BANKSTON III AND MIN ZHOU
The existing literature on immigrant education is essentially concerned with two pressing issues: one is whether or not the new second generation will be able to incorporate into middle-class America, following the path taken by the "old" second generation arriving at the turn of the century, and to advance beyond their parents' generation; and the other is what accounts for interethnic differences in educational outcomes. Recent research suggests that the pessimistic renderings of second generation scenarios are unwarranted (Zhou, forthcoming). Studies in the education area have consistently revealed that children of immigrants tend to out-perform natives in American schools and that children of Asian immigrants tend to do especially well (Fuligni, 1997; SuarezOrozco & Suarez-Orozco, 1995; Rumbaut, 1995, 1998; Steinberg, 1996; Sue & Okazaki, 1990; Zhou, forthcoming; Zhou & Bankston, 1998). These trends suggest that many of the children of today's immigrants are making it - and doing so better and faster than immigrant children have ever done before. Some of these differences are clearly associated with class differences among the immigrants themselves, who are far more diverse in socioeconomic circumstances than previous waves. But if class backgrounds explain why the children of foreign-born physicians, engineers, or computer specialists show up in elite universities, it is more difficult to explain divergent outcomes among groups that begin under modest circumstances. The concept of "social capital" is often employed to explain the latter phenomenon (Zhou & Bankston, 1994: Bankston, Caldas, & Zhou, 1997). The conceptualization of social capital, however, is frequently vague, and the term has been used in such a wide variety of ways that it is often difficult to define precisely (Portes, 1998). Nonetheless, as Portes (1998, p. 4) observes, "despite these differences (in usage), the consensus is growing that social capital stands for the ability of actors to gain benefits by virtue of membership in social networks or other social structures." One of the clearest expressions of the network approach to social capital may be found in the work of James S. Coleman. Prior to Coleman's writings on the subject, Loury (1977, 1981) used the term social capital to refer to the resources inherent in family relations and in community organizations for the development of children. Reacting to neoclassical, individualistic theories of racial income inequality, Loury argued that black families would tend to pass poverty on to their children, even in the face of equal opportunity programs, because poorly educated black parents could not contribute adequately to the educational advancement of children and because young black workers had weaker connections than whites to organizations and labor markets (Portes, 1998). Loury's term became the basis of Coleman's more systematic formulation (Astone et al., 1999). Coleman's formulation gave a central place to structures of family relations. In Foundations of Social Theory (1990a) and
Social Capital and Immigrant Children's Achievement
15
other. works, Coleman defines social capital in terms of network closure. Social capital exists, in Coleman's view, when there are close and closed networks among a set of individuals, promoting advantageous behavior. When parents, for example, maintain close contacts with their own children and with other adults who affect the lives of their children, parents and other adults in the parental networks can impose consistent norms and standards to direct the behavior of young people. Coleman employed this network closure model of social capital to explain the academic achievement of Asian children, the overwhelming majority of whom are children of immigrants. As it is formulated by Coleman, the network closure explanation assumes that immigrant children do well in school because they and their families are interconnected in closed, intimate social networks. There are good reasons to question this assumption, though. First, immigrant parents frequently face special difficulties in forming tight networks with other adults (Zhou, 1997a). Second, relations between immigrant parents and children growing up in the United States are often strained by cultural gaps (Zhou, 1997a). In this study, we test whether immigrant parents do exhibit greater inter-parental network closure than non-immigrant parents. We then examine the evidence for the relationship between network closure and academic achievement. We also offer preliminary indications of a model of social capital that is an alternative to the network closure model. Using the example of participation in religious institutions, we suggest that influences on the academic performance of the children of immigrants may be thought of in terms of parallel institutions, rather than in terms of network closure. From this perspective, immigrant social settings can generate social capital when immigrants participate in institutions or organizations that promote advantageous outcomes. We point out that if ethnicity is to be thought of as a source of social capital for immigrants (Bankston et al., 1997), then it may be helpful to look beyond the atomistic connections among sets of individuals, as suggested by the network closure model. The ties of individuals, especially family ties, are frequently disrupted by migration. Ethnic group membership, however, is frequently enhanced and intensified by settlement in a new homeland. Although the perceived ancestral character of ethnicity gives it its "primordial" character and therefore its emotional appeal, ethnicity is largely a matter of a collective sense of common ancestry (Geertz, 1973; Isaacs, 1975), and not a matter of actual ancestry inherited through families. Therefore, it is logical to think of the social settings that give support and encouragement to immigrants in terms of participation in institutions that arise
16
out of a sense of communal identity and that compensate for the incompleteness of ties among individuals and families.
THEORETICAL BACKGROUND
In the writings of Coleman and those in his tradition, social capital consists of closed systems of social networks inherent in the structure of relations between persons and among persons within a collectivity (Coleman, 1988; Coleman, 1990a; Portes & Sensenbrenner, 1993; Fernandez Kelly, 1995). The essence of this approach is that a "dense set of associations" (Coleman, 1990a, p. 316) within a social group can promote cooperative behavior that is advantageous to group members. Social capital is held by individuals by virtue of their membership and participation in groups. Adaptation does not result from isolated attributes of individuals, such as education or cultural characteristics, but from the structure of relationships among individuals. The idea of social capital is not limited to explanations of immigrant adaptation, but it can be used as one explanation of why different immigrant groups show different rates of success in adapting to host societies. On the issue of education, Coleman cites evidence from Asian families that parental interest in children's learning can promote academic achievement even when the parents have little human capital. Moreover, Coleman finds that the stability and the strength of a community's social structure plays a vital role in supporting the growth of social capital in the family. Social capital in a community allows parents " . . . to establish norms and reinforce each other's sanctioning of the children" (Coleman, 1990a, p. 318). From this point of view, immigrant children do well in school because of closure in their network connections with their parents and other adults and in the network connections between their parents and other adults. "The adults are able to observe the child's actions in different circumstances, talk to each other about the child, compare notes, and establish norms" (Coleman, 1990a, p. 593). The more closely parents are connected to other relevant adults, such as the parents of their children's friends and their children's teachers, the more social capital exists. Coleman defines a "functional community" as one in which this type of closure can be found (Coleman, 1990b, pp. 318-319). The network closure model of social capital has been an influential one in studies of the adaptation and academic performance of members of immigrant minority groups (Clark & Ramsay, 1990; Fernandez Kelly, 1995; Furstenburg & Hughes, 1995; Zhou & Bankston, 1994; Bankston, Caldas & Zhou, 1997; Sun, 1998; Hao & Bonstead-Bruns, 1998: Zhou & Bankston, 1998). Teachman et al. (1996, 1997), looking at influences on dropping our of school for students
17
in general, see patterns of parental interaction as a major indicator of social capital. Despite the influence of the parental network closure model of social capital, though, it may be difficult to use this as an explanation of school achievement by immigrant students. Migration can both disrupt family relationships and network connections between family members and those outside the family (Landale, 1996; Zhou, 1997a). Gold (1995) has found that among Israeli immigrant families in the United States, migration results in a loss of the close community bonds among adults that facilitate child-rearing in Israel. However, Gold's research indicates that this loss of network closure is partially compensated by the development of greater intimacy between mothers and children, made possible by the lower female labor force participation among Israelis in the United States than in Israel. Even this second type of social capital, based on dyadic parent-child relations rather than on triadic parent-other adult-child relations, may not be available to most immigrant families, though. Research on parent-child relations in immigrant families has generally shown that migration tends to disrupt relations between foreign born parents and children growing up in the United States and create problems for parent-child intimacy (Berrol, 1995; Kibria, 1993; Rumbaut, 1994; Sung, 1987; Waters, 1996; Zhou, 1997b). Working outside the home, the norm for most immigrant women as a family strategy to obtain financial resources, puts further strains on spousal relations and parent-child bonds since parents tend to work long hours and on different shifts away from home (Zhou, 1992, chap. 7). If migration truncates intimate family networks and interaction patterns in immigrant households, does it mean that the concept of social capital as network closure is not useful for investigating immigrant school achievement? In our view, the network closure model and other models of social capital in existing literature are helpful, but need clearer delineation and alternatives to the family network closure model require greater emphasis. Banks (1997) has argued that voluntary organizations are successful in promoting their goals when the organizations serve to link members to each other and thereby develop shared norms of collective action. Along somewhat similar lines, Zhou (1997b) considers community-based organizations as means of generating social capital. From this point of view, social capital does not inhere immediately in the social relations among individuals, but in the formal organizations and institutions that structure and guide those social relations. A number of researchers have maintained that religious institutions are among the most important formal structures for the social relations of immigrants. Thomas and Znaniecki, in their classic work The Polish Peasant in Europe and America (1927), found that immigrant religious institutions were focal points
18
for ethnic identity. Barton (1975), argues that religion has provided basic social units to communities of Italian, Rumanian, and Slovak immigrants in the United States. According to Warner (1998, p. 193), immigrants tend to be even more religious in their new home than they were before migration, "because religion is one of the important identity markers that helps them preserve individual self-awareness and cohesion in a group." Immigrant-based religions can be important for the adaptation of new arrivals because these formal structures compensate for the tightly integrated informal social networks that immigrants may have lost in the process of geographic mobility. From this point of view, immigrant religions and other formal social structures arise precisely because migration is so unsettling and disruptive. If such institutions do promote adaptation to the new homeland, they do so by providing a compensatory form of social capital, in which the support and encouragement of individuals are consequences less of dense social networks than of formal structures created to take the place of dense social networks. The literature on immigrant religious institutions has identified a number of mechanisms by which these institutions can act as sources of social capital. First, immigrant churches, temples, and mosques frequently establish formal community centers to serve the needs of members or participants (Bankston & Zhou, 2000; Haddad & Lummis, 1987; Hurh & Kim, 1990; Kashima, 1977; Kurien, 1998; Le6n, 1998; Min, 1992). Ebaugh and Saltzman (2000a, b), in comparative studies of immigrant congregations, found that the community center model is a common pattern. Religious community centers help young immigrants and children of immigrants, in particular, by providing classes teaching both parental languages and skills for American schools. Second, even when immigrant religions do not include formally established community centers, they act as network foci for social relations among immigrant group members. Feld (1981, p. 1016) has defined a network focus as " . . . a social, psychological, legal, or physical entity around which joint activities are organized (e.g., workplaces, voluntary organizations, hangouts, families, etc.)." Since social capital inheres in relations among individuals, religious institutions can generate advantageous outcomes by providing bases for organizing sets of relations. Kwon (1997) has found that Korean churches can create business opportunities for Korean entrepreneurs, even though these churches are not founded or maintained in order to supply members with opportunities for profit. The churches act as focal points for personal contacts and the exchange of information. Similarly, Zhou, Bankston, and Kim (forthcoming) have found that Buddhist temples and Christian churches among Southeast Asian refugees in the United States create opportunities for adherents. Young people, who are usually seeking to adapt to schools rather
19
than to job markets, can find in their religious institutions information about the American schools and about the best ways to succeed in these schools (Bankston & Zhou, 1996; Zhou & Bankston, 1998). This second point raises the question of why information about American schools within immigrant networks would have more value than the information young people could obtain elsewhere. As newcomers to American society, immigrants frequently have strong orientations toward achieving upward mobility. Ogbu (1974, 1990) has argued that one of the crucial differences between immigrant minority groups and native-born minority groups is that the latter often tend toward upward mobility, while the former are frequently oriented toward maintaining psychological defenses against historically imposed social disadvantages. Ogbu maintains that this difference is a primary reason for the comparative success of immigrant children in American schools. Information about education in immigrant networks, then, is information obtained in an environment dominated by expectations for achievement. Immigrant religious organizations provide alternative social worlds, in which succeeding in school is strongly valued. Thus, many immigrant adults see their religious institutions as places where their children can be encouraged to adjust constructively to American society while avoiding "Americanization" (i.e. conformity to American youth cultures) (Ebaugh & Chafetz, 2000c; Sullivan, 1998; Zhou & Bankston, 1998). The closely-knit, dense network ties within immigrant religions help to provide a normative mechanism for promoting academic success. At the church, temple, or mosque, young people form ties with other ethnic group members that establish normative supports and normative constraints (Greeley, 1997). Frequent contact with other members of the ethnic group reinforces the view that school achievement is an accomplishment to be valued and sought. Community centers at institutions can supply encouragement both by formally established classes and awards ceremonies and by informal means, such as informal expressions of approval for high-performing students (Ebaugh & Chaftez, 2000b; Zhou & Bankston, 1998). Conversely, gossip and other expressions of disapproval can constrain behaviors that are likely to undermine school performance (Bankston, 1995; Bankston, 1997). The normative contribution of religious institutions is not limited to immigrant children. Sanders (1998) found that church involvement, as well as parental academic support, positively affected the achievement of black urban children by bolstering academic selfconcepts and promoting constructive school behavior. As an ethnic network focus, the literature suggests that religious institutions can be particularly effective in promoting constructive behavior on the part of adolescents because these institutions help to overcome the problem of age
20
segregation. The institutions incorporate young people into groups that include adults and create ties between adults and adolescents (Ebaugh & Chafetz, 2000c; Bankston & Zhou, 1995a). Thus, relationships between minors and their elders in religious organizations can complement and serve many of the functions of relationships between parents and children in families.
DATA AND METHODS

The data in this study come from interviews conducted in 1995 as part of the National Longitudinal Study of Adolescent Health. The Public Use Sample that we used contains a random sample of 6,504 cases from these interviews. The Appendix gives a description of all variables included. We begin our analysis by presenting selected characteristics of children of immigrants and children of people born in the United States. This enables us to make some preliminary observations about how these two groups differ. In particular, it enables us to say whether children of immigrants do experience greater or less intimacy with parents and greater or less parental involvement in their social networks than native-born children do. Next, we use OLS regression to examine predictors of parental network closure; that is, predictors of the likelihood that parents will have contacts with the friends of their children and with the parents of the friends of their children. We enter the predictors in a series of models. By looking at how the coefficients of variables increase or decrease across models, we can, in the words of Clogg, Petrovka, and Haritou (1995, p. 1263), "test whether 'controlling' for a variable suppresses or enhances the relationship between two variables." We then turn our attention to predictors of involvement in religious institutions. We look at predictors of involvement in both ethnically based religions and non-ethnically based religions, without and then with the indicators of parent-child intimacy and parental network closure. This enables us to make observations about factors that are associated with involvement in religious organizations and it enables us to say whether ties of parents to children and to those surrounding children are related to involvement in religious organizations. It also enables us to say whether immigrant children or children of minority groups appear to participate in religious organizations as a consequence of parent-child ties. Finally, we look at predictors of academic achievement. Again, we proceed in a series of models. We begin with indicators of racial/ethnic group membership in a first model. In a second model, we include immigrant status. This enables us to say whether immigrant status is associated with academic achievement,
21
controlling for the race and ethnicity of immigrants, and it makes it possible to see whether any part of the associations between racial/ethnic group membership and academic achievement may be statistically explained by family immigrant status. A third step includes indicators of family socioeconomic status. In a fourth step, we include reported closeness to parents and parental network involvement in an equation. In this way, we can determine whether parent-child ties contribute to school performance and whether the school performance of immigrants and members of minority groups can be plausibly attributed to the closeness of ties between parents and children and the involvement of parents in the social networks of children. A fifth step brings in parental stress on higher education, a key normative predictor of the educational progress of children. Finally, in two models, we bring in involvement in non-ethnically based religious institutions and in ethnically based religious institutions.
RESULTS
Table 1 presents selected characteristics of children of immigrants and of children of U.S. natives. Respondents in both groups are approximately the same ages: The mean age for both sets of high school students is about 16 (the range is from 13 to 21 for children of immigrants and from 12 to 21 for children of natives). The two groups are also very similar in gender composition: just over half of both are female. As we mentioned above, children of immigrants come from diverse backgrounds, although Latinos make up fully half the category. About 20% of them are white, 18% are Asian, 11% are black, and 51% are Latino. Whites, of course, are much more numerous, so even though they make up one out of every five immigrants, Asian and Latino children are more likely to be immigrants. Only 3.5% of all whites are children of immigrants, while 72.8% of Asians and 50.8% of Latinos are from immigrant families (not shown in this table). The children of immigrants tend to come from slightly more limited socioeconomic backgrounds: their median family income is $5,000 per year less than that of the non-immigrant children. Children of foreign-born parents are also slightly less likely to come from families in. which at least one parent had a college degree and in which at least one parent was employed in a professional, technical, or managerial occupation. In considering these socioeconomic characteristics, it should be kept in mind that immigrant family status is highly correlated with Latino ethnicity. The other major group of immigrants, Asians, is often found to enjoy some socioeconomic advantages, particularly in parental educational status (Min, 1992). Our data also show that children in immigrant families are slightly less likely to come from single parent families.
22
CARL L. BANKSTON III AND MIN ZHOU Table 1. Selected Characteristics of Children of Immigrants and Children of U.S. Natives.
Immigrant children Children of natives 15.9 51.3 67.7 0.8 25.3 6.2 40,000 28.1 17.2 29.2 3.5 15.2 94.1 82.0 11.2 40.3 4.1 32.3
Age (mean) Sex (% female) White (%) Asian (%) Black (%) Latino (%) Median family income ($) Parent with college degree (%) Parent in professional/technical/managerial job (%) Single female family (%) Single male family (%) Less than quite close to parents (%) Parents have met best friend (%) Parents have met best friend's parents (%) Parents talked with friends' parents 6 or more times (%) Parents who would be very disappointed if child does not attend college (%) Highly involved in ethnically based religion (%) Highly involved in non-ethnically based religion (%) Percent of all immigrant children who are: First generation immigrants 1.5 generation Second generation N
16.2 51.3 19.7 17.9 11.1 51.3 35,000 25.8 15.7 26.8 2.8 20.6 85.3 68.2 4.9 61.4 23.6 19.1 43.3 7.4 49.3 635
5002
The measures of closeness between parent and child and of parent-child network closure show the greatest differences between the two groups. Children of immigrants are more likely than children of native-born families to report that they are less than "quite close" to their parents: 20% of the former report that they are not at all close to somewhat close to their parents, compared to only 15% of children of the native-born. Foreign-born parents are, moreover, less likely than the U.S. born to report in the parent interviews that they had met the best friends of their children. The difference in inter-parental network closure is even more marked. Only 68% of immigrant parents report having met the parents of their children's best friends, compared to 82% of non-immigrant parents. U.S.-born parents, moreover, are twice as likely to have met with parents of their children's friends six or more times in the previous four weeks. The relatively lower intimacy between immigrant
23
children and their parents and the apparently lower levels of parental network closure are particularly impressive when we consider that immigrant families were more likely to be two-parent families, which should have a structural advantage in maintaining parent-child relations and in communicating with other parents. In contrast to the comparatively limited intimacy and network closure of immigrant families, immigrant parents appear to have higher educational aspirations for their children than native-born parents do. Fully 61% of the foreign-born parents say that they would be "very disappointed" if their children did not attend college. Only 40% of parents born in the United States give this answer. The children of immigrants show greater involvement in religious institutions. Among them, 43% give responses that indicated high involvement in some kind of religious institution (i.e. their responses give a score of 7 or above on our 0-9 point scales). By contrast, 36% of children of natives indicate such high levels of religious involvement. The difference is even more marked if we look specifically at involvement in ethnically oriented religious institutions. Nearly a quarter (24%) of the immigrant children are highly involved in such institutions, compared to only 4% of children of native-born families. While it is true that many of these ethnically oriented institutions are by definition immigrant religious organizations (such as Vietnamese Catholicism, Asian Buddhism, Korean Presbyterianism, Eastern Orthodox Christianity), others are based on native-born ethnicity or race (such as the African Methodist Episcopal denomination, the National Baptist denomination, and Nation of Islam). Hispanic Catholicism, which we argue can be considered as ethnically based in its parochial institutions, can be connected both to native-born and immigrant Latino ethnicity. Although immigrant parents are less involved in social networks involving their children, then, the immigrant children themselves do tend to be more involved than others in a major form of social concentration: the religious institution. Finally, Table 1 presents the generational categorization of immigrant children. Most children of immigrants are either first generation or second generation; that is, they either arrived in the U.S. after the age of 5 or they were born in the United States. Only 7% of the adolescents included in this study can be classified as "1.5" generation. Table 2 attempts to look at predictors of parental network closure by regressing our network closure scale on a number of theoretically important predictors, including race and ethnicity, child immigrant status, family socioeconomic characteristics, family structure, and the degree of intimacy children report with their parents. Model 1 presents age and sex (0 = male, 1 = female)
24
as control variables and three major racial/ethnic categories in the United States. White, non-Hispanics are the reference category. Asians, Blacks, and Latinos all show significantly less parental network closure than Whites. Asians and Blacks had less closure by nearly a full point on our 0 to 8 point scale; Latinos had less parental network closure by more than one point. Parents in all minority groups were, in other words, less likely than Whites to be involved in the type Table 2. Unstandardized Coefficients of Predictors of Parental Involvement in the Social Networks of Children. (Standard Errors in Parentheses)
Model 1 Age Sex Asian Black Latino First generation 1.5 generation Second generation Family income Parental education Parental occupation Single female family Single male family Closeness to parents Constant a R2 6.335** (0.273) 2.078 0.066 6.077** (0.273) 2.067 0.076 4.906** (0.281) 2.022 0.119 5.083** (0.281) 2.012 0.128 -0.135** (0.016) 0.123* (0.056) -0.840** (0.167) -0.888** (0.067) -1.207** (0.091) Model 2 -0.124** (0.016) 0.129* (0.055) -0.375* (0.178) -0.886** (0.067) -0.890** (0.100) -1.200** (0.162) -0.480* (0.234) -0.431** (0.131) Model 3 -0.124** (0.016) 0.137** (0.055) -0.498** (0.177) -0.793** (0.068) -0.647** (0.102) -1.214** (0.165) -0.665** (0.231) -0.436** (0.130) 0.003** (0.001) 0.194** (0.015) 0.010* (0.005) Model 4 -0.122** (0.016) 0.124* (0.055) -0.503** (0.176) -0.727** (0.069) -0.623** (0.101) -1.240** (0.164) -0.699** (0.230) -0.454** (0.130) 0.003** (0.001) 0.182** (0.015) 0.008 (0.005) -0.319** (0.065) -0.957** (0.153) Model 5 -0.104** (0.016) 0.168** (0.055) -0.471** (0.176) -0.749** (0.069) -0.632** (0.101) -1.234** (0.163) -0.700** (0.229) -0.434** (0.129) 0.003** (0.001) 0.181** (0.014) 0.008 (0.005) -0.326** (0.065) -0.938** (0.152) 0.229** (0.034) 3.722** (0.344) 2.003 0.136
25
of inter-parental connections that Coleman identifies with social capital. We note that parents tended to be more involved in networks surrounding daughters than sons and that parental involvement decreased as children grew older. In Model 2, we bring in the immigration status of children. Children of immigrants had significantly less parental network closure than children of nonimmigrants. The children who were themselves first generation immigrants showed the least closure of all. We note that the coefficient for Asians decreases by over half its value (d = -0.465) when we control for immigrant status, suggesting that one major reason Asian children apparently have so little parental supervision is that their parents are immigrants. Using the method suggested by Clogg et al. (1995), we can determine that the standard error in the difference of the slopes for Asians is s(d) = 0.033 (i.e. s2(d) = s2(b z) -s 2 (b yx )a 2 v / o-2e, with civ = 2.06732 and ae = 2.07781). Thus, the coefficient for Asians decreases by a statistically significant quantity when we control for immigrant status. The coefficient for Latinos, the other group with a large proportion of children of immigrants also decreases substantially and significantly (d =-0.317, s(d) = 0.030) when we control for immigrant status. For both of the groups that are largely composed of children of immigrants, then, it appears that relatively limited parental contact with friends and parents of friends of children is partly explained by the fact that the fact that the parents are immigrants. Since the race/ethnic categories and the categories of immigration status are all dichotomous variables, we can use the unstandardized coefficients to compare relative strength of association with parental network closure. Being a first generation immigrant is more strongly related to lower parental oversight through network connections than is race, ethnicity, or being male. The two other categories of children of immigrants also are also significantly less likely to experience the involvement of their parents in their social networks in a manner that could facilitate supervision. Model 3 brings in the socioeconomic characteristics of families to see if the more limited socioeconomic background of immigrant children or minority groups could account for differences in parent-child network closure. This does not seem to be the case, although the coefficient for Latino adolescents does decrease somewhat when we control for family socioeconomic status. As we see in Model 4, family structure also does not appear to account for group differences in parental network closure, even though single parent families do have fewer connections with the children's friends and the parents of their children's friends than two-parent families do. Reported closeness to parents, considered in Model 5, is positively and significantly associated with parental contacts of the sort that promote parental supervision according to Coleman.
26
After controlling for parent-child intimacy, though, the minority groups and the children of immigrants still show less parental network closure than other children do. Table 2 suggests that there is little reason for supposing that immigrant children do well in school because of close contacts among parents that produce high levels of parental supervision. Children of immigrants, in fact, appear to be subject to less inter-parental control and direction than other children are. Immigrant parents are notably unlikely to be "able to observe the child's actions in different circumstances, talk to each other about the child, compare notes, and establish norms" (Coleman, 1990a, p. 593). Families, however, constitute only one part of the social setting surrounding immigrants. Immigrants are also part of institutions and organizations based on communal identities that extend beyond families and contain families. We have identified religious institutions as particularly important expressions of immigrant ethnic identity. In Table 3, we look at evidence regarding how ethnicity and immigrant status may be connected to involvement in religious institutions. Model 1 in the far left-hand column presents unstandardized coefficients of predictors of involvement in ethnic religious institutions, before we include the measures of parent-child intimacy and parental involvement in the social networks of children. Members of all three racial/ethnic minority groups are more likely than whites to be highly involved in ethnic religious organizations. This may be true by definition, since these organizations are defined as those based on minority race or ethnicity. However, even when we control for minority group membership, children of immigrants are more likely than other children to be involved in ethnically-based religious institutions. Only those in the "1.5" generation do not show a significant coefficient, since there are so few adolescents in this category. Introducing the parent-child relations variables, in Model 2, results in little change in the coefficients of the ethnic groups or in the coefficients of the immigrant status variables. It does not appear, in other words, that the involvement of immigrant adolescents or adolescents of minority groups in ethnically based religious institutions is connected to the relations between the young people and their parents. This would be consistent with the argument that participation in ethnic religious organizations is not a consequence of tight parent-child relations. From the perspective of this argument, any positive influence that such religious participation might have on the lives of children can be seen as largely independent of familial relations, and even as a compensation for problems in familial relations. When we look at predictors of religious institutions without an identifiable ethnic basis, in the last two models of this table, we see a somewhat different
27
Table 3. Unstandardized Coefficients of Predictors of Involvement in Ethnic and Non-Ethnic Religious Institutions. (Standard Errors in Parentheses)
Ethnic Religious Institution Model 2 Model 1 Age Sex Asian Black Latino First generation 1.5 generation Second generation Family income Parental education Parental occupation Single female family Single male family Closeness to parents Parental network involvement Constant a R2 0.574* (0.246) 1.764 0.258 -0.041** (0.014) 0.142** (0.049) 0.444** (0.152) 0.561** (0.060) 3.096** (0.088) 0.922** (0.142) 0.283 (0.198) 0.537** (0.113) 0.001 (0.0005) 0.003 (0.013) -0.0001 (0.011) -0.177** (0.056) 0.075 (0.132) 0.042** (0.014) 0.145** (0.049) 0.471** (0.155) 0.560** (0.062) 3.136** (0.089) 0.954** (0.145) 0.220 (0.202) 0.568** (0.114) 0.0002 (0.0005) -0.003 (0.014) -0.0001 (0.011) -0.170** (0.057) 0.100 (0.134) 0.023 (0.030) 0.002 (0.012) 0.491 (0.308) 1.765 0.262 Non-Ethnic Institution Model 1 Model 2 -0.131** (0.024) 0.264** (0.082) -0.089 (0.263) 1.018** (0.104) -2.299** (0.152) -0.448 (0.245) -0.051 (0.341) -0.517** (0.194) -0.003** (0.001) 0.163** (0.023) 0.016 (0.019) -0.674** (0.097) -1.306** (0.229) -0.079** (0.024) 0.327** (0.083) 0.026 (0.263) 1.115** (0.105) -2.271** (0.152) -0.170 (0.246) 0.183 (0.343) -0.486** (0.193) -0.004** (0.001) 0.135** (0.023) 0.011 (0.019) -0.645** (0.097) -1.082** (0.229) 0.349** (0.051) 0.175** (0.020) 2.586** (0.524) 3.001 0.141
5.625** (0.425) 3.045 0.116
picture. Asians are no more likely to participate in such institutions than whites are. While Latinos are much more likely to participate in ethnic religions and denominations than whites, the former are much less likely to participate in non-ethnic institutions. It should be recalled that we classified Hispanic
28
Catholicism as an ethnically based faith, and this is probably the reason for the drastic difference in the coefficients of Latinos in the two first models. Black adolescents, however, are more likely to engage in religious activities in both types of religious institutions. Among the immigrant categories, only second generation immigrant children are significantly different from the children of non-immigrants. However, in contrast to the columns including predictors of ethnically based institutions, all of the coefficients of the immigrant categories here are negative. Model 2, in the far left column, brings in the measures of closeness to parents and parental involvement in the social networks of children. Both parent-child intimacy and parental network closure are significantly and positively associated with the religious involvement of children. Perhaps the most interesting finding in this table is that parent-child bonds of the type that Coleman identifies with social capital do appear to make it more likely that non-immigrant children will participate in non-ethnically based religions. However, these kinds of bonds do not appear to increase the likelihood that children of immigrants will be involved in ethnically based religions. It is, of course, difficult to make statements about causality. However, we can pose alternative causal patterns and ask which one is consistent with the findings. An atomistic view of social capital in families would maintain that tight interconnections among parents, children, and other adults provide oversight for children and guide the children into larger supervisory social structures, such as religious institutions. The evidence that we see in the last two columns of this table, which deal with non-ethnically based institutions, is indeed consistent with this causal argument. The first two columns, though, are not consistent with it at all. Instead, the columns that deal with ethnically based institutions, the type of religion that tends to involve immigrants and minority group members, are consistent with a view of religious organizations as compensatory sources of social capital. In other words, it does not appear to be the case that close ties between immigrant parents and children steer the children into supportive ethnic organizations. Instead, the organizations arguably compensate for weakened ties among parents, children, and other adults. Table 4 presents evidence regarding the academic achievement of members of ethnic and racial minority groups and of children of immigrants. This table focuses on the predictors that are of central interest to us: race/ethnicity, immigration status, parent-child relations, and involvement in religious institutions. In Model 1, we include age and sex as control variables and the three racial/ethnic categories. Older students report lower average grades than younger students and females report higher average grades than males. Asians report
29
Table 4. Unstandardized Coefficients of Predictors of Average Grades. (Standard Errors in Parentheses)

Model 1 Age Sex Asian Black Latino First generation 10.5 generation Second generationi Family Income Parental education Parental occupation Closeness to parents Parental network involvement Parental stress on education Involvement in non-ethnic relig ion Involvement in ethnic religion 3.350** Constant (0.087) 0.704 CT 0.064 R2 -0.055** (0.005) 0.224** (0.018) 0.164** (0.048) -0.188** (0.021) -0.237** (0.029) Model 2 -0.056** (0.005) 0.223** (0.018) 0.099 (0.052) -0.188** (0.021) -0.280** (0.032) 0.155** (0.049) -0.027 (0.071) 0.093* (0.044) Model 3 -0.053** (0.005) 0.237** (0.018) 0.064 (0.059) -0.145** (0.022) -0.194** (0.034) 0.113* (0.055) -0.076 (0.076) 0.091* (0.043) 0.001** (0.000) 0.080** (0.005) 0.017** (0.004) Model 4 -0.044** (0.005) 0.243** (0.019) 0.107 (0.059) -0.140** (0.023) -0.172** (0.034) 0.171** (0.056) -0.042 (0.076)' 0.112* (0.044) 0.001** (0.000) 0.070** (0.005) 0.021** (0.005) 0.066** (0.012) 0.043** (0.005) Model 5 -0.041** (0.005) 0.236** (0.019) 0.092 (0.058) -0.134* (0.023) -0.179** (0.034) 0.149** (0.055) -0.042 (0.076) 0.111* (0.043) 0.001** (0.000) 0.065** (0.005) 0.017** (0.005) 0.059** (0.012) 0.041** (0.005) 0.070** (0.008) Model 6 Model 7 -0.040** (0.005) 0.229** (0.018) 0.092 (0.058) -0.156** (0.023) -0.127** (0.035) 0.152** (0.055) -0.047 (0.076) 0.121** (0.043) 0.001** (0.000) 0.061** (0.005) 0.017** (0.005) 0.051** (0.012) 0.036** (0.005) 0.068** (0.008) 0.022** (0.003)
3.368** (0.087) 0.704 0.066
2.773** (0.095) 0.679 0.134
2.185** (0.120) 0.671 0.154
-0.038** (0.005) 0.224** (0.018) 0.079 (0.058) -0.177** (0.023) -0.202** (0.038) 0.127** (0.055) -0.054 (0.076) 0.084 (0.044) 0.001** (0.000) 0.060** (0.005) 0.018** (0.005) 0.049** (0.012) 0.036** (0.005) 0.066** (0.008) 0.029** (0.003) 0.028** (0.006) 1.964** 1.926** 1.904** (0.122) (0.121) (0.121) 0.666 0.662 0.661 0.167 0.175 0.179
grades that are significantly higher than those of whites and blacks and Latinos report grades that are significantly lower. Model 2 includes the immigrant status of adolescents as predictors of school performance. Both first and second generation immigrants report significantly
30
higher grades than children of non-immigrants. The non-significant coefficient for members of the "1.5" generation may, once again, be due to the fact that there are so few respondents in this category. The coefficient for Asians is not statistically significant in Model 2 and there is a statistically significant difference of 0.065 in the unstandardized coefficients of Asians in Model 1 and model 2 (s(d) = 0.020). This suggests that the higher grades of Asians, compared to whites, are largely explained by the fact that most Asian adolescents tend to be children of immigrants. Similarly, for the other group with large numbers of children of immigrants, Latinos, we also see a significant change in coefficients (d = 0.043, s(d) = 0.014). Again, while it would be overly speculative to base claims about causation on these statistical findings alone, these results are consistent with the view that the school performance of Asian and Latino children is frequently enhanced by the fact that many of these children are children of immigrants, and immigrant children tend to out-perform the children of the native-born. In Model 3, we take family socioeconomic levels into consideration. The gaps between blacks and whites and between Latinos and whites narrow substantially (d = 0.043, s(d) = 0.011 for blacks and d = 0.086, s(d) = 0.014 for Latinos). Thus, a large portion of the explanation for the difference in school achievement between whites and members of these two minorities appears to be a matter of relative socioeconomic advantage. It is interesting that taking socioeconomic factors into account also shrinks the difference between whites and first generation immigrants, although the difference between the coefficients of first generation immigrants in models 2 and 3 is not statistically significant (d = 0.042, s(d) = 0.028). The most reasonable explanation for a possible socioeconomic source of an immigrant-native school performance gap is that many of the socioeconomic disadvantages of contemporary immigrants are due to the fact that they tend to be members of minority groups and that immigrant parents have relatively strong educational credentials once we control for the disadvantages of minority status (Portes & Rumbaut, 1996; Min, 1992 ). Even after we control for both minority status and family socioeconomic level, though, first and second generation immigrant children still appear to out-perform native born children. If there is something about being part of an immigrant family that promotes the school achievement of children in those families, what is it? The network closure model that we have described above suggests that close ties among parents, other adults, and children facilitate the reinforcing of sanctions to direct the behavior of children. However, we have already seen that immigrant families show relatively low levels of network closure and intimacy in immigrant families. Model 5 includes the measures of parental network closure and reported closeness between parents and children in the analysis.
31
Adolescents who report feeling close to their parents and adolescents with parents who maintain ties with children's friends and the parents of children's friends tend to do better in school than other adolescents do. Once we control for the fact that first generation immigrant children have relatively little intimacy with their parents and experience relatively little parental network closure, the achievement gap between first generation adolescents and children of the native-born grows even greater (between Model 3 and Model 4, d = 0.058, s(d) = 0.009). This is also the case for second generation immigrants (d = 0.020, s(d) = 0.005). It appears, in other words, that first generation and second generation immigrant students would out-perform the children of natives by an even greater margin if the former did not have so little intimacy with their parents and so little parental network closure. If ties to parents do not explain the achievements of immigrant children, but limit these achievements, to what can we attribute immigrant school performance? One possible answer is the mobility orientation of new adult arrivals in a country, an orientation they pass on to their children. Model 5, therefore, includes parental stress on higher education as a predictor of the average grades of children. The importance of higher education for parents, indicated by the extent of their disappointment if their children do not graduate from college, is indeed strongly and positively related to the school performance of children. Moreover, our findings here are consistent with the view that one reason immigrant children out-perform the children of natives is that the parents of the former place more importance on education as an avenue of advancement for children. There is a significant decrease in the coefficient of first generation immigrant children (d = -0.022, s(d) = 0.009), indicating that part of the association between immigrant status and educational achievement can be statistically explained by the higher educational aspirations of immigrant parents. Relations between immigrant children and their parents, then, do not appear to be irrelevant to the school performance of the children. Being close to parents, having parents who are closely connected to friends and other parents, and high parental educational aspirations are all positively related to the educational achievement of adolescents in general and to the achievements of adolescents from immigrant families in particular. Still, while immigrant families do have norms consistent with school performance, they do not exhibit the kinds of network ties that the parental network closure model of social capital posits as means of maintaining and enforcing those norms. Thesefindings,in themselves, suggest that some reconceptualization of social capital may be needed to explain the academic outcomes of immigrant children and of ethnic groups closely associated with contemporary immigration.
32
Ethnically-based religious institutions are also positively related to school performance, as we see in Model 6. It would appear that any kind of religious involvement is associated with academic achievement. Ethnically based institutions, though, are the ones most likely to affect the lives of children from immigrant families. A significant part of the superior performance of first generation immigrants adolescents can be attributed to the fact that these adolescents tend to be more active in ethnically based religions (between Model 6 and Model 7, d = 0.045, s(d) = 0.004). The coefficient for second generation immigrants also decreases by a statistically significant amount between Models 6 and 7 (d = 0.037, s(d) = 0.009). Moreover, there ceases to be any significant difference in school performance between second generation immigrants and children of natives after we control for involvement in ethnic religious institutions. While close ties to parents and supervision through parental network ties do not seem to promote the academic achievement of immigrant adolescents, there is some evidence that ethnically based religious institutions do help to fulfill this function.
CONCLUSIONS
This study, by itself, cannot provide definitive claims regarding the nature of the relationship between immigrant family status and family network closure or regarding the causes of immigrant school achievement. As we have observed above, one should always exercise caution in making causal claims based on cross-sectional data. It is, for example, entirely possible that high-achieving students self-select into religious activities, notwithstanding the social support mechanisms we discuss above. Along the same lines, some types of families may be more likely to involve their children in religious organizations than others. In addition, one may argue that highly motivated immigrant children bring social capital to religious organizations, as well as obtain capital from the organizations. Nevertheless, we think the study does offer some interesting implications for future research regarding the concept of social capital, as applied to the school outcomes of immigrant children. One of the central implications of this study is that the form of social capital that has been most precisely described, tight networks that establish norms and reinforce sanctioning, does not explain academic achievement among children of immigrant families. These families, in fact, show fewer and looser connections among parents, children, and other adults than other families do. Rather than discard the idea of social capital, though, we have attempted to address this problem by defining the idea more carefully. We
33
distinguish between the social capital of network closure and institutional social capital. The former is produced by close and closed sets of ties among individuals that facilitate the cultivation of financial or human capital. The latter is produced by formal institutions or organizations that provide support and control leading to the cultivation of financial or human capital. We should be careful to differentiate our distinction between these two forms of social capital from James Coleman's own thoughts regarding the institutional fostering of social capital. In his work on Catholic and private schools, Coleman did recognize that desirable social outcomes can result from the support and direction provided by the communities surrounding these schools (Coleman, 1990b, c; Coleman & Hoffer, 1987). In Catholic schools, in particular, Coleman argued that religious institutions could help to create social capital by countering the trends toward age-segregation in modern education by creating ties between children in schools and adults in religious communities (Coleman, 1990c). However, these views on educational and religious institutions were actually extensions of Coleman's individualistic network closure approach. In this approach, parental involvement in school or community extends ties among parents and children into extra-familial settings, so that adults in the church or school "as well as parents" are bound in broad, but still relatively closed networks (see, in particular, Coleman's discussion of functional communities with intergenerational closure in Coleman, 1990b). For the immigrant and ethnic minority children in this study, institutional settings are associated with the generation of human capital (educational performance), even though familial network connections are relatively open in two senses. The first sense is that of parental involvement with the friends of their children and parents of the friends of their children. The second sense is that of ties between institutional involvement and familial network ties. As we saw in Table 3, closeness to parents and parental involvement in the social networks of children are associated with children's participation in non-ethnic religious institutions, not with participation in the ethnic religious institutions that are most relevant to immigrant children and children of minority groups. Not only are the social ties of immigrant Asian and Latino children relatively open within their families, these ties are relatively open between immigrant families and ethnic religious institutions. The parental network closure model simply does not fit the phenomenon of immigrant children in American schools. If extra-familial institutions can provide an independent source of social capital, how does this happen? Our data do not provide insight into this question. In theory, though, we can identify at least two possible answers. It may be that ethnically based religious institutions (or other key group institutions)
34
provide alternative social networks. Adults and children active in the temple or mosque, rather than parents, establish and reinforce norms. This would be similar to a form of compensatory social capital that Coleman (1990c) does touch on, when he describes the social capital generated in Catholic schools as particularly important for children from economically disadvantaged or single-parent families. This explanation of the role of ethnic religious participation calls attention, though, to the need to give greater scrutiny to institutional settings and relationships between institutional settings in formulating a theory of social capital. Under what conditions is the family a source of productive behavior? Under what conditions do institutions outside the family provide such sources? How do institutions shape and direct network ties among individuals? The second theoretically possible answer to the question of how extra-familial institutions can provide social capital is that network ties among members of a social group are less important than the members' common normative orientations. We have seen that immigrant parents do apparently tend to place more emphasis on educational attainment that non-immigrant parents do. Moreover, parental stress on education does have a positive association with the educational performance of children that is largely independent of reported intimacy between children and parents and involvement of parents in the social networks of children. The nature of the ties that convey norms and values, then, may be less important than the content of the norms and values. If the adults and children that communicate with an adolescent all value education highly, it may not be critical for all those adults and children to have close ties or even know each other, as long as they are all repeating and reinforcing the same normative messages. From this point of view, the positive influence of an institution would depend less on network ties among members of the institution than on the common normative commitment that binds members to an organization or a social group. The term "social capital" has come into such common usage in the social sciences that efforts to describe and define it precisely have a particular theoretical urgency. Future investigation of the nature of social capital, in our view, would do well to concentrate on the role of institutional settings and on how these settings produce outcomes. Do the settings shape network ties among individuals or do they produce environments in which the same normative messages are repeated and emphasized? On the issue of immigrant academic achievement, in particular, we should recognize that there are special problems in applying the network closure model to people who are uprooted, dislocated, and struggling to establish themselves in a new homeland.
Social Capital and Immigrant
Children's
Achievement
35
REFERENCES
Astone, N. M., Nathanson, C. A., Schoen, R., & Kim, Y. J. (1999). Family Demography, Social Theory, and Investment in Social Capital. Population and Development Review, 25, 1-19. Banks, E. (1997). The Social Capital of Self-Help Mutual Aid Groups. Social Policy, 28, 30-38. Bankston, C. L. (1995). Gender Roles and Scholastic Performance Among Adolescent Vietnamese Women: The Paradox of Ethnic Patriarchy. Sociological Focus, 28, 161-176. Bankston, C. L. (1997). Education and ethnicity: Community and academic performance in an urban Vietnamese village. In: L. Weiss & M. S. Seller (Eds), Beyond Black and White: New Faces in United States Schools (pp. 207-230). New York: State University of New York Press. Bankston, C. L., & Zhou, M. (2000). De facto Congregationalism and Socioeconomic Mobility in Laotian and Vietnamese Communities: A Study of Religious Institutions and Economic Change. Review of Religious Research, 41, 453-470. Bankston, C. L., & Zhou, M. (1995a). Religious participation, Ethnic Identification, and Adaptation of Vietnamese Adolescents in an Immigrant Community. The Sociological Quarterly, 36, 523-535. Bankston, C. L., & Zhou, M. (1995b). The Effects of Minority Language Literacy on the Academic Achievement of Vietnamese Youth in New Orleans. Sociology of Education, 68, 1-17. Bankston, C. L., & Zhou, M. (1996). The Ethnic Church, Ethnic Identification, and the Social Adjustment of Vietnamese Adolescents. Review of Religious Research, 38, 18. Bankston, C. L., Caldas, S. J., & Zhou, M. (1997). The Academic Achievement of Vietnamese American Students, Ethnicity as Social Capital. Sociological Focus, 30, 1-16. Barton, J. L. (1975). Peasants and Strangers: Italians, Rumanians, and Slovaks in an American City, 1890-1950. Cambridge, MA: Harvard University Press. Berrol, S. C. (1995). Growing Up American: Children in America, Then and Now. New York: Twayne. Borjas, G. J. (1999). Heaven's Door: Immigration Policy and the American Economy. Princeton, N.J.: Princeton University Press. Clark, E. E., & Ramsay, W. (1990). The Importance of Family and Network of Other Relationships in Children's Success in School. International Journal of Sociology of the Family, 20, 237-254. Clogg, C. C , Petrovka, E & Haritou, A. (1995). Statistical Methods for Comparing Regression Coefficients Between Models. American Journal of Sociology, 100, 1261-1293. Coleman, J. S. (1988). Social Capital in the Creation of Human Capital. American Journal of Sociology, 94, S95-S120. Coleman, J. S. (1990a). Foundations of Social Theory. Cambridge, MA: The Belknap Press of Harvard University Press. Coleman, J. S. (1990b). Schools and the Community They Serve. In: J. S. Coleman (Ed.), Equality and Achievement in Education (pp. 315-324). Boulder, Co.: Westview Press. Coleman, J. S. (1990c). Families and Schools. In: J. S. Coleman (Ed.), Equality and Achievement in Education (pp. 325-340). Boulder, Co.: Westview Press. Coleman, J. S., & Hoffer, T. (1987). Public and Private High Schools: The Impact of Communities. New York: Basic Books. Ebaugh, H. R & Chafetz, J. S. (2000a). Structural Adaptations in Immigrant Congregations. Sociology of Religion, 61, 135-154.
36
Ebaugh, H. R., & Chafetz, J. S. (2000b). Structural Adaptations to the Immigrant Context. In: H. R. Ebaugh & J. S. Chafetz (Eds), Religion and the New Immigrants: Continuities and Adaptations in Immigrant Congregations (pp. 347-370). Walnut Creek, Ca.: Altamira Press. Ebaugh, H. R., & Chafetz, J. S. (2000c). Passing it on: The Second Generation. In: H. R. Ebaugh & J. S. Chafetz (Eds), Religion and the New Immigrants: Continuities and Adaptations in Immigrant Congregations (pp. 431-446). Walnut Creek, Ca.: Altamira Press. Feld, S. L. (1981). The Focused Organization of Social Ties. American Journal of Sociology, 86, 1015-1035. Fuligni, A. J. (1997). The Academic Achievement of Adolescents from Immigrant Families: The Roles of Family Background, Attitudes, and Behavior. Child Development, 68, 351-363. Furstenberg, F., & Hughes, M. E. (1995). Social Capital and Successful Development among At-Risk Youth. Journal of Marriage and the Family, 57, 580-592. Gans, H. (1992). Second Generation Decline: Scenarios for the Economic and Ethnic Futures of Post-(1965 American Immigrants. Ethnic and Racial Studies, 15, 173-192. Geertz, C. (1973). The Interpretation of Cultures. New York. Greeley, A. (1997). Coleman Revisited, Religious Structures as a Source of Social Capital. The American Behavioral Scientist, 40, 587-594. Haddad, Y. Y., & Lummis, A. T. (1987). Islamic Values in the United States: A Comparative Study. New York: Oxford University Press. Hao, L., & Bonstead-Bruns, M. (1998). Parent-Child Differences in Educational Expectations and the Academic Achievement of Immigrant and Native Students. Sociology of Education, 71, 175-198. Hurh, W. M., & Kim, K. C. (1990). Religious Participation of Korean Immigrants in the United States. The Journal for the Scientific Study of Religion, 29, 19-35. Isaacs, H. R. (1975). Idols of the Tribe: Group Identity and Political Change. Cambridge, Mass.: Harvard University Press. Kashima, T. (1977). Buddhism in America: The Social Organization of an Ethnic Religious Institution. Westport, CT: Greenwood Press. Kibria, N. (1993). Family Tightrope: The Changing Lives of Vietnamese Americans. Princeton, N.J.: Princeton University Press. Kurien, P. (1998). Becoming American by Becoming Hindu: Indian Americans Take Their Place at the Multicultural Table. In: R. S. Warner & J. G. Wittner (Eds), Gatherings in Diaspora: Religious Communities and the New Immigration (pp. 37-70). Philadelphia, Temple University Press. Kwon, V. H. (1997). Entrepreneurship and Religion: Korean Immigrants in Houston, Texas. New York: Garland Publishing. Landale, N. S. (1996). Immigration and the Family: An Overview. In: A. Booth, A. C. Crouter & N. Landale (Eds), Immigration and the Family: Research and Policy on U.S. Immigrants (pp. 281-291). New Jersey: Lawrence Erlbaum. Leon, L. (1998). Born Again in East Los Angeles: The Congregation as Border Space. In: R. S. Warner & J. G. Wittner (Eds), Gatherings in Diaspora: Religious Communities and the New Immigration (pp. 163-196). Philadelphia, Temple University Press. Loury, G. C. (1977). A Dynamic Theory of Racial Income Differences. In: P. A. Wallace & A. M. La Mond (Eds), Women, Minorities, and Employment Discrimination, (pp. 153-186). Lexington, MA: Heath. Loury, G. C. (1981). Intergenerational Transfers and the Distribution of Earnings. Econometrica, 49, 843-867.
Social Capital and Immigrant
Children's
Achievement
37
McClanahan, S., & Sandefur, G. (1994). Growing Up With a Single Parent: What Hurts, What Helps. Cambridge, MA: Harvard University Press. Min, P. G. (1992). The Structure and Social Functions of Korean Immigrant Churches in the United States. International Migration Review, 26, 1370-1394. Ogbu, J. U. (1974). Tlie Next Generation: An Ethnography of Education in an Urban Neighborhood. New York: Academic Press. Ogbu, J. U. (1989). Cultural Models and Educational Strategies of Non-Dominant Peoples. New York: City College Workshop Center. Ogbu, J. U. (1990). Minority Education in Comparative Perspective. Journal of Negro Education, 59, 45-57. Perlmann, J., & Waldinger, R. (1997). Second Generation Decline? Immigrant Children Past and Present - A Reconsideration. International Migration Review, 31, 893-922. Portes, A. (1998). Social Capital: Its Origins and Application in Modern Sociology. Annual Review of Sociology, 24, 1-24. Portes, A., & Rumbaut, R. G. (1996). Immigrant America: A Portrait (revised edition). Berkeley: University of California Press. Portes, A., & Sensenbrenner, J. (1993). Embeddedness and Immigration: Notes on the Social Determinants of Economic Action. American Journal of Sociology, M98,1320-1350. Portes, A., & Zhou, M. (1993). The New Second Generation: Segmented Assimilation and Its Variants. The Annals of the American Academy of Political and Social Science, 530(November), 74-96. Rumbaut, R. G. (1994). The Crucible Within: Ethnic Identity, Self-Esteem, and Segmented Assimilation among Children of Immigrants. International Migration Review, 28,748-794. Rumbaut, R. G. (1995). The New Californians: Research Findings on the Educational Progress of Immigrant Children. In: R. G. Rumbaut & W. A. Cornelius (Eds), California's Immigrant Children: Theory, Research, and Implications for Educational Policy. San Diego: University of California, San Diego, Center for U.S.-Mexican Studies. Rumbaut, R. G. (1998). Transformations: The Post-Immigrant Generation in an Age of Diversity. Paper presented at the Annual Meeting of the Eastern Sociological Society (Philadelphia, Penn., March). Sanders, M. (1998). The Effects of School, Family, and Community Support on the Academic Achievement of African American Adolescents. Urban Education, 33, 385-409. Steinberg, L. (1996). Beyond the Classroom. New York: Simon & Schuster. Suarez-Orozco, C , & Suarez-Orozco, M. M. (1995). Transformations: Migration, Family Life, and Achievement Motivation among Latino Adolescents. Stanford, Calif.: Stanford University Press. Sue, S., & Okazaki, S. (1990). Asian American Educational Achievement: A Phenomenon in Search of an Explanation. American Psychologist, 45,913-920. Sullivan, K. (1998). Immigrant Religion and the Challenge of the Second Generation: Religion as the Anti-Dote to 'Amoral' U.S. Society. Paper delivered at the annual meeting of the Society for the Scientific Study of Religion. Montreal, Canada. Sung, B. L. (1987). The Adjustment Experience of Chinese Immigrant Children in New York City. Staten, Island, N.Y.: Center for Migration Studies. Sun, Y. (1998). The Academic Success of East-Asian-American Students: An Investment Model. Social Science Research, 27, 432-456. Teachman, J. D., Paasch, K., & Carver, K. (1996). Social Capital and Dropping Out of School Early. Journal of Marriage and the Family, 58, 773-783.
38
Teachman, J. D., Paasch, K., & Carver, K. (1997). Social Capital and the Generation of Human Capital. Social Forces, 75, 1343-1359. Thomas, W. I., & Znaniecki, F. (1927). The Polish Peasant in Europe and America. New York: A. A. Knopf. Waters, M. (1996). Immigrant Families at Risk: Factors that Undermine Chances of Success. In: A. Booth, A. C. Crouter & N. S. Landale (Eds), Immigration and the Family: Research and Policy on U.S. Immigrants (pp. 79-87). New Jersey: Lawrence Erlbaum. Zhou, M. (1992). Chinatown: The Socioeconomic Potential of an Urban Enclave. Philadelphia: Temple University Press. Zhou, M. (1997a). Growing Up American: The Challenge Confronting Immigrant Children and Children of Immigrants. Annual Review of Sociology, 23, 63-95. Zhou, M. (1997b). Social Capital in Chinatown: The Role of Community-Based Organizations and Families in the Adaptation of the Younger Generation. In: L. Weiss & M. S. Seller (Eds), Beyond Black and White: New Voices, New Faces in the United States Schools (pp. 181-206). Albany, NY: SUNY Press. Zhou, M. (Forthcoming). Second-Generation Fate: Progress, Decline, Stagnation? In: R. Waldinger (Ed.), Strangers at the Gate: New Immigrants in Urban America. Berkeley: University of California Press. Zhou, M & Bankston, C. L. (1994). Social Capital and the Adaptation of the Second Generation: The Case of Vietnamese Youth in New Orleans. International Migration Review, 28, 821-845. Zhou, M., & Bankston, C. L. (1998). Growing Up American: How Vietnamese Children Adapt to Life in the United States. New York: Russell Sage Foundation. Zhou, M., Bankston, C. L., & Kim, R. (Forthcoming). Rebuilding Spiritual Lives in the New Land: Religious Practices among Southeast Asian Refugees in the United States. In: J. H. Kim & P. G. Min (Eds), Asian Immigration and Transplanting and Transforming Religions. Walnut Creek: Altamira Press.
APPENDIX: VARIABLES IN THE ANALYSIS

Race/Ethnicity: 0 (does not belong to the group in question), 1 (belongs to the group). First generation: Born outside the U.S. and entered at age 5 or after. Coded as Oor 1. 1.5 generation: Born outside the U.S. and entered after age 5. Coded as 0 or 1. Second generation: Born inside the U.S. Coded as 0 or 1. Single female family: Lives with a woman, without a man in the household. Coded as 0 or 1. Single male family: Lives with a man, without a woman in the household. Coded as 0 or 1.
39
Family income: Yearly dollar amount of income reported by parents. Parental education: Coded as 1 (8th grade or less), 2 (more than 8th grade but less than high school), 3 ( finished high school or received a GED), 4 (completed a trade or vocational school education after high school), 5 (some college), 6 (college graduate), 7 (graduate or professional degree). For two-parent families, the variable was the average of both parents. Parental occupation: 8 categories, from laborer to manager-professional. For two-parent families, the variable was the average of both parents. Parent-child closeness: Children's responses to the question "how close do you feel to your mother?" and "how close do you feel to your father?" Possible answers ranged from 1 (not at all) to 5 (very much). For two-parent families, the variable was the average of both parents. Parental network closure: A scale created from three items: (1) whether parents had met the best friends of their children (coded as 0 or 1), (2) whether parents had met the parents of the best friends of their children (coded as 0 or 1), and (3) how many parents of the friends of their children parents had met with in the previous six.weeks (from 0 to 6). The scale created by adding these items had values of 0 to 8. Parental stress on education: Parents' reports of how disappointed they would be if their children did not graduate from college (1, not disappointed; 2, somewhat disappointed; 3, very disappointed). Involvement in religious institutions: A scale created from three items: (1) church attendance (0, never attend, to 3, attend more than once a week), (2) importance of religion to the respondent (0, not important at all, to 3, very important), and (3) how often the respondent takes part in church activities (coded the same as attendance). We constructed separate scales for religious institutions with close conceptual connections to some ethnic group membership and for those without explicit ethnic connections. Average grades: the average of the grades respondents reported receiving the previous school year in English or language arts, math, history or social studies, and science. The values are ranked from 1 (D or F in all subjects) to 4 (A in all subjects).
FAMILY AND NON-FAMILY ROOTS OF SOCIAL CAPITAL AMONG VIETNAMESE AND MEXICAN AMERICAN CHILDREN
Kimberly A. Goyette and Gilberto Q. Conchas
ABSTRACT
This research considers the influence of both familial and non-familial social capital on the study habits of two recent immigrant groups: Vietnamese Americans and Mexican Americans. Using the 1988-1990 National Educational Longitudinal Study and qualitative data from a twoyear study of minority students' experiences in a California high school, we find that Vietnamese students study more than do Mexican students because they have more positive peer associations and more supportive relationships with teachers than do Mexican students. Contrary to popular belief, this research suggests that family practices do not sufficiently explain why Vietnamese students study more than Mexican students.
INTRODUCTION
Sociologists have long argued that unequal resources generate disparity in school achievement. School and family resources, in particular, are often considered necessary for positive educational outcomes. Useful resources may be evident, like
Schooling and Social Capital in Diverse Cultures, Volume 13, pages 41-72. Copyright 2002 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0817-6 41
42
KIMBERLY A. GOYETTE AND GILBERTO Q. CONCHAS
financial support, or they could be less tangible, such as norms, encouragement, and information gained from relationships and social networks. Following Bourdieu (1977) and Coleman (1988b), these less tangible resources have been called "social capital." Access to high quality social capital is a frequently invoked explanation for the educational success of students. It has been used to explain students' high school drop-out rates (Astone & McLanahan, 1991; Carbonaro, 1998; Coleman, 1987; Teachman et al., 1997, 1996), math achievement scores (Morgan & S0rensen, 1999), and grade point averages (Carbonaro, 1998; Valenzuela & Dornbusch, 1994). Although social capital has a wide-ranging definition and has been characterized in several different ways (Bourdieu, 1977; Burt, 1992; Granovetter, 1974; Loury, 1977; Portes, 1998), sociologists who study education often base their research on the work of Coleman (1987), who focused on the information, support, and supervision that closely-knit networks of relationships provide. Coleman was the first to formulate the notion of "social closure." Coleman noted that it was not just the sum of relationships, but also the interconnectedness of those relationships that mattered for children's educational outcomes. Many researchers who study social capital consider the family to be the most important supplier of it for children. Following Coleman, research on social capital focuses on the influence that parents, through their connections with children, schools, and other parents, have on the educational achievement of the general school population (e.g. Carbonaro, 1998; Coleman, 1988a; Hagan et al., 1996; Mehan et al., 1996; Teachman et al., 1997, 1996). Thus far, research on children's social capital has been less concerned with social capital that does not originate in the home; that is, social capital that comes from schools and relationships with peers.1 However, parents are not the only sources of children's social capital. Children spend large portions of their days in settings without their parents. Sources of "non-familial" social capital include friends, teachers and other school officials, and communities. Ignoring these non-familial sources of social capital is problematic because discrepancies in children's educational outcomes are attributed solely to the actions of parents without researchers also scrutinizing the practices of schools and communities. Research on social capital to date often concentrates on the school population in general and does not focus on particular sub-groups. This may not provide researchers with a complete understanding of how social capital operates among all students, however. For example, Stanton-Salazar and Dornbusch (1995) contend that children from low-SES and minority families make use of information and support from peers and institutions even more than of that provided
Family and Non-Family Roots of Social Capital
43
by families. Further, recent immigrant parents may have little knowledge of how the U.S. educational system works, and minority parents may feel alienated from it (Lareau & Horvat, 1999; Mehan et al., 1996; Valenzuela, 1999). Therefore, children may not always rely on parents for the resources they need to succeed. In this research, we choose two groups whose academic performance has been attributed to family relationships to illustrate the importance of considering the contribution of non-familial sources of social capital to children's achievement. These two groups are Vietnamese Americans and Mexican Americans. Vietnamese Americans and Mexican Americans provide an interesting comparison because these groups are socioeconomically and demographically similar. Vietnamese and Mexican Americans are both groups composed of many first- and second-generation immigrants, and both have less English proficiency and lower socioeconomic status than do whites. Despite this similarity, these two groups have disparate educational outcomes. Vietnamese students maintain better grades, graduate from high school at higher rates, and enroll in college more than do Mexican American students (Mehan et al., 1996). Popular press and academics alike pose the question: Why is this? Why do two racialized groups with similar demographic and socioeconomic profiles perform differently in school? A popular explanation for these differences is that Vietnamese families promote children's success through emphasis on Confucian values (Caplan et al., 1991; Nash, 1987), while Mexican families transmit norms that inhibit success. Some speculate that a tightly-knit family that demands a child's loyalty limits the commitment that Mexican American children make to further their education (Alvirez & Bean, 1976; Carter & Segura, 1979; Parsons, 1949). In both the research examining Vietnamese students' success and that exploring Mexican American students' educational achievements, the norms conveyed through family social capital are said to account for students' school success or failure. In this research, we show that it may be as or more important to consider the resources students gain from peers and from schools in accounting for the disparities between these two groups. Unlike other research on social capital, which leaves the mechanism through which social capital influences school outcomes unspecified, in this paper we measure the effects of social capital on a behavior that many believe is associated with positive educational outcomes. This behavior is the time students spend on homework outside of school per week. The investigation of time spent studying is consistent with the work of Coleman (1987) who argued that the social capital provided by Catholic schools fostered norms promoting studiousness and academic challenge among the students.
44
Throughout this research, we look at how differences in both the familial and non-familial social capital held by Vietnamese American and Mexican American students explains the variation in time spent on homework outside of school. Our results indicate that it is not only the involvement of families that accounts for the differences between Vietnamese and Mexican American study habits, but also the social capital that is available to students through school environments and interactions with peers and teachers.
DATA AND METHODS

In this analysis, we use a combination of quantitative and qualitative methods. The quantitative information is drawn from the National Educational Longitudinal Study (NELS) 1988-1990 panel, collected for the National Center for Education Statistics (NCES) by the National Opinion Research Center. In 1988, the NCES surveyed 24,599 United States eighth-graders. These same respondents were re-interviewed in 1990, 1992, and again in 1994. Information was collected from the sampled students and their parents, teachers, and school principals. The NELS 1988-1990 panel contains an over-sample of both Asian American and Hispanic students, including a sample of about 165 Vietnamese and approximately 1,250 Mexican American students. Although the number of Vietnamese students is comparatively small, it is a larger number than that available in any other data set. Because there are so few Vietnamese students, standard errors for this group will be large and statistical significance harder to achieve. Consequently, estimates of significance in this research may be conservative. We include responses from both 1988 and 1990 in order to capitalize on the parents' interview of 1988 (there was no parent interview in 1990) and to capture the school characteristics of students in 1990, when time spent studying outside of school is measured. Because the sample for analysis is a panel, the results can only be generalized to those students who were eighthgraders in 1988 who were also in school in 1990. The qualitative data is derived from a two-year research project (1996-1998) of the sociocultural mechanisms that construct school success for eighty racial minority students in a large, urban, comprehensive high school in California. Minority groups in this school include Latino, Vietnamese, and African American students. This particular paper draws from interviews and observations of the schooling experiences of twenty-seven Vietnamese students and thirteen Mexican American students. The sample consists of students from very low SES backgrounds, both boys and girls, immigrant and native-born. It also includes interview data from forty-five teachers and administrators at the high school.
45
We first present the differences in time that Vietnamese and Mexican American students spend on homework outside of school. We then discuss some possible explanations for these differences, focusing particularly on the roles that familial and non-familial capital may play in influencing study behavior. To investigate whether or not differences in social capital account for the variation in time spent studying between Vietnamese and Mexican American students, the analysis portion of this paper is divided into two parts. First, we compare the potential sources of social capital for Vietnamese and Mexican American high school students and use multivariate Ordinary Least Squares (OLS) regression models to explore the extent to which these sources of social capital, net of students' socioeconomic and demographic characteristics, influence students' time spent on homework. In the second section, using qualitative data, we specifically concentrate on students' interpretations of and reactions to the manner in which the resources derived from their relationships within one high school affect their school success.
DIFFERENCES IN STUDENTS' STUDYING BEHAVIOR

Using the NELS data, we present the mean number of hours that Vietnamese and Mexican American students spend on homework outside of school in the first line of Table 1. Vietnamese students spend approximately two hours more per week on homework (5.9 hours) than do Mexicans American students (3.9). One popular explanation for this is that Asian American families value schooling more than do Mexican families. Popular culture and academic research cite two mechanisms through which Asian American families are perceived to affect children's school norms and behaviors: parental expectations and parental satisfaction with school performance. One common belief about Asian American families is that Asian American parents have higher educational expectations of their children than do other parents. For instance, to explain the differences she sees between Asians and other students, one Chinese counselor from the high school we studied, Baldwin High School, said:
It all starts in the family... That is where students get their foundation . . . A lot of it has to do with the h o m e . . . Like when I think of the Chinese culture, even though they may not be at the higher or even the middle class economic level, there is this real thing about education. I remember once sitting at the dinner table with my two brothers and my parents, and my brother said he wanted to be a truck driver. I could just see the look in my father's face and he just said, "No, you are going to college!" (4/24/98, p. 26).
46
Some academic research supports the view expressed by this quote. Researchers have noted that Asian American parents hold higher expectations for children than do whites and other minorities (Chen & Stevenson, 1995; Goyette & Xie, 1999). These high expectations are said to compel students to work hard in order to excel in school. Another indicator of family values found in the academic literature is satisfaction with children's school performance. Some research suggests that Asian
Table 1.
Demographic and Socioeconomic Characteristics by Ethnicity, NELS 1988-1990 Panel.

Vietnamese Mexicans 3.9 3.9
Hours Spent on Homework Outside of School* Std. Dev. Immigration Generationp First generation Second generation Third generation Missing SES* Std. Dev. Father's Education Less than high school High school graduate College graduate Missing Mother's Education Less than high school High school graduate College graduate Missing N
5.9 4.8
67.8 15.5 0.4 16.3 0.44 0.93
13.0 35.4 38.8 12.8 -0.71 0.68
8.0 22.9 27.6 41.5
40.2 31.9 7.0 21.0
25.7 22.1 11.2 41.0 163
40.2 36.3 6.0 17.5 1,237
* These variables are means, not percentages as in the rest of the table. Missing values are excluded for the calculation of means. p This information comes from parents' questionnaires. Note: All variables except hours spent on homework are measured at the base year. Hours spent on homework are measured at the first follow up. Results are weighted according to the sample weights provided by NELS.
47
American parents are less satisfied with children's school performance than are white parents (Chen & Stevenson, 1995). Further, commentators both popular and academic, cite the pressure that Asian American children feel from their parents to succeed in school (Schneider & Lee, 1990). In contrast to Asian Americans, Mexican American families are often "blamed" for the academic failures of their children. Mexican families are said to have tight kin networks. These close bonds between nuclear and extended family members are perceived to limit the child's school success by placing obligations to the family unit above personal achievement. When conflicts arise between family responsibilities, such as providing income or childcare for other members, and school commitments, schoolwork suffers (Heller, 1966; Horowitz, 1981; Kuvelesky & Patella, 1971; SanchezJankowski, 1991; Vigil, 1988). While the popular and, to some extent, academic press places responsibility for children's educational outcomes on Vietnamese and Mexican American families, little research has identified the practices within these families that promote or inhibit educational success. One concept that has been used to explore the involvement of the family in the educational achievement of students is "social capital." Social capital is the useful resources that are gained from relationships with other people. In this case, these resources could be information, supervision and support, and/or norms that promote school success. Using this concept, popular perceptions of the differences between the two groups would suggest that Vietnamese families provide children with better resources with which to navigate their academic careers than do Mexican families. In short, Vietnamese families provide more advantageous social capital. Of course, there are other explanations for the differences in study habits between Vietnamese and Mexican American students. Although both groups have low socioeconomic status (SES) compared to whites, Vietnamese have higher average family SES than do Mexican Americans. This is, in part, due to the relatively high levels of education that Vietnamese parents hold. And, although Vietnamese and Mexican Americans are often recent immigrants, Vietnamese students are more likely than are Mexicans to be of the first or second immigration generation. Researchers have speculated that "immigrant optimism" may account for the high achievement of recently immigrated students who believe their effort will be fairly rewarded in the U.S. (Ogbu, 1991; Sudrez-Orozco & Suarez-Orozco, 1995). Descriptive information concerning these differences between Vietnamese and Mexican Americans in the NELS sample is included in Table 1. These socioeconomic and demographic differences are included in all multivariate models as controls.
48
A final, under-explored source of differences in study habits between Vietnamese and Mexican Americans may be found in what we call "non-familial" social capital. Non-familial social capital consists of those academically useful resources that are available to children from their school environments and relationships with peers, teachers, and others outside their families. Past research gives us reason to believe that this non-familial social capital may influence the study habits of these two groups. In this research, we divide non-familial social capital into two categories. We first explore social capital differences that may result from macro-institutional processes such as inequality among schools. It is possible that the school context may influence the information and support available to students; however, evidence concerning the impact of school environments on students' motivation and achievement is mixed. While many researchers find minimal effects of school context on students' achievement (Caterall, 1998; Jencks, 1972; Rumberger, 1987; Thornton & Eckland, 1980), others find that students who attend schools with predominantly low-SES students and high concentrations of minorities have less accurate information about educational and occupational opportunities (Conchas in press; Hoelter, 1982; Kemple & Snipes, 2000; Mehan et al., 1996; Stanton-Salazar & Dornbusch, 1995; Weiler, 2000).2 Differences in the acquisition of social capital may occur between schools. Other sources of non-familial social capital are found within schools. Students in the same schools may maintain different relationships with peers, and with teachers and other school officials. These relationships can be conceived as sources of social capital because peers and teachers provide encouragement, support, supervision, and information to the student. For example, students who have many friends who drop out of school are unlikely to receive support and encouragement from these friends to study. Further, these friends are unlikely to have academically useful information. Students whose friends consider studying important are likely to have support, encouragement, and information that motivate and enable them to study hard. The effects of peers on students' academic behaviors are well known. Educational research has long stressed how student's interactions with one another, and the meanings associated with these interactions, significantly shape patterns of academic behavior in school. Scholars posit that the peer group can serve as a mediating factor, either promoting compliance with or resistance to a school's rules for success (Conchas in press; De Vos, 1982; Lee, 1996; Mehan et al., 1994). For example, students who have many friends who drop out of school will be negatively influenced to do the same (Conchas in press; Fine, 1991; Gandara, 1995; Mehan et al., 1996; Valenzuela, 1999;
49
Vigil, 1988). Steinberg et al. (1992) further assert that, although peer groups are an important influence on all students' educational outcomes, for some racial groups, peers are an even more important influence on high school students than parents are. They contend that Asian Americans, who often associate with other high-achieving Asian American students, get most of their information, norms, and support from peers (Lee, 1996). Relationships with school personnel are another source of social capital. Teachers encourage students whom they believe are talented or hard working (Conchas, 1999; Farkas et al., 1990; Mehan et al., 1996; Stanton-Salazar, 1997; Stanton-Salazar & Dornbusch, 1995). Principals and guidance counselors may provide information about college preparation courses, applications, and financing for students who do not have access to this information at home. Guidance counselors may take students on trips to colleges to help them make informed choices. Education personnel may choose students whom they believe worthy to "sponsor" and thus enable their educational achievement (Conchas in press; Mehan et al., 1996). This type of helpful "sponsorship" may depend on the race of the student. Teachers and other school personnel may judge students differently based on their race (Conchas in press; Matute-Bianchi, 1986; Mehan et al., 1996; Valenzuela, 1999; Wong, 1980). Asian American students who are perceived to be "model minorities" may benefit from positive stereotypes because teachers provide them with information they would not otherwise have (Conchas in press; Lee, 1996). While Asian American students may gain from school relationships, Angela Valenzuela (1999) posits that schools devalue the familial social capital that Mexican American students bring into schools. She argues that the schooling process ignores the rich cultural and linguistic practices of Mexican youth in favor of assimilationist policies and practices. Schooling, in turn, functions to subtract social capital from Mexican American students, leading to academic failure. In this research, we do not suggest that families have no influence on children's study habits. Rather, we illustrate that typical measures of social capital in families cannot entirely explain the difference between the educational behaviors of Vietnamese and Mexican American students. We acknowledge that familial influence on children's study habits may operate through channels other than those we typically characterize as social capital. Conceptualizations and measures of familial social capital may have to be adapted to fit the varied situations and experiences of diverse minority ethnic groups in order to be usefully applied to those communities. However, in this research, we do demonstrate that non-familial social capital plays a critical role in influencing students' studying behavior.
50
DESCRIPTIVE RESULTS: THE SOURCES OF SOCIAL CAPITAL

Familial Sources of Social Capital Although familial social capital has been shown to influence the math learning and drop-out behavior of white children (e.g. Carbonaro, 1998; Furstenberg & Hughes, 1995; Teachman et alM 1997,1996), most indicators of it cannot explain the better study habits of Vietnamese compared to Mexican American students. Table 2 shows that, by most measures, Vietnamese students have less familial social capital than do Mexican American students, when familial social capital is measured as the frequency of interactions between parents and children, parents and schools, and parents and other children's parents (e.g. Carbonaro, 1998; Teachman et al., 1997, 1996). The exception to this is that Vietnamese parents tend to have higher educational expectations of their children than do Mexican American parents. We include several measures of familial social capital in our analyses to try to tap a variety of dimensions of the concept. The first of these is the frequency of interactions between parents and children. To measure parent-child interaction, we present results from the parent's report of how often she or he talks to the child about school experiences, high school plans, and post-high school intentions by race. This measure of parent-child interaction is consistent with that used by Teachman et al. (1997, 1996). Using this measure, we note that approximately half of Vietnamese parents report infrequent discussions about school experiences with children, compared to less than a third of Mexican parents. Mexican parents are also far more likely to report regularly talking to their children about their high school plans, at 46% compared to 31% of Vietnamese parents. Finally, when we report how often parents talk to children about their post-high school plans, the same pattern results. Fewer Vietnamese than Mexican parents regularly talk to children. When these measures are combined in a summary index with each component equally weighted, the results are not surprising. The mean for Mexican parents is higher, at 1.6, than the Vietnamese parents' mean of 1.1. Frequency of, or at least opportunity for, interactions has also been measured by family structure (Astone & McLanahan, 1991; Furstenberg & Hughes, 1995; Teachman et al., 1997, 1996). McLanahan and Sandefur (1994) note that children living in single parent families have less social capital available to them because they lack the information, support, and supervision of the absent parent. In this respect, Vietnamese and Mexican families do not seem to differ much. Vietnamese families are slightly more likely to be intact (with both natural
51
Table 2. Familial Sources of Social Capital among Vietnamese and Mexican Students, NELS 1988-1990 Panel.
Vietnamese Family Structure Intact Non-intact Missing Talk to Child about School Experiences1' Regularly Not regularly Missing Talk to Child about High School Plans'1 Regularly Not regularly Missing Talk to Child about Post-High School Plans'' Regularly Not regularly Missing Parent-Child Interaction Index*p Std. Dev. Missing Belong to the PTAP Yes No Missing Attend PTA MeetingsP Yes No Missing Attend PTA-Sponsored Activitiesp Yes No Missing Volunteer at the School11 Yes No Missing Mexicans
65.8 26.1 8.1
70.5 27.1 2.4
38.8 49.5 11.7
58.3 28.7 13.0
30.7 59.5 9.8
46.2 43.9 9.9
25.2 65.0 9.8 1.1 1.2 11.3 9.8 75.5 14.7
38.9 51.2 9.9 1.6 1.2 13.3
12.0 72.3 15.7
25.5 59.4 15.1
32.1 52.9 15.0
12.1 73.2 14.7 9.0 75.6 15.5
16.0 68.3 15.7
10.9 73.4 15.7
52
KIMBERLY A. GOYETTE AND GILBERTO Q. CONCHAS Table 2. Continued.

Vietnamese Mexicans 0.8 1.1 16.6 2.5 1.5 31.4 0.7 14.6 30.8 43.7 10.2 1,237
Parent-School Interaction Index*? Std. Dev. Missing Mean Number of Child's Friends' Parents Known*? Std. Dev. Missing Parents' Educational Expectations for Child Less than high school High school graduate Some college College graduate Missing N
0.7 1.0 15.8 2.0 1.6 55.8 1.4 4.5 12.3 72.0 9.8 163
These variables are means, not percentages as in the rest of the table. Missing values are excluded for the calculation of means. p This information comes from parents' questionnaires. Note: All variables are measured at the base year. Results are weighted according to the sample weights provided by NELS.
parents present) than are Mexican families, but the difference is less than five percentage points. Relationships that result in advantageous social capital can also be measured by family-school interactions. Based on the measures found in other research on social capital (Teachman et al., 1997,1996), we measure family-school interactions with an index composed of four variables. These variables account for whether parents had volunteered at the school, attended a school event, joined the PTA, or attended a PTA meeting. On all measures, Vietnamese parents were somewhat less likely than Mexican parents to participate in school events. For example, although the difference in PTA membership across groups is small (10% of Vietnamese parents compared to 12% of Mexican parents), 16% of Mexican parents report attending a PTA-sponsored activity and almost 11% report volunteering at the child's school compared to 12% and 9% of Vietnamese parents. When all measures are combined in a single index (constructed such that if a parent answered "yes" to any of the questions, the parent was scored " 1 , " and the scores were summed), the Vietnamese parents achieve a mean score of 0.7, while the mean for Mexican parents is 0.8. Clearly, parental interaction with schools does not explain variation in study behavior
53
between Vietnamese and Mexican American students, as it is approximately equal. Mexican American students appear to have similar access to familial social capital through the interactions of parents with schools as do Vietnamese Americans. Another commonly used dimension of familial social capital is social closure. This is often measured as the number of the child's friends' parents that a parent knows (Carbonaro, 1998). Table 2 shows that Mexican parents are more involved in children's school communities than are Vietnamese parents. The mean number of friends' parents known by Vietnamese parents is 2.0, compared to 2.5 parents known by Mexican Americans. From this measure, it appears that Mexican children are more likely than are Vietnamese children to benefit from social closure in the school community.3 Finally, we include parental expectations of students. While almost all Vietnamese and Mexican American parents expect their children to graduate from high school, Vietnamese parents are far more likely to expect their children to graduate from college. More than 70% of Vietnamese parents report that they expect their child to complete college compared to 44% of Mexican parents. While Vietnamese parents do not interact more with children, teachers, or other children's parents than do Mexican parents, they do hold higher expectations of their children than Mexican parents do. The relative influence of these higher expectations on the study habits of children is explored in later multivariate analyses. Non-Familial Sources of Social Capital Although most sources of familial social capital do not appear to be able to account for the variation between Vietnamese and Mexican American students' study habits, sources of non-familial social capital may. We begin by exploring the various school contexts of Vietnamese and Mexican American students. Table 3 shows that overall Vietnamese students attend schools with more favorable environments than do Mexican American students. For example, Vietnamese students attend schools in which, on average, 17.6% of the school body receives reduced lunch. In schools attended by Mexican American students in this sample, almost 41% of the student body receives reduced lunch, indicating that Mexican Americans attend schools with poorer students. Mexican Americans attend schools with higher drop-out rates, with 13% of tenth-graders having dropped out of school on average, compared to 8% in schools attended by Vietnamese students. Vietnamese students are more likely to be in schools with college-bound students, with 45% of the 1989 senior class heading to a
54
KIMBERLY A. GOYETTE AND GILBERTO Q. CONCHAS Table 3. Non-familial Sources of Social Capital among Vietnamese and Mexican Students, NELS 1988-1990 Panel.
Vietnamese Mexicans
Between Schools (School Context) % Receiving Reduced Lunch* Std. Dev. Missing % of 10th Graders who Drop Out* Std. Dev. Missing % of Students Going to a Four-year College* Std. Dev. Missing % of Parents who Volunteer* Std. Dev. Missing Teachers Push Students to Achieve Not accurate or somewhat accurate Accurate or very accurate Missing School Type Public Catholic Other private Missing School Urbanicity Urban Suburban Rural Missing Wi7/i//i Schools Number of Friends that Have Dropped Out None Some Most or all Missing Among Friends, How Important Is Studying? Not important Somewhat important Very important Missing
17.6 20.3 15.7 8.2 8.2 16.0 45.2 26.2 32.2 13.7 17.5 31.0 13.6 56.2 30.3 90.0 8.9 1.1 0.0 56.4 38.4 5.2 0.0
40.5 28.7 15.9 13.2 14.5 16.9 32.8 20.8 46.7 11.4 17.0 48.8 12.8 42.7 44.5 95.1 3.3 1.1 0.6 42.2 33.0 24.3 0.5
73.4 22.0 2.5 2.1 2.8 42.4 52.7 2.1
53.3 34.4 2.6 9.7 7.9 49.6 32.1 10.4
Family and Non-Family Roots of Social Capital Table 3. Continued. Vietnamese Teachers' Expectations No college College Teacher doesn't care Missing N 16.1 56.4 15.6 11.9 163 Mexicans 8.2 51.8 22.3 17.7 1,237
55
These variables are means, not percentages as in the rest of the table. Missing values are excluded for the calculation of means. Note: Results are weighted according to the sample weights provided by NELS.
four-year college, compared to 33% in the schools attended by Mexicans. Parents are slightly more likely to be involved in the schools Vietnamese students attend, with a 14% rate of volunteering compared to 11% in the schools Mexicans attend. When school administrators were asked whether teachers motivate their students to achieve in this school, the schools attended by Vietnamese had a slightly higher rate of agreement. About 56% of the administrators in schools attended by Vietnamese students described the statement as accurate or very accurate compared to 43% of the administrators of schools attended by Mexicans. Further, Mexican Americans are more likely than are Vietnamese to be in public schools at 96%, compared to 87% of Vietnamese. Vietnamese are much more likely than are Mexican Americans to be in Catholic schools, at 11%, compared to only 4%. Apart from the structural differences between public and Catholic schools, Catholic schools may also provide students with greater social capital than do public and other private schools. According to Coleman (1987), in Catholic schools, students are accountable to a value community, and thus receive more supervision and support than do students in other types of schools. We also distinguish between urban, suburban, and rural schools, though it is unclear how urbanicity affects students' study habits. The table shows that Mexican American students are more likely to be in rural schools than are Vietnamese.4 Next, we turn to sources of social capital that may be found within schools. One such source is the peer group. Our measures of peer relationships do not include the frequency of interactions with peers nor relationship quality, as the NELS does not ask questions that would allow us to address these two facets of peer social capital. Instead, we choose two variables that tap other dimensions of peer social capital. We employ the first, the number of friends
56
who have dropped out of school, because this could affect the information, academic support, and advantageous study norms available from peers. Our second variable, whether friends consider studying important, is intended to ascertain whether friends will support a student's educational goals and whether the friends' behavior teaches or reinforces good study habits. Table 3 shows that there are large ethnic differences in peer group norms and behaviors. Vietnamese students have friends who drop out of school less and consider studying more important than do the friends of Mexican Americans. Almost 73% of Vietnamese reported having no friend that dropped out of high school, while only 53% of Mexican American students reported having no friends drop out of high school. Another 34% of Mexican Americans report having some friends that have dropped out of high school compared to only 22% of Vietnamese. On the other hand, nearly 53% of the friends of Vietnamese report that studying is very important to them compared to 32% of the friends of Mexican Americans. These results are consistent with the findings of Steinberg et al. (1992), who report that Asian Americans' peers are more likely to encourage their academic success than are the peers of other groups.5 Another source of non-familial social capital that we explore is the relationship between the student and teachers. The NELS provides no information on the frequency of interactions between the student and his or her teachers, nor can we gauge the helpfulness of the teachers' resources for the educational careers of students. However, we do have the student's report of his or her favorite teacher's expectation for future education. Table 3 shows some differences between teachers' expectations for Vietnamese and Mexican American students, although at first glance, teachers' expectations of these students appear to be similar. Over 56% of Vietnamese students report that their teachers expect them to graduate from college, compared to 52% of Mexican American students. However, among students who report that they do not know their teachers' expectations or that their teachers do not care, there are bigger ethnic differences. Over 22% of Mexican American students report that they do not know teachers' expectations or teachers do not care, compared to less than 16% of Vietnamese students. This NELS variable that measures the student-teacher relationship is imperfect for several reasons. First, the variable is student reported. Second, it is difficult to separate the influence of students' unobserved characteristics from their teachers' expectations. Finally, these measures do not capture the nuances of the relationships between students and teachers. In multivariate models we rely on this imperfect measure, although we later present qualitative evidence to further clarify the quantitative results.
57
THE INFLUENCE OF FAMILIAL AND NON-FAMILIAL SOCIAL CAPITAL ON STUDY HABITS

From Table 1, we see that, without controlling for other variables, Vietnamese do approximately 2 hours more homework per week than do Mexican Americans. To test whether differences in the sources of social capital are related to variation in study habits across these two groups, we rely on multivariate models. Table 4 presents multivariate Ordinary Least Squares (OLS) regression models with time spent on homework outside of school per week as the dependent variable.6 The models are straightforward to interpret. Coefficients represent the increase in time spent on homework outside of school (in hours) that corresponds to a one-unit increase in the independent variable. Results from these models account for the stratified, cluster sampling design of the NELS through the use of weights in the statistical program STATA. Model 1 reproduces the bivariate results from Table 1. Here we see that the differences are significant at the 0.01 level of confidence. Without introducing any controls, Vietnamese students spend approximately two more hours per week on homework. Model 2 introduces the controls for immigration generation and family socioeconomic status. From this model, we see that second-generation students do more homework than first-generation students, but third-generation students do not. Neither of these differences is significant. SES, though, does seem to matter for time spent on homework. The higher the family SES, the more hours a student spends on homework. With the introduction of these controls, the differences between the groups are reduced. Vietnamese now average about an hour and a half more on homework per week than do Mexican Americans. However, the time Vietnamese and Mexican Americans spend on homework remains different at the 0.01 level of significance. In Model 3, we add familial social capital to the model. These measures further reduce the differences between the two groups. Vietnamese are predicted to do about an hour and a quarter more homework than Mexican Americans, and this difference is significant at the 0.10 level of confidence. Parents' interaction with schools has a positive and significant effect on time spent on homework, as does parents' interaction with children. In contrast, family structure and social closure do not seem to matter much for time spent on homework. Parents' expectations matter, though, particularly for those children whose parents expect them to graduate from college. These children are predicted to do almost two hours more homework per week than children whose parents do not expect them to complete high school.
58
KIMBERLY A. GOYETTE AND GILBERTO Q. CONCHAS Table 4. Coefficients from Ordinary Least Squares (OLS) Models Predicting the Hours Students Spend on Homework Outside of School; NELS 1988-1990 Panel.
Model 1 Model 2 4.716** 1.617** Model 3 2.868** 1.278* Model 4 4.133** 1.005+ Model 5 2.242+ 0.883
Constant Race (Mexican = excluded) Vietnamese Background Characteristics Immigration Generation (Firstgeneration = excluded) Second-generation Third-generation SES Familial Social Capital Family Composition (Intact = excluded) Not intact Parent-School Interaction Index Parent-Child Interaction Index Closure Parents' Expectations (Less than high school = excluded) High school graduate Some college College graduate
3.902** 1.986**
0.041 -0.261 1.029**
-0.200 -0.212 0.705**
-0.071 -0.370 0.884**
-0.250 -0.326 0.683'
0.440 0.349* 0.132+ 0.197
-0.471 0.285* 0.128* -0.185
0.791 1.655+ 1.974*
0.804 1.669* 1.779'
Non-familial Social Capital % Receiving Reduced Lunch % of 10th Graders who Drop Out % of Students Going to Four-year Colleges % of Parents who Volunteer Teachers Push Students to Achieve (Not accurate or somewhat accurate = excluded) Accurate or very accurate School Type (Public = excluded) Catholic Private School Urbanicity (Urban = excluded) Suburban Rural
-0.005 0.014 0.005 0.015
-0.004 0.016 0.005 0.016
0.002
0.004
0.138 1.042
-0.667 -0.564
0.115 0.354
-0.101 -0.290
Family and Non-Family
Roots of Social Capital Table Model 1 4. Continued. Model 3 Model 4 Model 5
59
Model 2
Friends Who Have Dropped Out (None = excluded) Some Most or all Importance of Studying to Friends (Not at all excluded) Somewhat Very Teachers' Expectations (College excluded) No college Doesn't care R2 0.017
+
-0.928** -1.367**
-0.764* -1.144*
1.079* 1.843**
1.041* 1.737**
0.046
0.090
-0.632 -1.110** 0.123
-0.499 -1.056** 0.152
p < 0.10 *p< 0.05 **p < 0.01. Note: The above models also include dummy variables representing missing values for immigration generation, family composition, parent-school interaction, parent-child interaction, closure, parents' expectations, percentage in school receiving free lunch, percentage of tenth graders who dropped out, percentage of students going to a four-year college, percentage of parents who volunteer, teachers' stress on achievement in the school, school type, school urbanicity, friends who have dropped out, importance of studying to friends, and teachers' expectations. The sample size is 1,400. Note: Results are weighted according to the sampling weights provided by NELS, and the inflation of the standard errors has been corrected using a procedure in the statistical package STATA that accounts for non-random sampling designs.
In Model 4, we add sources of non-familial social capital to the model with only controls. The sources of non-familial social capital include both school context variables and variables that pertain to students' relationships within schools. In this model, we see that school context variables have little significant effect on students' study habits. The percentage of the school that receives reduced lunch has a small negative effect, while the percentage of seniors attending a four-year college, teachers' motivation of students, and the rate of parental volunteerism are slightly positive. Surprisingly, the effect of the percentage of tenth-grade drop-outs is weak but positive, while the influence of attending a Catholic or private school compared to a public school is negative. However, none of these variables achieve significance, even at the 0.10 level of confidence.7 In contrast to the results concerning the school environment, relationships within schools appear to be strongly related to students' study behavior. In this model, we see that students' friends' norms and behaviors are linked to the
60
time they spend on homework. Children who have many friends who drop out of school do about 173 hours less homework per week than children who have no friends drop out of school. If studying is very important to a student's friends then that student is likely to do 2 more hours of homework per week than a student whose friends do not consider studying important at all. Both of these effects are significant at the 0.01 level of confidence. Further, teachers' expectations have a significant effect on students' study behavior. Students who do not know teachers' expectations or believe that teachers do not care do over an hour less homework than do those whose teachers expect them to attend college. With the addition of these non-familial social capital variables, differences between Vietnamese and Mexican Americans are reduced even more so than in the model with only familial social capital variables. Rather than doing l'A hours of homework more than Mexicans Americans, in this model Vietnamese are predicted to do just about an hour more homework per week. This difference is still significant at the 0.10 level of confidence, however. In thefinalmodel, Model 5, we include both familial and non-familial sources together in the same model. Here differences between Vietnamese and Mexican American students are reduced to insignificance. After controlling for both familial and non-familial social capital, Vietnamese students are predicted to do less than an hour more of homework per week than Mexican students. Connections between parents and schools, and parents and teachers; parents' expectations; friends' norms and behaviors; and teachers' expectations all remain significant in this model. Thisfinalmodel including both the familial and non-familial sources of social capital is also run separately for Mexican and Vietnamese students. These results are presented in Appendix Table A. Comparison of the two sets of results shows that there are differences between the groups. For example, immigration generation is a more important predictor of Vietnamese study habits than it is of Mexicans'. Second- and third-generation Vietnamese study significantly less than do their first-generation peers, while the opposite is true among Mexicans. Interactions between parents and schools seem to matter more for Mexican students' behavior than for Vietnamese students', but communication between students and parents appears to be more important for Vietnamese students. Parents' expectations achieve significance for Mexican students, but not Vietnamese. However, this difference in significance could reflect the difference in size between the Vietnamese and Mexican samples. In small samples, standard errors are often larger than in big ones, thus significance is harder to achieve. Non-familial capital also influences Vietnamese and Mexican students differently, though these differences do not seem great. School type appears to have opposite effects for Vietnamese and Mexican students, with Vietnamese students
61
benefiting more from private and Catholic schools than do Mexicans, but these variables do not achieve significance. Vietnamese students in rural schools appear to study harder than do those in urban schools and this variable is significant, in contrast to the effect of rural schools on Mexican Americans. Having friends who consider studying important has a positive effect on Vietnamese students, though this variable is not as significant as it is among Mexican Americans. Reporting that they do not know or that teachers do not care whether or not students go to college has a negative effect among both Mexican and Vietnamese students, though it is significant only among Mexican students. When we looked for these interaction effects in the combined sample, we found none that significantly improved the fit of the model. Again, a small sample size of Vietnamese students likely limited the explanatory power of these effects. Although the separate models show some differences between the two groups, the larger conclusions remain the same. While parents appear to have some influence on the study behavior of Mexican and Vietnamese students, so, too, do peers and teachers. Unlike studies that focus on differences in access to non-familial social capital between schools (Hoelter, 1982; StantonSalazar & Dornbusch, 1995), our research highlights the importance of relationships within schools. To further explore and emphasize these relationships, we present qualitative findings that illuminate the influence of peers and teachers on students' motivations and behavior within one urban, California high school in the next section of this research. We investigate the following questions: Do and how do the resources derived from relationships within schools influence students' educational norms and behaviors? Does this type of non-familial social capital shape student outcomes? Does/how does this influence their study behavior? Qualitative Results: Non-familial Social Capital within an Urban High School In the qualitative study, both Vietnamese and Mexican American students identify several important resources they gain from their relationships with teachers and friends. These resources include support and information for achieving their educational goals. The support and information they receive from peers and teachers influence their motivation to succeed in school, which in turn relates to their study behavior. Often, they report that the acquisition of these resources differs by ethnicity. Students note that peer support is vital for their motivation, behavior, and achievement. Data from this study reveal that students' friends give them
62
information, influence them to study hard, and encourage their school success. Although this is true for both the Vietnamese and Mexican American high school students, the Vietnamese students, more so than the Mexican American students, form relationships in school to support each other's academic achievement. For instance, a Vietnamese student remarks that "[Vietnamese] believe in all this brotherhood thing where if one [Vietnamese] needs help . . . the whole group of [Vietnamese] come out to help" (Fred, 3/4/97, p. 14). Even though mostly low-achieving, Mexican American students also suggest that peers have a powerful influence on their school success. One student in the study, Jorge, states that if he had positive friends to push him, then perhaps he would do well. Though both groups recognize that peers shape their study motivation and behavior, they also note that friendships within the school are segregated by ethnicity. Says Ana, a Mexican American junior at Baldwin, "Students segregate themselves based on how they are treated. Like you go out for lunch . . . and you see . . . a group of Asian people right there in the classes with some Whites, and then you see a group of Blacks in the front, and then you see a group of Latinos by the gym." This segregation limits the resources that Mexican American students are able to obtain from peers. Another source of social capital within schools is relationships with teachers. The results of this study show that these relationships also appear to vary by student's ethnicity. Overwhelmingly, teachers and administrators in Baldwin High School portray Vietnamese students positively, while Mexican American students are seen negatively. Says one white, female, Spanish teacher, Vietnamese students "are . . . the most motivated in this school and anyone here knows that . . . African Americans and the . . . Hispanics and Latinos are kind of split between who are least motivated" (Spanish Teacher, 4/23/97, p. 3). The same Chinese counselor cited earlier states, "We place students in academies based on what they are capable of doing . . . and it also depends on what the student wants . . . Many Asian students want to be engineers, as opposed to Black students who do not think about engineering, and as opposed to how many Latino students don't either . . . " (4/24/98, p. 26). Teachers and counselors in this study seem to believe that a student's ethnicity relates to his or her motivation and ability. Students are aware of these perceptions. Vietnamese students comment that teachers expect them to excel simply based on their ethnicity. "Teachers look at appearance and just because we are a certain thing, they expect us to do good all the time," states Genie, a Vietnamese student. Similarly, the following dialogue illustrates Vietnamese students' awareness of teachers' perceptions of them and how that influences students:
Roots of Social
Capital
63
GC: Do teachers have the same expectations of all students? Lisa: Not always . . . People around you, they just don't see us different, like teachers just assume we are all the s a m e . . . that means us, Asian people, have to set a higher standard, since you want to live up to what t h e y . . . are calling you or what they assume that you are. GC: So you think that Vietnamese are stereotyped in different ways? Kim: Okay, it's like they treat us in some different ways because we are Asian. We are supposed to be smart, so they, they want you to be smart in some ways and they think that you are smart by looking at y o u . . . all they want to see is the stereotype (3/4/97, pp. 5-6).
Mexican American students are also aware of teachers' perceptions of them, though these perceptions are often negative. A Mexican American junior, Diego, is sympathetic to his teachers when he states:
I don't think any of this is done directly. All this racial segregation in the school. I don't think there is any one person or group that are out to do this at school, but I think just the ways things have shaped up, things that happen... are the way they arc because the stereotypes that people hold and they get turned into who gets the best and most challenging things here. Teachers have also been influenced by this . . . I think that over the years, they have seen it over and over again and after a while they help in making stereotypes come true.
Less forgiving, Marisa, another Mexican American student, says, teachers "think we're all the same, they think we don't exist." Students notice that their interactions with teachers differ by ethnicity. Further, they are aware that important resources are differentially gained from these relationships. For example, Tran suggests that some teachers' negative attitudes toward African Americans and Mexican Americans result in them receiving less academic support than many Asian and white students have.
Some teachers treat students equally, but a lot of them treat us differently . . . Asians and whites get more attention and better classes... Most teachers treat Blacks and Mexicans kinda bad because they think they are lazy, that they don't do nothing (Tran, 3/10/97, p. 8).
Relationships with peers and teachers influence students' motivation to work hard in school. For example, Vietnamese students see the support they receive from teachers as integral to their academic success. Tran, for instance, believes that "the [school] program pushes us and motivates us and tells us by giving support and lets us know when we are doing well" (Tran, 3/10/97, p. 4). Similarly, Sandy expresses the importance of receiving information for students' motivation:
. . . the [program] makes you focus on what you really want to do, you know, like especially for students who want to do something in the medical field. They motivate you, they put you in the right classes . . . and in the right level. They put you in everything, they give you a whole bunch of packets, and they take care of you (Sandy, 3/10/97, p. 11).
64
Support such as this motivates students to work hard to achieve their academic goals. While Mexican American students also see such support as necessary, the majority of Mexican American students at Baldwin High School are not as academically motivated and successful as the Vietnamese students. Many of the Mexican American students at Baldwin High School found school boring and disengaging. They cut classes and did little or no homework because, as Blanca explains, "there ain't nothing else to do." Most attribute their lack of academic motivation and achievement to a lack of guidance from adults and other students (Stanton-Salazar, 1997; Stanton-Salazar & Dornbusch, 1995). Many of the Mexican American students express that teachers and counselors are only concerned about Asian students. Jorge, for instance, relates that counselors are not attentive to his needs: "I don't think they care because I have been filling out slips to go see my counselor. I sent like four from September and they still have not called me. Every time I go there, he's at lunch or is with other students and during class he has no time for me." Ricardo, a Mexican immigrant sophomore, comments that differing perceptions of students by ethnicity influence his ability to improve his schoolwork:
(Translated from Spanish.) For instance, my teacher, and I'm in Graphics, he goes up to an Asian student, looks at his work and says, 'you could do better.' With me, however, he simply s a y s . . . 'it's all right.' But he never says I can do better, right? He is like telling me, for me (as a Mexican), it is all right. Like if I cannot do better than that, that is the best I can do. And I do not like that.
In short, the Mexican students believe that teachers do not notice them or care about them. This is linked to their lack of motivation to study outside of school. These Mexican American students are given little guidance and support from what Stanton-Salazar (1997) calls key "institutional agents" such as peers, and adult teachers and counselors. During interviews and informal conversations, some express an interest in becoming computer technicians, nurses, doctors, astronauts, and small business owners, but, unlike the Vietnamese students, they do not know how to achieve these goals. Miguel, for example, wants to eventually marry, have children, and run a small business. "I have lots of goals," Miguel explains. "I want to be a lot of people and do lots of things . . . I want to have my own shop, like a high performance shop dealing with engines. I know a lot about that." However, he articulates his awareness that information and support are necessary to achieve his goals. "I have no support, man, no way of doing it." Students like Miguel understand the importance of positive peer relationships and caring teachers
65
but have little experience of either. Miguel explains that "good teachers and good school programs could help." Although there are plenty of good programs and teachers at Baldwin, these Mexican American students do not feel they have access to them. Their sense of alienation and invisibility translates into a lack of motivation, poor study habits, a failure to plan for college, and pessimism about life-long career goals. Together, the quantitative and qualitative data illustrate that non-familial social capital found within schools is linked to the differences in the study habits of Vietnamese and Mexican Americans. The quantitative data shows that indeed peer norms and behaviors are important influences on the time students spend on homework. Further, students who did not know their teachers' expectations or who believed their teachers did not care about their educational futures did substantially less homework than others. Using the qualitative data, we find that Vietnamese students more than Mexican students benefit from relationships with motivated peers and caring teachers. These interactions result in higher motivation and more studious behavior among Vietnamese Americans than among Mexican Americans.
CONCLUSIONS
The main goal of this study was to investigate the extent to which different sources of social capital explain difference in Vietnamese and Mexican American students' study habits. Although academic literature and popular media have argued that families are largely responsible for differences between these two recent immigrant groups, our research shows that relationships outside the family - i.e. non-familial social capital - also play a large role in explaining variation between the two groups. Our results suggest many directions for future research. Our findings, both quantitative and qualitative, point to the fact that the acquisition of social capital may vary by ethnicity. In this research, ethnicity is particularly important in influencing the resources that students are able to gain from non-family relationships. More work should be done to see how the sources of social capital, and the resources able to be gained from them, differ according to race and ethnicity. Another research direction to which our results point is the role of the family in the achievement of children of different races and ethnic backgrounds. Can we assume that familial social capital has the same effects on the educational outcomes of children of all races and ethnic backgrounds? Do minority and immigrant children rely on familial social capital less than white and native-born children? Are there measures of familial social capital that would better capture the ways in which minority and recent immigrant families shape children's
66
achievement? Our results suggest that family social capital, as it is measured in our research, does not account for a large share of the difference in the study habits of these two recent immigrant minority groups. Perhaps family influence could be measured differently to better capture the effects of families on children. A final issue that we do not address empirically is the activation of social capital. Lareau and Horvat (1999) contend that in order to gain resources from a source of social capital, students have to be able to use it in specific environments. Students' capacity to use their social capital may be circumscribed by their own race or ethnicity and the race or ethnicity of those who hold resources. For example, Lareau and Horvat (1999) show that African American parents attempting to intervene on children's behalf in schools may be prevented from doing so by teachers' perceptions that such parents are "pushy." In our research, it is possible that the social capital Mexican American students bring to schools is not activated as much as is the social capital brought by Vietnamese students. Perhaps the interventions of extended family members on behalf of Mexican American children are not valued because these advocates are not parents, and thus not perceived as the proper guardians with whom information should be shared. In contrast, the mannerisms that Vietnamese parents bring to their interactions with teachers may win teachers' approval. Vietnamese parents are said to treat teachers with great respect based on their Confucian heritage (Nash, 1987). Teachers may respond well to this treatment, and in turn, share information and give support to both parents and students. The perceptions of each of these ethnic groups by teachers and other school officials - Vietnamese as "model minorities" and Mexicans as lazy and unmotivated - may affect the extent to which they are able to use the social capital they receive from their families. Much more work needs to be done investigating how students are able to activate their social capital in various settings. Our challenge is to push those exploring social capital to think about how race and ethnicity influence how it is acquired and how it is expressed. Research done on the effects of social capital using predominantly white samples has led to generalizations among policy-makers and the public about steps that should be taken to enhance children's success in schools. Since most of the research done on social capital and educational achievement has focused on the role of the family in transmitting it, families are either praised or blamed for children's success in schools. Our results suggest that, at least among two recent immigrant minorities, intangible resources gained from families may not be as important in influencing their study habits as those gained from relationships within schools. Findings from our research suggest that educators should concentrate their efforts on equalizing access to such relationships by designing
67
policies to reduce ethnic segregation among peers and to encourage teachers to mentor students of all ethnic groups. Before we can truly understand how family and non-family practices influence children's school success, we need to investigate their influence on all social groups - immigrants, minorities, and low-SES students. It is not self-evident that what works for the majority, usually middle-class whites, will also work for others. Social capital, in this case, ultimately refers to the useful, though intangible resources students are able to utilize to enable their educational success. To provide a more complete understanding of inequality in schools, we need to further explore how these intangible resources are distributed and actualized, and how this affects the educational achievement of all children.
NOTES
1. For important exceptions see Stanton-Salazar and Dornbusch (1995), and Zhou and Bankston (1998). 2. To some extent, different school contexts could also reflect the social capital of parents. Parents with good information may choose to send their children to particular schools or move to areas where there are high quality schools. 3. Another possible interpretation is that Vietnamese students have fewer friends than do Mexican students. We contend, though, that the variable still measures community involvement, as students who have few friends are less likely to be involved in closelyknit, supportive networks of relationships. 4. We also explored the percentage of the student body that was white and school size as school context variables, but neither showed an effect on study habits in multivariate models. 5. There are several problems with using the reported values and behaviors of friends as a proxy for the social capital available from these sources. The biggest difficulty with this measure is that students select their friends. It is difficult to know the extent to which observed coefficients in multivariate models are due to the influence of friends and the extent to which they are due to the student's own unobserved characteristics. Studious students tend to choose those with like norms and behaviors as friends. Fortunately, we have qualitative information, which we present later, that demonstrates the importance of peer relationships on students' motivation and study habits. 6. We also did analyses with the dependent variable as logged number of hours spent on homework outside of school. The results from these models are available from the authors upon request. Since the results were similar, here we simply report coefficients from the model using hours spent on homework as the dependent variable for ease of interpretation. 7. It has been suggested that the influence of school environment on students' motivation and behavior may operate in opposing directions, and thus cancel out overall effects. Low achieving students may transmit poor study norms to others in their schools, as high achieving students may challenge their peers. However, some speculate that students in schools with low achievers may study hard because they
68
believe themselves to be more talented and motivated than their peers. Others note that students in schools with high achievers, comparing themselves to their peers, may become discouraged about their abilities, and therefore study less.
REFERENCES
Alvirez, D., & Bean, F. D. (1976). The Mexican American Family. In: C. H. Mindel & R. W. Habenstein (Eds), Ethnic Families in America: Patterns and Variations (pp. 271-292). New York: Elsevier Press. Astone, N. M., & McLanahan, S. S. (1991). Family Structure, Parental Practice, and High School Completion. American Sociological Review, 56, 309-320. Bourdieu, P. (1977). Reproduction in Education, Society, and Culture. London: Sage Publications. Burt, R. S. (1992). Structural Holes: The Social Structure of Competition. Cambridge MA: Harvard University Press. Carter, T. P., & Segura, R. D. (1979). Mexican Americans in Schools: A Decade of Change. New York: College Entrance Examination Board. Caplan, N., Choy, M. H., & Whitmore, J. K. (1991). Children of the Boat People: A Study of Educational Success. Ann Arbor MI: The University of Michigan Press. Carbonaro, W. J. (1998). A Little Help from My Friend's Parents: Intergenerational Closure and Educational Outcomes. Sociology of Education, 71, 295-313. Caterall, J. S. (1998). Risk and Resilience in Student Transition to High Schools. American Journal of Education, 106, 302-333. Chen, C , & Stevenson, H. (1995). Motivation and Mathematics Achievement: A Comparative Study of Asian-American, Caucasian, and East Asian High School Students. Child Development, 66, 1215-1224. Coleman, J. S. (1987). The Relations between School and Social Structure. In: M. T. Hallinan (Ed.), The Social Organization of School: New Conceptualizations of the Learning Process (pp. 177-204). New York: Plenum Press. Coleman, J. S. (1988a). 'Social Capital' and Schools. The Education Digest, 53, 6-9. Coleman, J. S. (1988b). Social Capital in the Creation of Human Capital. American Journal of Sociology, 94, S95-120. Conchas, G. Q. (1999). Structuring Educational Opportunity: Variations in Urban School Success among Racial Minority Youth. Unpublished doctoral dissertation, University of Michigan, Ann Arbor. Conchas, G. Q. (In press). Structuring Failure and Success: Understanding the Variability in Latino School Engagement. Harvard Educational Review. De Vos, G. A. (1982). Adaptive Strategies in American Minorities. In: E. E. Jones & S. Korchin (Eds), Minority Mental Health. New York: Praeger. Farkas, G., Grobe, R. P., Sheehan, D., & Shuan, Y. (1990). Cultural Resources and School Success: Gender, Ethnicity, and Poverty Groups within an Urban School District. American Sociological Review, 55, 127-142. Fine, M. (1991). Framing Dropouts: Notes on the Politics of an Urban Public High School. New York: State University of New York. Furstenberg, F. F., & Hughes, M. E. (1995). Social Capital and Successful Development Among At-Risk Youth. Journal of Marriage and the Family, 57, 580-592. Gandara, P. (1995). Over the Ivy Walls: The Educational Mobility of Low-Income Chicanos. New York: State University of New York.
Roots of Social
Capital
69
Goyette, K., & Xie, Y. (1999). Educational Expectations of Asian American Youths: Determinants and Ethnic Differences. Sociology of Education, 72, 22-36. Granovetter, M. S. (1974). Getting a Job: A Study of Contacts and Careers. Cambridge MA: Harvard University Press. Hagan, J., MacMillan, R., & Wheaton, B. (1996). New Kid in Town: Social Capital and the Life Course Effects of Family Migration on Children. American Sociological Review, 61,368-385. Heller, C. (1966). Mexican-American Youth: Forgotten Youth at Crossroads. New York: Random House. Hoelter, J. (1982). Segregation and Rationality in Black Status Aspirations Processes. Sociology of Education, 55, 31-39. Horowitz, R. (1981). Passion, Submission, and Motherhood: The Negotiation of Identity by Unmarried Innercity Chicanas. Social Science Quarterly, 22, 241-252. Jencks, C. (1972). Inequality: A Reassessment of the Effect of Family and Schooling in America. New York: Basic Books. Kemple, J., & Snipes, J. (2000). Career Academies: Impacts on Students' Engagement and Performance in High School. New York: MDRC. Kuvelesky, W. P., & Patella, V. M. (1971). Degree of Ethnicity and Aspirations for Upward Social Mobility among Mexican American Youth. Journal of Vocational Behavior, I, 231-244. Lareau, A., & Horvat, E. M. (1999). Moments of Social Inclusion and Exclusion: Race, Class, and Cultural Capital in Family-School Relationships. Sociology of Education', 72, 37-53. Lee, S. J. (1996). Unraveling the "Model Minority" Stereotype: Listening to Asian American Youth. New York: Teachers College Press. Loury, G. C. (1977). A Dynamic Theory of Racial Income Differences. In: P. A. Wallace & A. LaMund (Eds), Women, Minorities, and Employment Discrimination (pp. 153-186). Lexington MA: Lexington Books. Matute-Bianchi, M. E. (1986). Ethnic Identities and Patterns of School Success and Failure among Mexican-Descent and Japanese-American Students in a California High School: An Ethnographic Analysis. American Journal of Education, 95, 233-255. McLanahan, S., & Sandefur, G. (1994). Growing Up with a Single Parent: What Hurts, What Helps. Cambridge MA: Harvard University Press. Mehan, H., Hubbard, L., & Villanueva, I. (1994). Forming Academic Identities: Accommodation without Assimilation among Involuntary Minorities. Anthropology and Education Quarterly, 25,91-117. Mehan, H., Villanueva, I., Hubbard, L., & Lintz, A. (1996). Constructing School Success: The Consequences of Untracking Low-Achieving Students. Cambridge: Cambridge University Press. Morgan, S. L., & S0rensen, A. B. (1999). Parental Networks, Social Closure, and Mathematics Learning: A Test of Coleman's Social Capital Explanation of School Effects. American Sociological Review, 64, 661-681. Nash, J. W. (1987). Vietnamese Values: Confucian, Catholic, American. Unpublished doctoral dissertation, Tulane University, New Orleans. Ogbu, J. U. (1991). Minority Coping Responses and School Experience. Journal of Psychohistory, 18, 433-456. Parsons, T. (1949). The Social Structure of the Family. In: R. N. Ashen (Ed.), The Family: Its Functions and Destiny (pp. 173-201). New York: Harper. Portes, A. (1998). Social Capital: Its Origins and Applications in Modern Sociology. Annual Review of Sociology, 22, 1-24.
70
Rumberger, R. W. (1987). High School Dropouts: A Review of Issues and Evidence. Review of Educational Research, 57, 101-112. Sanchez-Jankowski, M. (1991). Islands in the Street: Gangs and American Urban Society. Berkeley: University of California Press. Schneider, B., & Lee, Y. (1990). A Model for Academic Success: The School and Home Environment of East Asian Students. Anthropology and Education Quarterly, 21, 358-377. Stanton-Salazar, R. D. (1997). A Social Capital Framework for Understanding the Socialization of Racial Minority Children and Youths. Harvard Educational Review, 67, 1-39. Stanton-Salazar, R. D., & Dornbusch, S. M. (1995). Social Capital and the Reproduction of Inequality: Information Networks among Mexican-Origin High School Students. Sociology of Education, 68, 116-135. Steinberg, L., Dornbusch, S. M., & Brown, B. B. (1992). Ethnic Differences in Adolescent Achievement: An Ecological Perspective. American Psychologist, 47, 723-729. Sudrez-Orozco, M. M., & Sudrez-Orozco, C.E. (1995). The Cultural Patterning of Achievement Motivation: A Comparison of Mexican Immigrant, Mexican American, and Non-Latino White American Students. In: R. G. Rumbaut & W. A. Cornelius (Eds), California's Immigrant Children: Theory, Research, and Implications for Educational Policy (pp. 161-190). San Diego: The Center for U.S.-Mexican Studies, University of California. Teachman, J. D., Paasch, K., & Carver, K. (1996). Social Capital and Dropping Out of School Early. Journal of Marriage and the Family, 58, 773-783. Teachman, J. D., Paasch, K., & Carver, K. (1997). Social Capital and the Generation of Human Capital. Social Forces, 75, 1343-1359. Thornton, C. H., & Eckland, B. K. (1980). High School Contextual Effects for Black and White Students: A Research Note. Sociology of Education, 53, 247-252. Valenzuela, A. (1999). Subtractive Schooling: U.S.-Mexican Youth and the Politics of Caring. New York: State University of New York. Valenzuela, A., & Dornbusch, S. M. (1994). Familism and Social Capital in the Academic Achievement of Mexican Origin and Anglo Adolescents. Social Science Quarterly, 75,18-36. Vigil, D. J. (1988). Barrio Gangs: Street Life and Identity in Southern California. Austin: University of Texas Press. Weiler, J. W. (2000). Codes and Contradictions: Race, Gender, Identity, and Schooling. New York: State University of New York. Wong, M. G. (1980). Model Students? Teachers' Perceptions and Expectations of their Asian and White Students. Sociology of Education, 53, 236-246. Zhou, M., & Bankston, C. L., III. (1998). Growing Up American: How Vietnamese Children Adapt to Life in the United States. New York: Russell Sage Foundation.
Roots of Social
Capital
71
Appendix Table A. Coefficients from Ordinary Least Squares (OLS) Models Predicting the Hours Students Spend on Homework Outside of School Separately for Mexican and Vietnamese Students; NELS 1988-1990 Panel.
Mexican Constant Background Characteristics Immigration Generation (First-generation = excluded) Second-generation Third-generation SES Familial Social Capital Family Composition (Intact = excluded) Not intact Parent-School Interaction Index Parent-Child Interaction Index Closure Parents' Expectations (Less than high school = excluded) High school graduate Some college College graduate Non-familial Social Capital % Receiving Reduced Lunch % of 10th Graders who Drop Out % of Students Going to Four-year Colleges % of Parents who Volunteer Teachers Push Student to Achieve (Do not agree or somewhat agree excluded) Agree or strongly agree School Type (Public = excluded) Catholic Private School Urbanicity (Urban = excluded) Suburban Rural Friends Who Have Dropped Out (None = excluded) Some Most or all Importance of Studying to Friends (Not at all = excluded) Somewhat Very Teachers' Expectations (College = excluded) No college Doesn't care R2 2.177 Vietnamese 2.965
0.155 0.179 0.648*
-2.783* -2.403 0.502
0.490 0.369* 0.107+ 0.232 0.768 1.673* 1.754+ -0.005 0.014 0.005 0.013
0.240 -0.683 1.214** 0.501 1.551 0.025 1.721 -0.025 0.048 -0.026 0.002
0.059 -0.767 -0.934 -0.414 -0.497 -0.672* -1.026* 1.116* 1.737** -0.377 -1.056** 0.150
0.720 2.792 3.708 1.604 3.113* -1.743* -1.496 0.146 1.444 1.988+ -1.651 0.393
72
+ p < 0.10 *p< 0.05 **p < 0.01. Note: The above models also include dummy variables representing missing values for immigration generation, family composition, parent-school interaction, parent-child interaction, closure, parents' expectations, percentage in school receiving free lunch, percentage of tenth graders who dropped out, percentage of students going to a four-year college, percentage of parents who volunteer, teachers' stress on achievement in the school, school type, school urbanicity, friends who have dropped out, importance of studying to friends, and teachers' expectations. The sample size is 1,237 Mexicans and 163 Vietnamese. Note: Results are weighted according to the sampling weights provided by NELS, and the inflation of the standard errors has been corrected using a procedure in the statistical package STATA that accounts for non-random sampling designs.
Commentary: USES AND MISUSES OF SOCIAL CAPITAL IN STUDYING SCHOOL ATTAINMENT

Commentary on Goyette-Conchas and Bankston-Zhou papers.
Patricia Fernandez Kelly
INTRODUCTION
In the early summer of 2001, I attended a small graduation ceremony at Princeton University, where I teach. The yearly event anteceded, by a few days, regular commencement exercises and was organized to honor the achievements of students of Mexican descent. The moment had singular meaning because there were several among the celebrants whom I had known since their freshman year. Along the way some hadfloundered,but in the end most had prevailed. At Princeton, Chicanos, as they like to call themselves, include many children of recent immigrants. Now they stood in the company of their families and friends, beaming and choked-up by emotion. At one table was a father wearing a cheap but formal suit and leather sandals. His parched complexion bore testimony to years of stoop labor in the fields of Texas. Only recently had he obtained legal residency. His son, one of eight brothers and sisters, sat nearby wiping his eyes surreptitiously. He was bound for Yale Medical School - a monumental leap from his humble origins.
Schooling and Social Capital in Diverse Cultures, Volume 13, pages 73-83. 2002 Published by Elsevier Science Ltd. ISBN: 0-7623-0817-6 73
74
PATRICIA FERNANDEZ KELLY
The sight of youngsters overcoming daunting obstacles thanks to the sacrifice of parents and their own personal resilience is compelling - a vivid reenactment of heroic narratives at the core of our culture. But families do not succeed alone. Counselors and teachers who had aided the young students in their journey to Princeton also sat at various tables. Mothers and fathers had afforded their children the reassurance of strong ties, but it had been the conscientious mentors who had forged the weak ties, necessary for them to find a way into the ivy-league. Displayed before my eyes were the tangible effects of repeated local transactions between individuals with differing capacities to negotiate the slings and arrows of the educational system. It is against that context that I wish to reflect upon the two preceding papers, "Family and Non-family Roots of Social Capital Among Vietnamese and Mexican Children" by Kimberly A. Goyette and Gilberto Q. Conchas, and "Social Capital and Immigrant Children's Achievement" by Carl L. Bankston II and Min Zhou. Both touch upon key issues concerning the effects of kinship and friendship on the behaviors of immigrant students. Before reviewing their content, I provide a short conceptual overview.
HOW TO CAPTURE THE IDEA OF SOCIAL CAPITAL?

Concepts are the building blocks of theory. They are constructed to make visible phenomena that otherwise would remain undetected. To be useful, conceptual definitions must be restricted in focus and limited in scope. Only then can they properly serve their descriptive and explanatory purposes. In other words, to endure, a concept must be intellectually restrained. Discipline, however, is difficult to uphold. Successful concepts are forever running the risk of dissolving into banality through overuse and careless expansion of their purview. So it was with social capital, the idea uniting the two articles under consideration. As originally defined by Coleman (1988, 1990) and Bourdieu (1986), in separate scholarly contributions, social capital was meant to designate situational advantages derived by individuals from their membership in particular networks. This was in analytical contrast to the skilling benefits accruing from formal instruction (human capital) or wealth (material capital). Mark Granovetter (1985) offered the complementary notion of embeddedness to underscore the social underpinnings of economic action. In consonance with ideas first articulated by Max Weber, his objective was to expose the limitations of orthodox
Commentary: Uses and Misuses of Social Capital
75
economics by emphasizing the organic interrelationship between social and economic phenomena. The result of these efforts was to greatly expand the range of the New Economic Sociology. As the field grew, social capital came into vogue and its definitional limits were stretched further and further. Robert Putnam (2000) used social capital to designate every possible asset available to individuals and groups on their way to rectitude and prosperity. In Putman's circular schema, successful societies thrive because they possess successful norms and values. Almost anything and, simultaneously, almost nothing of worth can be explained from that point of view. Diminished conceptual rigor coincided with the popularization of social capital as a panacea for the ills of society (Putnam, 1993). Ghetto poverty was especially susceptible to misuses of the term. With characteristic enthusiasm, policy wonks discussed strategies to infuse social capital into the dark corners of the inner city. Robert Putman became an honored guest at the White House and a favorite adviser during the Clinton years. A cynical observer might have concluded that it was the lack of precision in its uses that accounted for the popularity of social capital. Identifying the term with norms and values was doubly advantageous, for it placed the responsibility of economic failure directly upon the shoulders of the poor themselves. In other words, the abduction of social capital and its dilution represented the most recent iteration in a familiar tendency: using deviant morality as a principal explanation for inequality. It didn't have to be that way. Lost in the thinning out of social capital were two important elements of its early theorization. One was Granovetter's (1985, 1974) distinction between strong and weak social ties, especially significant in that it clarified the ways in which varying types of interactions mediate the relationship between persons and vital resources. Strong social ties forged by kinship or friendship give individuals a sense of belonging, security, and identity. Nevertheless, it is the weak ties formed through intermittent contact with comparatively more knowledgeable or powerful actors that allow individuals to tap into the workings of the larger society. Reciprocity, obligation, and economic resources play a part in both cases, but the effects are not the same. Upright parents who are poor and ignorant cannot provide on their own the means for their children to achieve an improved social station. To change their circumstances it would be necessary to transform not so much their values and beliefs as the character of their dealings with the groups that control information and wealth (Fernandez Kelly, 1995). This conceptual treatment thus allows for the understanding of situations in which people remain in squalor not primarily as a result of their moral liabilities, but as a consequence of their diminished capacity to bridge over to actors who hold access to social and economic resources.
76
The second overlooked aspect of social capital, brought to attention by Portes and Landolt (1996), is its paradoxical and some times devastating effect. Contrary to rosier interpretations, social capital has its down side. Reciprocal transactions on the basis of kinship or friendship can be a powerful deterrent to economic progress and political participation. The benefits of group membership - trust, safety, and self-definition - may harden the barriers between individuals and resources of high quality. Street gangs are a case in point. Their confederates stand in a tight relation of obligation and mutuality to one another. Together they back and support one another. It is not for lack of social capital that they languish at the bottom of the economic and political hierarchy.
IS CULTURE A FORM OF SOCIAL CAPITAL?

The paper by Goyette and Conchas makes a significant contribution to the study of extra-familial factors affecting varying degrees of educational success among immigrant children. The authors compare two groups with apparently similar characteristics: Vietnamese and Mexican students. They rely on two data sources, the 1988-1990 National Educational Longitudinal Study and qualitative data from a two-year study of minority students in a California high school. The overarching purpose of the research was to identify and explore one mechanism through which social capital affects educational outcomes: these students' study habits. At first blush, the paper's aim seems puzzling. There is little dispute that the length and intensity of after-school study has positive repercussions in student performance and prospects. Yet, the authors' main concern is not solely to confirm what other researchers have said before but to explain the role of peer networks in the development of study habits. In the authors' view, too much has been made of parental responsibility, to the neglect of outside influences, as a factor shaping children's school performance. The paper thus begins with a similar set of observations as those presented at the beginning of this commentary. The similarity ends there, however, and the paper's strong empirical contribution becomes sharply divided from its theoretical underpinnings. This does not detract from its importance, but it provides an opportunity to discuss in greater detail the uses and misuses of social capital. Goyette and Conchas begin with a dangerously elastic definition. Although they cite known names in the relevant literature, they deploy a weakened version of Putnam's already watered-down concept. Here social capital consists of "less tangible [resources] such as norms, encouragement, and information gained from relationships and social networks." Several questions and potential problems
77
arise from that conceptualization. First, it allows for educational success to be imputed a posteriori to the group's value system, just as in Putnam's case. This creates the opportunity for tautological reasoning. Desirable outcomes in'standardized testing and top grades are automatically attributed to a regard for study and intellectual improvement without a need to review the factors that link values with outcomes. A high regard for education may lead to educational success but the opposite is also true: mastery of the educational contest spins narratives about the worth of education. Is it that collectivities with the right values thrive educationally, or is it that those who succeed in education then cultivate mental and behavioral habits that enhance educational success? Both possibilities exist and are part of a larger iterative process across generations. Second, social capital is reified to be recast as a resource, not the panoply of mechanisms that connect individuals to resources. Surely almost anything can be construed as an asset given the right circumstances. But such amplification hampers understanding. Coleman is known for his enduring faith in the role of social networks in the educational achievements of the young, but his theoretical writings make clear that social capital is realized at the point of connection between individuals not in the realm of their ideas or normative convictions. Individuals may cherish education without attaining educational success. This is particularly true among some racial and ethnic minorities and many working-class people in general. Even the most cursory inquiry reveals that it is not the regard for education that is unevenly distributed among groups but the means to make such regard "work" effectively Finally, equating social capital with culture renders the first term superfluous. Goyette and Conchas provide evidence of this. "Culture" or "norms and values" can be substituted in every single statement where the authors use "social capital" with minimal or no effect on meaning. Why inflict another esoteric term into the already crowded sociological repertory unless there is proper justification? Terminological ambiguity leads to other confusions. For example, the purported similarities between Vietnamese and Mexicans may be as significant as the differences between the two groups. Mexicans constitute the longest uninterrupted migration to the United States, and while it is true that the National Educational Longitudinal Study covers "many children of recent immigrants," many more are youngsters whose forbears have been in this country for much longer than the Vietnamese have. That is important because extended residence in the United States, in tandem with aborted or limited socioeconomic mobility, is associated with a reduced belief in education as a path toward success. Poor study habits among Mexicans echo this reality.
78
In the same vein, the resemblance in economic status and racial bearing between the two groups may be deceptive. Among Vietnamese refugees, a low standing in American society is a reflection of their recent arrival but not necessarily of their class background prior to migration. The opposite is true for Mexicans: the vast majority were humble laborers before arriving in the United States, and many retain a lowly class status in the United States. Class differences between the two groups may more strongly explain their disparate educational performance than the inferred variation in their mores. Findings from the Children of Immigrants Longitudinal Study (Portes & Rumbaut, 2001) backs, to some extent, the previous statements. CILS is the largest and most reliable research conducted about generational variations among recent immigrant families. Many of its findings coincide with those reported by Goyette and Conchas. Mexican children lag well behind their Vietnamese counterparts in standardized test scores and grade point averages. On the other hand, the aspirations of parents and children are high in both cases. The main divergence is between aspirations and expectations. While the Vietnamese count on succeeding at the highest educational levels, Mexicans anticipate that their longings will meet with limited success. This may be interpreted not as a function of varying social capital, or mores and values, but as the effect of each group's accumulated human capital and its form of incorporation into American society. Especially (but not exclusively) in California, where Goyette and Conchas conducted the qualitative part of their study, many Mexicans stand at the bottom of the social ladder. A recent study shows that they face greater disdain from white residents than do African Americans (Steffensmeier & Demuth, 2000). Historically, they have endured prejudice and served as scapegoats during economic downturns. Punitive legislation passed in the 1990s further soured the climate in which children go to school and develop expectations. By contrast, the Vietnamese, most of who came to this country as refugees, have faced considerable hostility and financial duress at the local level, but they also have received government support of their settlement process. This paired with a higher socioeconomic status in their country of origin, their educational success in the United States is admirable but not surprising. Although both groups occupy a minority status in their adopted country, the character and scale of their racial experience couldn't be more dissimilar. Goyette and Conchas say as much in one of the most enlightening sections of their paper. They dispel the misimpression that Mexican parents are not involved in their children's education. By every measure - frequency of interaction with teachers, participation in PTAs, active encouragement of youngsters' educational yearnings and knowledge of children's friends - Mexicans score
79
higher than Vietnamese. Yet the authors also note that Vietnamese children are more likely to attend schools with smaller numbers of poor students than do Mexicans children. In other words, Mexicans experience higher levels of neighborhood segregation in terms of class and ethnicity than do the Vietnamese. As shown by various scholars, Massey and Denton (1993) and Kasinitz (1996) in particular, spatial segregation is tantamount to the severing of ties with external groups whose members may have greater riches, knowledge, and expertise. It is primarily segregation that accounts for the scarcity of weak ties available to Mexican students. The same is true in even sharper relief for inner-city blacks. Beyond generational differences what matters most is the quality and quantity of contacts outside one's primary group. Against this backdrop it is understandable that many Mexican youngsters have given up on education as a path toward advancement. I liken this shift to the experience of the fox in Aesop's celebrated fable: exhausted, dejected, and unable to reach the lustrous grapes that had aroused his desire in the first place, he pretended not to want them in order to salvage his pride. It is easy to despise what you cannot have. So it is for many students in vulnerable groups. Without outside contacts to help them actualize their aspirations, they seek the assurance of strong ties among family and friends. The pressure to leave the educational system is strong when further complicated by the dubious allure of teen culture and the need to contribute to family income. Mexican Americans are the most likely of all groups to abandon school and seek employment prematurely. In the larger picture, it is not the corrosive influence of adolescent mores but the dearth of external connections that shapes the experience of mexican youths.
PARALLEL INSTITUTIONS AND THE STRENGTH OF WEAK TIES

The paper by Bankston and Zhou aptly complements Goyette and Conchas's contribution. They too question the overriding power of families to determine the educational performance of children, highlighting instead the impact of immigrants' religious institutions. Using data from the National Longitudinal Study of Adolescent Health, they find that average grades achieved by immigrant children can be explained by their religious involvement. The study's comparative reach is impressive: beyond immigrants it covers native-born groups, including African Americans and Hispanics. I will leave aside an obvious question in the framing of this study, that of endogeneity. It is entirely possible that the correlation between religious participation and successful educational performance is the result of self-selection. This,
80
however, does not discount the important role that religion plays in how immigrant parents or children adapt. One of the paper's chief virtues is to focus upon a subject that has received too little attention. Although early sociologists highlighted the part played by religion in immigrant adaptation (Thomas & Znaniecki, 1927), two generations of scholars exposed to the radical changes of the 1960s neglected it. Only recently are the pivotal roles of religious narratives and church affiliation among immigrants being reassessed. The authors begin by recounting Coleman's (1988) idea that network closure, expressed through persistent and intimate relations between parents and children, leads to educational success, especially among Asian immigrants. Not necessarily, say Bankston and Zhou, noting an insufficiently acknowledged fact: migration can profoundly subvert gender roles, class standing, and even the authority of parents vis-a-vis their children. Adults may be forced to depend on their offspring to interpret and negotiate their way in American society. This is especially true, and at times traumatic, among Asians who struggle more than other groups with mastery of the English language. The need for immigrant women to work outside the home also may lessen maternal intimacy. Aware of their parents' limited knowledge and power, youngsters may turn defiant. Reduced parental stature can thus push adolescents toward alternative sources of guidance and information. What makes up for the reduced capacity of families to bolster the educational fortunes of their children? Bankston and Zhou point to the existence of parallel institutions, such as religious organizations, that supplement and reinforce parental messages. As in the Goyette and Conchas paper, here too some easy assumptions about immigrants are dispelled. Contrary to what Coleman suspected, immigrants are not more likely to experience closure than other groups. In fact, the opposite is true. Judging by several criteria, immigrant children are the least supervised of all the groups considered by Bankston and Zhou. In fact, being a first generation immigrant is more strongly related to lower parental oversight through network connections than is race or ethnicity. It is against this backdrop that the authors examine religious participation. Helpfully, the authors point out that immigrants tend to be even more religious in their adopted countries than in their countries of origin. It deeply tries the soul to leave behind kith and kin, to confront the hazards of unknown lands and to face displacement without the guarantee of acceptance in their new home. Religious narratives, with their emphasis on suffering and redemption, their images of roads traversed and labor ultimately rewarded, give meaning to the immigrant experience. It is not unusual, therefore, for immigrants to cling to religiosity or to experience conversion in their adopted countries. Immigrant children are religiously involved to a much larger extent than natives. Nearly
81
a quarter of the immigrant children studied by Bankston and Zhou participate in religious institutions, compared to only four percent of the native born. Native born Asians are no more likely to be religiously active than are whites. The opposite is true for Latinos, who exhibit a higher propensity to participate in ethnic religions and denominations than whites. What makes Bankston and Zhou's analysis intriguing is the discovery that religious involvement may have different effects depending on the kinds of institutions into which individuals are sorted. Involvement in ethnic religious institutions is correlated with lack of network closure as defined by Coleman, but closure among immigrants is related to children's participation in nonethnic religious institutions. Similarly, participation in ethnic religiosity is not necessarily correlated with outstanding educational performance -joining in mainstream institutions is. The authors discuss several possible reasons for the dichotomy and remain inconclusive. It is possible, they suggest, that ethnic religiosity operates as a compensatory mechanism among youngsters who experience undue stress at home. The correspondence of educational success and partaking in mainstream churches may be the effect of selfselection. Surprisingly, however, Bankston and Zhou do not consider a third plausible alternative: the differential effects of ethnic and mainstream institutions may be related to the varying kind and quality of connections that they offer participants. Ethnic institutions, the authors note, provide settings for sociability, a strong sense of belonging and, equally important, business opportunities. Churches, mosques and temples are abuzz with what is commonly known as networking. This insight brings strong and weak ties to mind again. In his important contribution to the understanding of the mechanisms that connect individuals with jobs, Granovetter (1974) was among the first to note that the most successful employment searches occur by word of mouth, through casually extended tips in social settings outside the labor market. Religious institutions thus fulfill more than spiritual functions; they also create an atmosphere of trust and reciprocity where individuals can exchange tidbits of information about the workings of the economic system. On the other hand, ethnic institutions almost by definition do not provide a large number of external links. Their role is to maximize internal resources and to supply the needs of groups that have either recently arrived, as in the case of Asians, or been left behind by the hardening of ethnic demarcations, as in the cases of Mexican and African Americans. By contrast, participation in mainstream religious institutions points to membership in flourishing segments of American society. Thus, it's understandable that this kind of religious engagement is positively correlated with network closure, as specified by Coleman,
82
and with enhanced school achievement. The authors are right in calling for a conceptualization of social capital that gives attention to institutions. After all, institutions are, by far, the mother lode of weak ties.
CONCLUSION
I began this commentary with a personal remembrance that underscores the varying character of interrelationships at play in youngsters' educational journeys. Chicanos at Princeton depart from the normative profile of Mexican Americans in the extent to which they have been able to forge ties with individuals, groups, and institutions external to their own ethnic group. Their exceptional stamina, faith, and intelligence surely affected their trajectory. The support and encouragement offered by parents was not insignificant. But decisive in the constellation of factors that determined their educational success was the active presence of persons mediating between themselves and institutionally structured opportunities. As we ponder the uses and misuses of social capital, let's note again that it is in the relational character of the concept that its considerable worth rests. When we speak of social capital as if it were a thing, or when we extend its definitional bounds in excess, we throw away the possibility of better understanding how social networks function. Personal character and collective culture are important, but they are realized or thwarted only through the presence or absence of bridges that span individuals and economic resources. On the basis of outstanding empirical research, the two papers give life to that assertion.
REFERENCES
Bourdieu, P. (1986). The Forms of Capital, In: J. G. Richardson (Ed.), Handbook of Theory and Research for the Sociology of Education. New York: Greenwood Publishers. Coleman, J. S. (1988). Social Capital in the Creation of Human Capital. American Journal of Sociology, (Supplement) 94, S95-S120. Fernandez Kelly, P. (1995). Social and Cultural Capital in the Urban Ghetto: Implications for the Economic Sociology of Immigration. In: A. Portes (Ed.), The Economic Sociology of Immigration: Essays on Networks, Ethnicity, and Entrepreneurship. New York: The Russell Sage Foundation Press. Granovetter, M. (1985). Economic Action and Social Structure: the Problem of Embeddedness. American Journal of Sociology, 91, 481-493. Granovetter, M. (1974; 1995). Getting a Job: A Study of Contacts and Careers. Chicago: Chicago University Press. Kasinitz, P. (1996). Missing the Connection: Social Isolation and Employment on the Brooklyn Waterfront. Social Problems, 43(2) (May), 180-196.
Commentary:
Uses and Misuses of Social
Capital
83
Massey, D. S., & Denton, N. A. (1993). American Apartheid: Segregation and the Making of the Underclass. Cambridge MA: Harvard University Press. Portes, A., & Landolt, P. (1996). The Down Side of Social Capital. The American Prospect, 7(14)(May/June), 80-87. Portes A., & Rumbaut, R. (2001). Legacies: The Story of the Immigrant Second Generation. Berkeley and New York: University of California Press and Russell Sage Foundation Press. Putnam, R. (1993). The Prosperous Community: Social Capital and Public Life. The American Prospect, 4(13) (March), 21-30. Putnam, R. (2000). Bowling Alone: The Collapse and Revival of American Community. New York: Simon and Schuster. Steffensmeier, D., & Demuth, S. (2000). Ethnicity and Sentencing Outcomes in U.S. Federal Courts: Who is Punished More Harshly? American Sociological Review, 65(5)(October), 705-729. Thomas, W. I., & Znaniecki, F. (1927). The Polish Peasant in Europe and America. New York: Alfred A. Knopf.
ETHNIC DIFFERENCES IN PARENTS' EDUCATIONAL ASPIRATIONS*

Grace Kao
ABSTRACT
/ examine race and ethnic patterns in parental educational aspirations and savings behavior for their children using the National Education Longitudinal Study of 1988 (NELS). Despite the well-documented role that parents play in encouraging academic achievement, little is known about parental aspirations and savings behavior and variation by race and ethnicity. Some have speculated that the lower educational outcomes of minority youth may result from a lack of educational and material resources. This paper shows that minority parents have extremely high educational aspirations for their children, and these aspirations exert considerable influence over youths' aspirations and the maintenance of these aspirations over time. However, only Asian American parents have consistently high aspirations for college buttressed by a high level of savings compared with others with similar family incomes. Immigrant parents are extremely optimistic about their children's eventual educational attainment.
85
86
GRACE KAO
INTRODUCTION
This paper is motivated by two well-known relationships. First, we know from the status attainment model that parents' educational attainment contributes to their children's eventual educational attainment through educational aspirations (Blau & Duncan, 1967). Thus, in general, we would expect parents with high levels of education to have higher educational aspirations for their children. Presumably, parental aspirations for their children influence children's aspirations for themselves. Second, we also know that, despite the lower socioeconomic status of blacks and Hispanics, black and Hispanic youth tend to have extremely high aspirations relative to whites (Hauser & Anderson, 1991; Kao & Tienda, 1998). The explicit prediction of the status attainment framework is that minority parents would have lower aspirations for their children due to their lower socioeconomic backgrounds. However, few studies have actually examined minority parents' aspirations, savings behavior and their influence on children's aspirations (for studies on immigrant versus native-born youth, see Hao & Bonstead-Bruns, 1998; Glick & White, 2000). Specifically, this paper examines several questions. First, are minority parents, by virtue of their lower socioeconomic status, less optimistic about their children's eventual educational attainment than white parents? Alternatively, it may be that the relatively high educational aspirations of minority youth may come from parents' optimism about their children's eventual attainment. Second, do parental educational aspirations affect the maintenance of children's aspirations over time? To unravel the relationship between parents' and children's aspirations, I first present a brief review of previous research on how parental socioeconomic status translates into high educational aspirations and performance among children. Here, I argue that while parental socioeconomic status influences the financial resources available for educational purposes, parental optimism can still affect children's aspirations and their maintenance over time. I argue that while many 8th graders report high educational aspirations, most children do not maintain these aspirations throughout high school. Youth who report high educational aspirations and maintain them over time are most likely to be successful in reaching their goals (for a theoretical discussion of this idea, see Alexander & Cook, 1979). Finally, I summarize the primary findings of this paper.
PREVIOUS RESEARCH
The association between parents' socioeconomic status and children's educational outcomes is one of the most widely believed and thoroughly tested
Ethnic Differences in Parents' Educational Aspirations
87
propositions in the social sciences. Parents provide both material and interpersonal resources in promoting their children's educational outcomes. It is clear that one of the advantages high socioeconomic status affords is parents who are better equipped and informed about educational choices. What is less clear is how race, ethnicity, and immigrant status additionally influence parental aspirations for youth and savings for college. Cultural differences in understandings about parental and children's responsibilities can differ among ethnic groups. For example, beliefs about parental versus children's responsibilities among Asians may further motivate parents to save for their children's education, regardless of income. Just as children are expected to achieve in school, Asian parents may feel that financing college is primarily a parental responsibility. However, Hispanic and Asian parents are most likely to be LEP (limited english proficient); hence they are at the greatest risk of being uninformed and overwhelmed by the daunting task of applying for financial assistance. Minority and immigrant status can also affect the degree of optimism parents have about their children's odds of success. I am interested in how ethnicity and immigrant status influence parental educational aspirations, and how these influence the fluctuation of children's own aspirations as they progress through high school. While family savings for college and aspirations are certainly correlated with socioeconomic status, they are by no means equivalent - families with comparable incomes can make dramatically divergent choices about how much money to reserve for educational expenses. Presumably, parents who have begun thinking about how to finance their children's college educations in 1988 will be better prepared to assist children in 1992. Their preparation also sends a strong signal to youth at a relatively early age (8th grade) that parents expect them to attend college. Parental attitudes towards college finances may also be driven in part by parents' own experiences as children, which are affected by their parents' socioeconomic background, race, and personal preferences. For instance, Steelman and Powell (1991) found that parents whose own parents provided financial assistance for college were more likely to pay for their children's college expenses. Given the higher educational levels of Asian American parents relative to their minority counterparts, it may be that parental networks help convey information about the cost of higher education. In order to further understand how minority and immigrant parents influence their children's educational aspirations, I examine ethnic and immigrant differences in home environments, knowledge aboutfinancingcollege, actual savings behavior, and finally, parents' aspirations for youth.
88
GRACE KAO Educational Resources at Home
One of the recent outgrowths of status attainment research has been the examination of educational resources. A resource more indirectly related to income, as Bourdieu (1977) and DiMaggio (1982) postulate, is cultural capital that is distinct from class position. Such resources stem from increasing familiarity with middle class activities, such as visiting art museums or going to the symphony. DiMaggio (1982) argues that exposure to "middlebrow" activities can increase children's grades simply by conveying somewhat of a privileged status of children to teachers. Parents can also promote educational outcomes more directly. One method is to create a home environment conducive to educational activities; this might include books, a computer, and reserving a place specifically for studying (Murnane, Maynard & Ohls, 1981; Teachman, 1987). Of course, not all parents can afford such goods, but most children in NELS reported that they have their own room, which suggests that most households can reserve a place for studying (see Kao, 1995). Parents can also directly manage their children's academic careers. For instance, Baker and Stevenson (1986) found that mothers with college educations were more likely to choose college preparatory classes for their children, regardless of their children's grade performance. High SES mothers were also better informed than their lower SES counterparts about their children's scholastic activities. Knowledge about Financial Costs and Savings Behavior For children who aspire to attend college, parental savings for college can be an important indicator of the likelihood of reaching their goals. As youths progress through high school, the knowledge that parents have not saved any money for their college expenses (especially when they are able to do so) may lower college aspirations. First and foremost, a family's ability to save is determined by their income and the number of children who share that income (Steelman & Powell, 1991; Hossler & Vesper, 1993). Previous studies of parental savings for post-secondary schooling have also found that such behavior is associated with parental motivation to save, and parental aspirations for their children (Hossler & Vesper, 1993). Parental savings can also be indicative of parental awareness of the expense of post-secondary education, and signal the degree to which parents are acquainted with higher education more generally. For instance, Case and McPherson (1987) argue that some families do not act rationally in their savings behavior because they are unaware about financial aid and fail to anticipate the
89
costs of a college education (from Hossler & Vesper, 1993). It is logical to suspect that immigrant parents, especially those whose English skills are lacking, may be most uninformed or intimidated about financial aid. However, it is unclear how such a lack of informational resources would affect parental savings, or if such fears influence savings at all. One reaction among families with high incomes is to save more than their counterparts with comparable incomes so as to not rely on (or rely less on) financial aid, as Hossler and Vesper found (1993). Of course, those families who are most intimidated by and least knowledgeable about college costs may elect to forgo any desire to attend college and decide not to save for such an unlikely event. Finally, it is unclear from earlier studies whether or how race and ethnicity affects savings behavior. One recent study, which examined racial differences between blacks and white when their children were in 11th grade, found no significant differences in the propensity to save once they controlled for other parental attitude variables in willingness to pay and knowledge of costs (Hossler & Vesper, 1993). Another study by Steelman and Powell (1993) found that minority parents were more likely to accept the burden offinancingtheir children's education and were also more likely to support policies about government financial aid than white parents. How these attitudes translate to actual savings behavior is not explicit. Race and Ethnic Variation in Parental Aspirations Despite the long-standing tradition in status attainment research that links parental aspirations with children's own aspirations along with several decades of research in exploring racial variation (between blacks and whites) in children's educational aspirations, there is a great paucity of research on ethnic differences in parental aspirations. There is some evidence that black parents are more optimistic than their white counterparts about their children's aspirations, but there has been little discussion of why this is the case. In addition, Ogbu's (1991) research suggests that black parents may send mixed signals to children about whether education is the most rational path to obtaining socioeconomic mobility for blacks - however, what this implies for the level of black parents' aspirations is unclear. Some notable exceptions to the general lack of attention to minorities include an early exploration of immigrant aspirations that compared rational assessments (based on prior socioeconomic status) to motivational (optimistic) explanations (Portes, McLeod & Parker, 1978). They found that the variation in adult immigrant aspirations for themselves could be explained primarily via rational assessments. However, in a more recent paper, Kao and Tienda (1995)
90
GRACE KAO
argued that immigrant parents' optimism was responsible for the higher achievement levels of their children since both first and second generation children earned higher grades, test scores, and were more optimistic about their future attainment than their third generation counterparts. In other words, adult migrants have modest expectations for their own attainment, but the possibility of socioeconomic mobility in the next generation makes the hardship of immigration to the U.S. worthwhile. In fact, these two findings together resonate with the idea of the "American Dream" where sacrifice and hard work of the first generation will reap rewards for generations to come.
DESCRIPTIVE FINDINGS
Because socioeconomic background plays a vital role in determining parental aspirations and the ability to save for children's educations, I begin by examining race and ethnic differences in parental socioeconomic status in Table 1. As is well documented elsewhere, Hispanic and black families have much lower family incomes than whites or Asians. These differences are mirrored in mother's and father's educational levels. About 40% of Hispanic mothers and fathers never completed high school, while about 20% of black mothers and fathers and Asian mothers have such low levels of education. This is in stark contrast to about 10% of white mothers and fathers and Asian fathers who do not have high school diplomas. Overall, Asian parents have the most education, although they do not earn more money than white parents. Significant race and ethnic differences in family immigrant status also exist. About 84% of Asian youth in NELS have immigrant parents, compared to about 50% of Hispanic, 8% of black and 6% of white youth. Hispanic youth have the most siblings (2.8 on average), closely followed by African American, Asian, and whites. Finally, there are also significant race and ethnic differences in educational aspirations of parents. About 80% of Asian American parents aspire to at least a 4 year college degree for their children compared to about 62% of whites, 58% of blacks, and 50% of Hispanic parents. More relevant to our concerns is how race and immigrant status influence parental aspirations and savings once we control for socioeconomic levels. Table 2 presents tabulations by race and ethnicity overall and within socioeconomic groups. The three socioeconomic groups labelled "low" "mid" and "high" are parents who fall into the lower, middle two, and upper quartiles of the socioeconomic scale, as calculated by NCES. The scale is composed by equally weighing mother and father's education, family income, and occupational scores.1
Ethnic Differences
in Parents'
Educational
Aspirations
91
Table 1. Descriptive Characteristics of Parental Socioeconomic Status (Means and percentages; standard deviations in parentheses).
Asian Mother's Education" Less than HS HS Grad Some college College Grad Graduate School Meanb Father's Education 3 Less than HS HS Grad Some college College Grad Graduate school Meanb Family Income Immigrant Parent? Number of siblings Parents aspire college? N Hispanic Black White
16.6 17.0 30.5 23.6 12.2 13.957*** (2.499) 9.7 13.6 30.7 22.3 23.8 14.738*** (2.499) 48675 (43936) 0.841*** 2.327*** (1.610) 0.804*** 1409
40.9 20.7 30.9 4.7 2.9 12.159*** (2.151) 41.6 18.4 27.3 6.4 6.3 12.348*** (2.436) 26706*** (24940) 0.495*** 2.835*** (1.730) 0.498*** 2844
19.0 25.6 43.5 6.7 5.2 13.072*** (2.075) 21.1 32.0 30.6 8.7 7.5 12.988*** (2.277) 23334*** (23334) 0.082** 2.745*** (1.811) 0.578*** 2461
9.9 29.2 38.7 14.6 7.6 13.612 (2.098) 11.0 23.8 32.6 16.3 16.3 14.060 (2.443) 46964 (39579) 0.064 2.10 (1.491) 0.619 15593
*** Mean is statistically significantly different from whites at p < 0.001 level. ** Mean is statistically significantly different from whites at p<0.01 level. * Mean is statistically significantly different from whites at p < 0.05 level. a Education Measures compiled by using parental responses when available and filling in missing item response with 8th grader's responses to these items. b Mean of educational levels calculated by setting "Less than HS"= 10, "HS Grad"= 12, "Some college" = 14, "College grad" = 16, and "Graduate school" = 18. Source: The National Education Longitudinal Study of 1988.
Overall, Asians have the highest and Hispanics have the lowest parental aspirations for their children, while black and white parents fall somewhere in between. Over 80% of Asian parents hope that their 8th grader will at least graduate from college, while only about 60% of whites and blacks, and 50%
92
GRACE KAO
Table 2. Race and Ethnic Differences in Parental Aspirations for Children and Propensities for Reserving Resources for Education: Overall and Within SES Levels.
Asian OVERALL Parent's Aspirations for Child Less than HS 0.1% HS Grad 6.3 Some College 13.2 College Grad 33.8 Masters Degree 17.4 M.D./Ph.D. 29.2 Mean Aspiration 16.987*** (2.447) Hispanic Black White
0.7% 15.9 33.6 26.5 9.5 13.7 15.388*** (2.520)
0.4% 15.0 26.9 32.1 11.7 13.9 15.625 (2.485)
0.4% 11.0 26.7 41.4 11.3 9.1 15.589 (2.175)
Financial Resources Earmarked for College 0.381*** Money saved? 0.620*** 1340.920*** If so, how 4145.698*** (5593.010) (3212.123) much now? 7693.959*** If so, how 10531.783*** (5074.144) much later? (5005.137) N LOW SES Parent's Aspirations for Child Less than HS 0.4% HS Grad 13.0 Some College 27.6 College Grad 31.0 Master's Degree 10.5 M.D./Ph.D. 17.6 Mean Aspiration 15.816*** (2.560) 1341 2714
0.417*** 1364.670*** (3162.961) 8197.576*** (4966.733) 2700
0.526 2934.759 (4760.318) 9639.981 (4948.157) 15237
1.1% 25.1 39.1 19.5 5.8 9.5 14.641*** (2.427)
0.6% 25.2 34.1 25.0 7.4 7.6 14.727*** (2.349)
1.6% 29.4 44.9 17.0 3.6 3.5 14.042 (1.984)
Financial Resources Earmarked for College 0.426*** 0.243 Money saved? 1326.923** 323.621*** If so, how (2991.738) (1226.383) much now? 7098.361 4367.347*** If so, how (4988.169) much later? (4023.999)
0.237 416.058*** (1532.319) 6002.857 (4714.400)
0.264 751.943 (2417.330) 6190.141 (4882.060) 2432
239
1311
1129
93
Table 2. Continued.
Asian MID SES Parent's Aspirations for Child 0.2% Less than HS 8.1 HS Grad Some College 14.9 37.7 College Grad Master's Degree 19.3 19.8 M.DVPh.D. 16.545*** Mean Aspiration (2.370) Financial Resources Earmarked for College 0.576*** Money saved? If so, how 2771.205*** much now? (4454.543) If so, how 8958.678* much later? (4893.508) N HIGH SES Parent's Aspirations for Child Less than HS 0.0% HS Grad 1.5 Some College 5.0 College Grad 30.9 Master's Degree 18.4 M.D./Ph.D. 44.1 Mean Aspiration 17.974*** (2.086) Financial Resources Earmarked for College Money saved? 0.725* If so, how 6358.491*** much now? (6335.457) If so, how 12257.310*** much later (4376.978) N 537 0.0% 2.9 6.2 41.7 19.2 30.0 17.342*** (2.120) 0.641 4087.413** (5346.031) 10873.596 (4713.489) 307 0.0% 2.8 6.2 35.9 20.4 34.7 17.560*** (2.154) 0.685 3891.941** (4976.253) 11076.087 (4303.661) 323 0.0% 2.1 8.0 53.6 20.1 16.1 16.800 (1.848) 0.676 5068.158 (5814.751) 11345.909 (4511.518) 5214 565 0.4% 8.6 34.8 30.7 11.2 14.3 15.735*** (2.361) 0.444 1552.980* (3226.608) 7925.876 (4672.291) 1096 0.4% 8.9 25.7 37.7 13.3 14.1 15.938*** (2.320) 0.479 1450.851*** (3134.172) 7885.193 (4787.328) 1248 0.3% 11.3 33.7 40.9 7.6 6.1 15.252 (2.000) 0.475 1869.975 (3567.778) 8195.667 (4728.265) 7591 Hispanic Black White
*** Significantly different from whites at p< 0.001. ** Significantly different from whites at p< 0.01. * Significantly different from whites at p < 0.05. Source: The National Education Longitudinal Study of 1988 (NELS).
94
GRACE KAO
of Hispanics do so. About 30% of Asian parents expect their child to earn a doctorate degree, while this figure ranges from 9% to 14% for the other groups. Even within socioeconomic status groups, Asian parents have higher aspirations for their children than all other groups. Similarly, black and Hispanic parents within all SES groups also have higher aspirations for their children than their white counterparts. These tabulations indicate that minority parents are more optimistic about their children's eventual attainment than would be expected from their socioeconomic background. While high parental aspirations alone may promote their children's own academic goals, without financial resources or financial planning for college expenses, their children's likelihood of attending and graduating from college can be severely diminished. Parents' actual savings as well as planning for college expenses also corroborate their educational aspirations for children. Their college savings and planning behavior four years prior to their children's entrance to college also demonstrate the extensiveness of their commitment to their educational aspirations for their children. Asian parents are notable for their high aspirations for children along with their greater propensities to earmark resources for education (Kao, 1995). Kao (1995) found that Asian and white families have comparable incomes, yet a greater proportion of Asian parents have started saving money for their children's educational goals. This pattern of increased savings behavior among Asians is verified in subsequent panels, which show Asians' greater propensity to save for school expenses within each SES subgroup. In addition, overall and within SES subgroups, compared to others who have begun saving money, Asians have saved more money, and plan to have more money saved by the time their 8th grader enters college. This pattern hints at the overwhelming extent to which Asian parents are motivated towards educational goals for their children (Kao, 1995). On the other hand, black and Hispanic parents overall are less likely than white parents to have begun saving money, although their average family incomes fall substantially below that of white parents. In fact, these tabulations show that there are no differences in the likelihood of having any savings for college between black, Hispanic, and white parents within SES subgroups. However, black and Hispanic parents have saved less money by 8th grade (1988) than white parents from comparable SES backgrounds. Despite their lower savings, black and Hispanic parents are optimistic, and plan to have savings for college comparable to their same SES counterparts by the time their child begins college (1992). At least two mechanisms might account for why black and Hispanic parents have fewer savings for college than whites. First, because financial resources
95
are susceptible to resource dilution as the number of siblings increases, and since black and Hispanic families have greater numbers of children than white families, these differences may be completely accounted for by ethnic differences in family income and the number of children who must share these limited resources (Steelman & Powell, 1993; Downey, 1995).2 Second, minority families may also differ from white parents in their attitudes about how to finance their children's post-secondary education.
EMPIRICAL ANALYSES
I examine the determinants of parental aspirations for children to finish college in Table 3. Results confirm findings from the descriptive tabulations. Race, ethnic and immigrant differences in parental aspirations were most striking. In Model Table 3. Logistic Regression Estimates of Effects of Family Background and Immigrant Status on Parental Aspirations for Child to Complete College.8
Model 1 Constant Asian Hispanic Black Mother's Education Father's Education Family Income Number of siblings Immigrant Parent Psuedo R2
a
Model 2 -2.391*** (0.053) 0.548*** (0.099) 0.115 (0.060) 0.460*** (0.056) 0.354*** (0.015) 0.356*** (0.014) 0.150*** (0.008) -0.010*** (0.003) 0.506*** (0.061) 0.20
-2.380*** (0.052) 0.927*** (0.087) 0.327*** (0.054) 0.457*** (0.056) 0.349*** (0.015) 0.357*** (0.014) 0.152*** (0.008)
0.19
N = 19,132 ***p<0.001; **j><0.01; * / J < 0 . 0 5 . Source: The National Education Longitudinal Study of 1988.
96
GRACE KAO
1, black, Hispanic, and Asian parents are more likely to have college aspirations for their children than their'white counterparts with comparable education and income.3 Specifically, after inclusion of measures for parental education and income, I find that Asian American parents are 2.5 times as likely to aspire to a college degree as white parents, while Hispanic parents are 1.4 times as likely and African American parents 1.6 times as likely as white parents. Model 2 adds measures of sibship size and parents' immigrant status to the model. Note that inclusion of parental immigrant status wipes out the Hispanic effect, suggesting that the positive impact of Hispanicity is due more to Hispanic parents who are immigrants than their counterparts who are U.S.-born. The effect of Asian Americans is also reduced; but recall that about 85% of Asian American youth in NELS have immigrant parents. Hence, the total effect for the typical Asian American youth in Model 2 is similar to Model 1. Specifically, immigrant parents are 1.7 times as likely as nonimmigrant parents to aspire to a college degree. Of course, parental behavior towards educational goals are important only in as much as they affect their children's aspirations. Thus, to examine how race and ethnicity mediate the effects of parental aspirations on children's aspirations, I begin by using descriptive tabulations of race and ethnic differences in the likelihood youth raise, lower, or maintain their aspirations from 8th to 10th and 10th to 12th grade, given their parents' aspirations for them in 8th grade in Table 4. The table is presented as follows. Each panel documents the maintenance of college aspirations from 8th to 10th grade (or 10th to 12th grade) in light of parental aspirations for college. This table allows for least 3 types of comparisons for our purposes. First, I can examine how ethnic youth differ in their responses to parents' aspirations. Second, I can examine the relative strength of negative versus positive parental reinforcement on aspirations. Finally, I can examine how parental effectiveness changes over time (from 8th to 10th versus 10th to 12th grades). The first row (in bold) represents the percent of 8th graders who had college aspirations and maintained these aspirations in 10th grade whose parents also had college aspirations for them. Specifically, of those Asian youth who had college aspirations in 8th grade (8th = College) and whose parents had such aspirations for them (Parents = College), 91.2% continued to have such aspirations in 10th grade. On the other hand, among Hispanics who aspired to complete college and whose parents had similar aspirations for them in 8th grade, about 75.5% persisted in having such goals. Similarly about 86% of white youth and 79% of black youth who had college aspirations in 8th grade and whose parents had such aspirations maintained them in 10th grade.
97
Table 4. Likelihood of Maintaining College Aspirations by Parental Aspirations to Graduate from College: 8th to 10th and 10th to 12th Grade.
Asians 8th = College Parents = College 10th = College 10th = No College Parents = No College 10th = College 10th = No College 8th = No College Parents = College 10th = College 10th = No College Parents = No College 10th = College 10th = No College 10th = College Parents = College 12th = College 12th = No College Parents = No College 12th = College 12th = No College 10th = No College Parents = College 12th = College 12th = No College Parents = No College 12th = College 12th = No College
a
Hispanics
Blacks
Whites
91.2% 8.8 70.1% 29.9
75.5% 24.5 47.6% 52.4
78.8% 21.1 48.3% 51.7
86.3% 13.7 52.3% 47.7
50.0 50.0 26.4 73.6
34.1 65.9 22.8 77.2
45.3 54.7 22.3 77.7
42.9 57.1 15.7 84.3
95.5% 4.5 81.6 18.4
90.1% 9.9 71.8 28.2
88.5% 11.5 75.9 24.1
92.8% 7.2 73.4 26.6
64.2 35.8 40.6 59.4
52.2 47.8 38.9 61.1
59.7 40.3 45.8 54.2
51.4 48.6 28.5 71.5
"8th" refers to 8th grade aspirations to graduate from college; hence "8th = College" means that R's 8th aspirations was to at least graduate from college. Thus, "10th" and "12th" refers to such aspirations in 10th and 12th grades, respectively. b "Parents" refers to parent's aspirations for R when R was in 8th grade. "Parents = College" means that parents planned for R to at least finish college when R was in 8th grade.
In the transition from 8th to 10th grade, Asians were by far more likely to maintain their high aspirations from 8th grade, whether or not their parents agreed with such goals. About 91% of Asian youth whose parents agreed with their high goals, maintained their aspirations from 8th to 10th grade, while
98
GRACE KAO
comparable figures for Hispanics, blacks, and whites were 75.5%, 78.8% and 86.3% respectively. Note that these figures do not include those who had dropped out between 8th and 10th grade, but that such inclusion would accentuate the racial gap, since black and Hispanic youth are more likely to drop out of high school than Asian or white youth. Even among high aspiring 8th graders with pessimistic parents, 70% of Asian youth managed to maintain their aspirations, while only about one-half of blacks, Hispanics, and whites managed to do so. Among those with lower aspirations in 8th grade (panel 2), parents' positive reinforcement motivated about one-half of Asian youth to raise their aspirations but such parental support only influenced about one-third of Hispanic youth to change their minds in 10th grade. Hispanic youth with low aspirations in 10th grade are least likely to be affected by positive parental reinforcement (note that there is only a 11% (34.1-22.8%) difference in likelihoods of having high aspirations in 10th grade, given low aspirations in 8th grade, between those with positive versus negative parental reinforcement compared to about 25% for Asians, blacks, and whites). This pattern of low parental leverage persists for low-aspiring Hispanics in affecting the change in aspirations from 10th to 12th grade as well as for blacks in the change in 10th to 12th grade aspirations. In the transition from 10th to 12th grade, there were no apparent racial differences in the likelihood of maintaining positive aspirations in light of parental support. Specifically, about 90% of all youth who aspired to finish college in 10th grade and whose parents supported such high aspirations in 8th grade, maintained their high aspirations by 12th grade (in bold). Perhaps a more dramatic transformation is the overall strength of positive aspirations in 10th grade. Even among those youth whose parents had low aspirations for them in 8th grade, about three-fourths managed to maintain their aspirations. Hence, youth whose outlook in 10th grade was positive were quite likely to maintain these aspirations. However, positive parental reinforcement in 8th grade was still quite effective in transforming low 10th grade aspirations. The last panel shows that about 50% to 65% of youth increased their low aspirations in 10th grade by the time they reached 12th grade. Overall, youth in 12th grade were much more likely to maintain their college aspirations (or raise them to such levels) regardless of their parents' aspiration levels or their own aspirations in 10th grade. There are at least two reasons for this pattern. First, developmentally, adolescents who are 17 or 18 years old are much more independent of their parents than adolescents who are 2 years younger. Second, there is some selection bias, since those children least likely to have college aspirations have already dropped out by 12th grade.
99
Race and ethnic differences in 10th grade educational aspirations are quite prominent. In the first 2 panels, which represent the transition in aspirations from 8th to 10th grade, Asians were most likely to have high aspirations in all possible pairings of 8th grade aspirations of children and their parents. There also seemed to be hints of varying degrees of parental influences. Specifically, Hispanic youth with low 8th grade aspirations were least likely to be affected by parents' aspirations in determining their 10th grade aspirations. By 12th grade, ethnic groups became more similar in their likelihood of maintaining high aspirations, given that their parents supported such goals in 8th grade. Notably, blacks, Hispanics, and Asians were more likely than whites to form high aspirations (in both 10th and 12th grade) despite an unsupportive parental environment and their own low aspirations in the past. In 10th grade (bottom of second panel), less than 16% of whites increased their aspirations despite the lack of parental support, while such figures ranged from 22 to more than 26% for minority youth. Comparably in 12th grade (bottom of 4th panel), note that about 40% of black, Asian, and Hispanic youth aspired to graduate from college despite low parental support and low 10th grade aspirations, while less than 29% of white youth managed to turn their expectations around. To further clarify the effects of parental aspirations on children's aspirations, I use logistic regression to estimate the effects of family background, parental aspirations, and immigrant status on 10th grade aspirations to complete college in Table 5. Model 1 only includes parental SES, race, immigrant status, and sibship size. Overall, Asians and African Americans are more likely than their white counterparts of comparable SES background to have 10th grade aspirations to complete college. Specifically, Asian 10th graders are 1.7 times as likely, while African American youth were 1.4 times as likely to have college aspirations than their white counterparts. These race and ethnic effects somewhat mask the actual differences, as most Asian 10th graders in NELS are children of immigrants (about 85%), and half of Hispanic youth are also children of immigrants. Hence the typical Asian American youth of immigrant parents is about 2.6 times as likely as his/her white counterpart (of native-born parents) to have college aspirations. Specifically, 10th graders of immigrant parents are about 1.5 times as likely to aspire to a college degree as their counterparts with native-born parents. Then, I add parental aspirations for their child to complete college and parental savings in Model 2. Students whose parents aspired to a college degree for their children in 8th grade are more likely to have college aspirations than their counterparts with low-aspiring parents. Specifically, 10th graders whose parents aspired to a college degree are more than 4 times as likely to have college aspirations themselves than their counterparts with low-aspiring parents.
100
GRACE KAO
Table 5. Logistic Regression Estimates of Effects of Family Background, Parental Aspirations and Immigrant Status on 10th Grade Aspirations to Complete College.
Model 1 Constant Asian Hispanic Black Mother's Education Father's Education Family Income Immigrant Parent Number of Siblings Parents' college aspiration Money saved? Student's 8th grade college aspiration Psuedo R2 0.14 0.20 -1.593*** (0.065) 0.553*** (0.127) -0.013 (0.075) 0.304*** (0.076) 0.229*** (0.019) 0.283*** (0.017) 0.146*** (0.010) 0.415*** (0.078) -0.009* (0.004) Model 2 -1.905*** (0.069) 0.421** (0.130) -0.039 (0.078) 0.220** (0.080) 0.146*** (0.020) 0.206*** (0.018) 0.111*** (0.010) 0.327*** (0.080) -0.006 (0.004) 1.429*** (0.047) 0.167*** (0.046) Model 3 -2.552*** (0.077) 0.395** (0.135) 0.007 (0.083) 0.227** (0.084) 0.119*** (0.021) 0.174*** (0.019) 0.102*** (0.010) 0.305*** (0.084) -0.005 (0.004) 1.021*** (0.051) 0.125* (0.049) 1.596*** (0.053) 0.26
N = 10,696 Source: The National Education Longitudinal Study of 1988.
In addition, parents who had begun saving money for college by 8th grade also had children with higher aspirations. However, the presence of money for college was less important than parental optimism about their child's eventual educational attainment. Both factors reduce the Asian and black effects but they do not eliminate their significance. In addition, the positive effect of having immigrant parents persists even after controlling for parents' own aspirations and savings. However, parental aspirations may simply reflect youth's own plans in 8th grade (when parent's aspirations were measured). To evaluate the independent effect of parental aspirations, Model 3 adds youth's own aspirations in 8th
101
grade. While youth's own aspirations in 8th grade is a significant factor in determining their aspirations in 10th grade, all of the previous effects found in Models 1 and 2 persist. Hence, parental background characteristics and their optimism about their children significantly influence high educational aspirations among youth. In sum, race and immigrant status influence parental and youth's aspirations in distinct ways. Immigrant parents are more likely to have high aspirations for their children than native parents with comparable levels of education and income. Parental aspirations and the presence of savings each exert its own influence on 10th grade aspirations. Moreover, parents can raise their children's low aspirations in 8th grade by their financial aspirations and readiness in 8th grade.
CONCLUSIONS
Overall, these analyses support the notion that race and ethnic differences in parental aspirations, if anything, coincide with the high aspirations of minority youth. However, not all high aspiring youth have high aspiring parents (or vice versa). Youth whose parents are optimistic about their children's future help t o maintain high aspirations from 8th to 10th grade and even raise low aspirations in 8th grade to higher aspirations in 10th grade. Descriptive analyses showed that within socioeconomic subgroups, blacks, Hispanics, and Asians had higher aspirations for their children than white parents. In particular, Asian parents had extremely high aspirations for their children. Hence, minority parents manage to partly overcome the adversity from the lack of material and educational resources simply through their optimistic attitudes about their children's eventual attainment. Empirical analyses suggest that African American, Hispanic, and Asian American youth all benefit from having extremely optimistic parents, and that this optimism helps booster the aspirations of minority youth. Moreover, some minority youth additionally benefit from having immigrant parents. Immigrant parents tend to be more likely to have college aspirations for their children than native-born parents, but they do not differ from nativeborn parents with respect to their savings behavior. Having immigrant parents also increases the likelihood of maintaining high aspirations from 8th to 10th grade and this relationship is independent of parental aspirations. While the greater effectiveness of immigrant parents in sustaining high aspirations has something to do with high parental educational aspirations, there is more to it than simply optimism on the part of parents. Further research on the distinct parent-child relationships in immigrant and ethnic families may uncover why
102
GRACE KAO
these parents are more effective in maintaining high aspirations than white native-born families. Finally, these results suggest that Asian American parents are better prepared to support their children's college aspirations financially, and that the Asian advantage persists across all socioeconomic groups. In other words, within socioeconomic levels, Hispanics, blacks, and whites behave similarly in their savings behavior while Asian parents are more aggressive in their savings patterns. This may stem from a lack of knowledge of or the hesitation to rely on financial aid for college. This pattern may also account for the growth of Asian American enrollment in higher education, which cannot be explained simply through socioeconomic status differentials.
NOTES
*An earlier draft was presented at the 1997 Annual Meeting of the American Sociological Association, Toronto, Canada. This research was supported by a 1998-2000 National Academy of Education and Spencer Foundation Post-Doctoral Fellowship. 1. The occupational categories and their scores are as follows: (1) Clerical = 56.58; (2) Craftsperson = 27.41; (3) Farmer = 28.00; (4) Homemaker = Missing; (5) Laborer = 7.33; (6) Manager/Administrator = 67.73; (7) Military = 38.00 (8) Operative = 19.18; (9) Professional 1 = 70.21; (10) Professional 2 = 70.21; (11) Owner/Small Business = 49.70; (12) Protective Service = 38.00; (13) Sales = 54.42; (14) Teacher = 70.21; (15) Service = 15.90; (16) Technical = 16.40; never worked or don't know were coded as missing. 2. Another source of the savings difference in black and Hispanic families is their likely difference in wealth within these rough SES categories. In fact, other studies have found that the minority-white gap in wealth is much more substantial than race differences in income. However, I cannot measure wealth in NELS because it is not designed to capture detailed household economic information. 3. Note that "college aspirations" is defined as aspirations to complete college.
REFERENCES
Alexander, K. L., &Cook, M. A. (1979). The Motivational Relevance of Educational Plans: Questioning the Conventional Wisdom. Social Psychology Quarterly, 42, 202-213. Baker, D. P., & Stevenson, D. L. (1986). Mothers'.Strategies for Children's School Achievement: Managing the Transition to High School. Sociology of Education, 59, 156-166. Blau, P. M., & Duncan, O. D. (1967). The American Occupational Structure. New York: Wiley. Bourdieu, P. (1977). Reproduction in Education, Society, Culture. Beverly Hills: Sage Publications. DiMaggio, P. (1982). Cultural Capital and School Success: The Impact of Status Culture Participation on the Grades of U.S. High School Students. American Sociological Review. 47, 189-201.
Ethnic Differences
in Parents'
Educational
Aspirations
103
Downey, D. B. (1995). When Bigger Is Not Better: Family Size, Parental Resources, and Children's Educational Performance. American Sociological Review. 60, 746-761. Glick, J. E., & White, M. J. (2000). Parental Aspirations and Post-Secondary School Participation among Immigrant and Native Youth in the United States. Unpublished manuscript presented at 2000 Meetings of the American Sociological Assocation. Hao, L., & Bonstead-Bruns, M. (1998). Parent-Child Differences in Educational Expectations and the Academic Achievement of Immigrant and Native Students. Sociology of Education, 71, 175-198. Hauser, R. M., & Anderson, D. K. (1991). Post-High School Plans and Aspirations of Black and White High School Seniors: 1976-86. Sociology of Education, 64, 263-277. Hossler, D & Vesper, N. (1993). An Exploratory Study of the Factors Associated with Parental Saving for Postsecondary Education. The Journal of Higher Education, 64, 140-165. Kao, G. (1995). Asian-Americans As Model Minorities? A Look At Their Academic Performance. American Journal of Education, 103, 121-159. Kao, G., & Tienda, M. (1995). Optimism and Achievement: The Educational Performance of Immigrant Youth. Social Science Quarterly, 76, 1-19. Kao, G & Tienda, M. (1998). Educational Aspirations among Minority Youth. American Journal of Education, 106, 349-384. Murnane, R. J., Maynard, R. A., & Ohls, J. C. (1981). Home Resources and Children's Achievement. Review of Economics and Statistics, 63, 369-377. National Center for Educational Statistics (1990). National Education Longitudinal Study of 1988: Base Year Student Component Data File User's Manual. Washington, D.C.: U.S. Department of Education, Office of Educational Research and Improvement. Ogbu, J. U. (1991). Immigrant and Involuntary Minorities in Comparative Perspective. In: M. A. Gibson & J. U. Ogbu (Eds), Minority Status and Schooling: A Comparative Study of Immigrant and Involuntary Minorities (pp. 3-33). New York: Garland Publishing. Portes, A., McLcod, S. A., & Parker, R. A. (1978). Immigrant Aspirations. Sociology of Education, 51, 241-260. Steelman, L. C , & Powell, B. (1991). Sponsoring the Next Generation: Parental Willingness to Pay for Higher Education. American Journal of Sociology, 96, 1505-1529. Steelman, L. C , & Powell, B. (1993). Doing the Right Thing: Race and Parental Locus of Responsibility for Funding College. Sociology of Education, 66, 223-244. Teachman, J. (1987). Family Background, Educational Resources and Educational Attainment. American Sociological Review, 52, 548-557.
SCHOOLING ALTERNATIVES, INEQUALITY, AND MOBILITY IN ISRAEL

Yossi Shavit, Hanna Ayalon and Michal Kurlaender
ABSTRACT
This paper examines the value of second chance opportunities for obtaining a matriculation diploma in Israel in attenuating social inequalities. Our results indicate that men, students of vocational versus academic tracks, and individuals from the lower middle range of the socioeconomic hierarchy benefit from the existence of second chance structures. We find reductions in social origin inequalities on the odds of obtaining a matriculation diploma. However, obtaining a matriculation diploma via a second chance route is associated with lower post-secondary educational attainment and weaker occupational attainment, compared to traditional routes.
INTRODUCTION
This paper examines the role of "second chance" educational opportunities in attenuating educational and occupational inequalities in Israel. Second chance opportunities in educational attainment rest on the idea that through an organized structure an individual can actualize an educational opportunity missed or failed
105
106
YOSSI SHAVIT, HANNA AYALON AND MICHAL KURLAENDER
the first time around. Specifically, we define second chance education as the existence of a formal structure by which individuals seek a second chance to either complete their secondary schooling or improve their completion status. The notion that failure is not final and that mobility can be achieved through nontraditional routes is central to educational structures that stress equal opportunity (Inbar & Sever, 1989). The opportunity for individuals to correct a past failure or to actualize their potential is often targeted at those for whom the traditional route has not been successful. The existence of second chance education has been considered a vehicle for moderating educational inequalities in society (Ayalon, 1990; Brint & Karabel, 1989; Inbar & Sever, 1986). It is important to distinguish "second chance" education from "alternative" education. Alternative education takes the form of separate curricular tracks offered at various stages of the educational system. Most systems offer a main route, in the form of a relatively demanding academic track leading to study in a university, and alternative routes, intended for less able or less ambitious students. Alternative education is not intended for those aspiring to complete a four-year college or university education. Rather, alternative routes are often vocational and prepare students for immediate entry into the labor market as skilled workers. Researchers agree that students who embark on the alternative tracks are disproportionately drawn from the lower social strata, and are diverted away from a university education and from the professions (Shavit, 1984; Oakes, 1985). Moreover, the prospects of early entrance into the labor market are said to lure some academically able, lower-class students away from college and the professions (Saha, 1985). As such, alternative education is often blamed for contributing to the reproduction of social inequality.1 The idea of second chance education is different. Second chance education is based on the belief that both schools and students often err when deciding to drop out from the main route, and that such errors can and should be corrected (Yogev, 1997). It is intended to provide students who did not succeed on the main road with another chance to do so, not via a parallel road, but rather through re-entry portals into the main track. Inbar and Sever (1989) offer the following three basic criteria by which second chance structures should be evaluated: (1) accessibility; (2) effectiveness; and (3) equivalence. They argue that second chance structures must be non-selective, effective in terms of actually improving the educational attainment of the participants, and lead to similar if not the same rewards for success as the first chance provides. The promise of educational attainment via a second chance route should not merely mirror that which is offered by the traditional educational structure in society. Rather, second chance programs
Schooling Alternatives, Inequality and Mobility in Israel
107
can offer a renewed sense of potential for disadvantaged groups in society, whose chance for mobility may be limited in the traditional route.2 Second chance education has been researched as a potential mechanism for moderating inequalities that exist within the educational structure (Ayalon, 1990; Brint & Karabel, 1989; Inbar & Sever, 1986). Previous research has explored the mechanisms themselves, the target population, organizational structure, and who makes use of second chance opportunities to meet their educational goals. The research has attempted to answer two main questions. First, to what extent do second chance programs actually enable students to reach higher education, and second, to what extent do they reduce educational inequality between social strata? Previous studies were more successful in providing answers to the first question than the second. In many cases these studies either concentrated on students of specific second chance programs, or excluded early dropouts from their samples, thus failing to perform the appropriate comparisons for the analysis of educational inequality. By using nationally representative data on birth cohorts, we can compare second chance students to both students on the traditional route as well as to dropouts, and thus address both of these questions. Scholars are generally skeptical about the success of second chance education in reducing inequality. First, it is often the case that educational institutions claiming to offer a second chance are, in practice, offering alternative routes, which tend to exacerbate class inequalities in educational attainment. For example, community colleges in the U.S. were first established with the intent that they would be second chance institutions, and most emphasized a liberal arts education. Gradually, they expanded their vocational programs and shifted to an alternative education focus (Brint & Karabel, 1989). In recent decades, the majority (about 60%) of community college students are enrolled in vocational programs (see, for example, Dougherty, 1994). Thus, many community colleges were, in fact, transformed from second chance into alternative forms of education. Second, even when second chance institutions adhere to their initial calling, they do not seem to further the subsequent attainments of students. The General Educational Development test (GED) in the United States, which serves as a high school completion certificate for individuals who have dropped out, is a case in point. When compared to those who graduate from high school, individuals who obtain a GED are less likely to continue with their education at the post-secondary level. In addition, GED recipients do not fare as well in the labor market as those who complete high school in the traditional way (Boesel et al., 1998; Murnane et al., 2000). Third, research has shown that second chance educational programs often miss their target population - students of underprivileged social groups - and instead, they serve middle-class students whose
108
family resources enable them to take advantage of these additional opportunities. For example, Lee and Frank (1990) report that users of community colleges as a pathway to four-year universities in the U.S. are not socially disadvantaged on average. Ayalon (1990) shows that in Israel, members of the more privileged Jewish ethnic group take the best advantage of second chance opportunities. Second Chance Education in Israel There are two major forms of second-chance education in Israel: external matriculation examinations and university preparatory programs (Mechinot). Students who do not succeed in all matriculation examinations while in secondary school, or those who had dropped out, can take them on their own at any time. When taken outside the formal secondary school framework, the examinations are referred to as external matriculation examinations. External examinations are similar to internal ones except that the final grade of the internal examination is computed as an average of the exam score and the school grade in the subject; in the external examination, the final score is computed solely from the exam score. Most students who take external examinations do so within the external school systems. These are usually private schools that specialize in preparing students for the examinations. Their clientele consists of two primary categories: teenagers who failed in academic secondary schools and young adults, usually in their twenties, who did not pass some or any of the matriculation examinations (Inbar & Sever, 1986). The university academic preparatory programs, Mechinot, are one-year intensive study programs sponsored by the universities in collaboration with the Israeli Ministry of Education, which also subsidizes them (Ayalon et al., 1992). The goal of the Mechinot is to offer young men and women a second chance at university admissions, or to improve their prospects for admission to the more selective faculties therein. Earlier empirical research on second chance education in Israel has been rather limited in scope. Inbar and Sever (1986) studied the social composition of students enrolled in external secondary schools and found that their socioeconomic characteristics are very similar to the general distribution of the population. Neither the top nor the bottom strata are over-represented in this group of students. They suggest that the real common quality among these students is that despite having not completed high school, they hold high educational aspirations. However, Inbar and Sever argue that these programs have not been successful in providing external school students entry into post-secondary education. They found that only a small percentage of the
109
students who enroll in external schools actually receive the matriculation diploma they desired. Ayalon (1990) compared the social composition of external day schools to that of academic and vocational students at traditional secondary schools, as well as to dropouts. She too found that the socioeconomic profile of external students is similar to the composition of average Israeli high schools, except that Ashkenazim, the privileged Jewish ethnic group, are over-represented among them. She concludes that external schools are employed as a second chance by academically weak Ashkenazim, a socio-economically successful ethnic group, rather than by those who are at the very bottom of the social hierarchy. Thus, despite their intended goals of moderating educational inequalities, students of advantaged groups are those who benefit most from these second chance mechanisms. Therefore, this second chance structure actually widens social gaps in education by providing additional opportunities to students from privileged strata. In a later study, Ayalon et al. (1992) examined the social composition and success rates of students enrolled in Mechinot in the early 1980s. In this case, the authors found that educational inequalities between social strata that were observed at the beginning of the program diminished by graduation. Yet, this study did not explore what leads to enrollment in Mechinot or their impact, relative to other educational destinations, nor was the focus on tertiary educational outcomes. Compared to previous studies of second chance opportunities in the Israeli educational system, ours has two main advantages and one major limitation. The first advantage is that we have at our disposal nationally representative data based on birth cohorts. Previous studies were in many cases not representative of the entire population, either because they excluded early dropouts or because they studied students at specific second-chance programs. The second advantage is that our data provide us with the opportunity to address the two main questions on second chance education: a second chance for whom and a second chance to where? Thus far, the Israeli research has primarily focused on the first question. The limitation is that we do not have information on the institutional structures within which respondents may have attempted to obtain a second chance. We do not know if they enrolled in Mechinot or in external schools. However, our focus is not on the specific type of second chance taken, but rather on the comparison of second chance students to other "traditional" students. The study has three specific components. First, we study the social determinants of the educational route taken by students and distinguish between three main routes: (1) early dropout of secondary education; (2) main road: obtaining the matriculation diploma in time (by age 19) and going on to tertiary
110
education, and (3) second chance. Second, we compare the degree of gender, ethnic, and socioeconomic inequality in the log odds of matriculation, before and after the use of second chance education. Third, we explore the postsecondary educational and occupational attainment associated with having taken a specific educational route. Specifically, we look at the attainment of second chancers in comparison to individuals who took other educational routes.
DATA AND VARIABLES

Our data are drawn from the national population censuses held by the Israeli Central Bureau of Statistics (CBS) in 1983 and in 1995. The Census consists of a short questionnaire, which is administered to all persons living in Israel during the week of the census, and a more detailed one, which is completed by a 20% sample of households and their members, who are fifteen years old and over. The census master files, which are kept under tight security at CBS, include respondents' national ID numbers. This allows CBS to merge individual records across censuses. Subject to various confidentiality constraints, the research community can request sub-files from this rich database. In this study we employ merged records for persons who were 19-20 years old in 1983. We refer to them as our primary subjects. The 19-20 age group was selected because it was old enough in 1983 to have graduated from secondary school, completed the matriculation examinations, and to have been informed by the Ministry of Education of the results. For primary subjects who still lived with their parents at the time of the census, data was merged with parents' 1983 census records.3 In the 1983 census there were 16,114 detailed questionnaires completed for 19-20 year old individuals. For 3,212 of these individuals, detailed questionnaires were completed in the 1995 census. This is exactly the proportion that we would expect on the basis of equal probability (20%). Of those, we identified parental records for 90%. The first dependent variable in our analysis is labeled Route. This variable represents the educational route that primary subjects followed. It is measured by cross-tabulating the highest diploma attained by 1983 and by 1995, and it consists of five categories. The first category includes subjects who did not obtain the matriculation diploma in 1983 or in 1995. We label this category "leavers" in that they have opted out of any post-secondary education. Indeed, many of them did not complete secondary school. This category comprises 35.3% of the sample. The second category consists of
111
all those who obtained a matriculation diploma on time (in 1983 when they were 19-20 years old) and then obtained post-secondary educational qualifications by 1995. A quarter of the sample took this route, which we label the "main road." The third category consists of those who did not obtain a matriculation diploma in 1983, but who by 1995 obtained the diploma or even completed a post-secondary degree. We assume that most students who took this route did so via the external schools or the Mechinot and we label this category "second chance." The second-chance route is taken by about thirteen percent of the cohort. A fourth category is labeled "halters." These are people who obtained the matriculation diploma in 1983 but did not complete any form of post-secondary education. About 11% of the sample fall in this category. Finally, the last category "Other," which consists of 17.8% of the cohort, includes subjects with missing data on route, as well as 6.3% who were coded as having a matriculation diploma in 1983 but not in 1995.4 Other dependent variables will be described later as we bring them into the analysis. Six variables were included as predictors for this analysis. Track in secondary school included three mutually exclusive dummy variables: academic track, vocational track, and no secondary education. Subjects were asked to indicate how many years they attended each form of education. Those who attended any vocational secondary education were coded as vocational track students; those who attended at least 3 years of academic secondary education were coded as academic track students; those who attended two or fewer years of academic education, and no vocational education, were considered having no secondary education. Parental education is computed as the mean number of school years attended by mother and father. Standard of living is an indicator of the economic circumstance of subjects' family of origin. The census form includes a list of yes/no questions about household possessions, including such items as an air conditioner, car, television, phone, etc. We computed an index summing positive responses to the possession questions, and for ease of interpretation standardized it to a mean of zero and a standard deviation of one. Ethnicity consists of three dummy variables: "Ashkenazim" (Jews whose fathers were born in Europe, America or South Africa as well as second generation Israelis5); "Mizrahim" (Jews whose fathers were born in the Middle East or in North Africa.); and "Arabs" (Non-Jews whose fathers were born in Israel). Very few respondents did not fit in any of the three categories and were excluded from the analysis. Gender is coded 1 for men and 0 for women. The Census form asks women to indicate the number of births they have had. We employ mother's response to this question as a proxy for the variable number of siblings.
Table 1.
Means (SD) Gender Track in Secondary School (row percentages)
Proportions and Means of Variables by Educational Route, Men and Women 19-20 Years Old in 1983.
Proportions
Ethnicity (row' percentages)
Socioeconomic Variables
Route 0.14 0.53 0.29 0.01 0.01 0.00 0.08 0.55 0.12 0.33 0.48 0.60 0.35 0.57 0.25 0.43 0.67 0.44 0.56 0.56 0.64 0.62 0.37 0.55 0.22 0.34 0.25 0.29 0.11 0.60
Total
Mizrahim
Arabs
Ashkenazim
Males
Academic Track
Vocational Track
No Sec Ed
Parental Education
No. of Standard of Siblings Living
Leavers
0.35
0.53
0.33
Main Road
0.25
0.36
0.09
0.13
0.54
0.24
Second Chance Halters
0.11
0.50
0.16
Other
0.18
0.54
0.21
Total
N=2909
0.49
0.22
8.3 (6.1) 16.5 (6.2) 10.5 (6.4) 13.5 (5.9) 10.7 (7.0) 11.5 (7.1)
6.3 (2.5) 3.8 (2.0) 5.4 (2.4) 4.2 (2.0) 5.4 (2.5) 5.2 (2.5
-0.36 (0.91) 0.65 (0.88) -0.08 (0.88) 0.33 (0.92) -0.04 (0.95) 0.00 (0.99)
113
RESULTS
Educational Routes Table 1 includes descriptive sample information on the independent variables by educational route. Arabs and Mizrahim are over-represented among leavers, as are men and students of vocational tracks. The mean parental education and standard of living are lowest for this category, and the number of siblings is highest. Thisfindingis consistent with numerous studies of the socioeconomic background and social origin of high school dropouts and those who do not advance to post-secondary education (Shavit, 1990; Ayalon, 1990). Main roaders are disproportionately Ashkenazi (55%) and students of the academic track (62%), and their social origins are more favorable on average than those of the other categories. For example, the standard of living is much higher for this group than for any other group, and their mean parental education (16.5 years) is nearly twice that of the leavers. Turning to the second chance category, we see that in comparison to main roaders, it consists of more men, Arabs and Mizrahim, and its mean parental education and standard of living are lower. Finally, second chance students are disproportionately vocational track students. We now turn to a multinomial logit analysis of the determinants of educational route. Since virtually no dropout has ever obtained a matriculation diploma, when estimating the logit models for Route, we ran into singularity problems when estimating our models. Therefore, we dropped all subjects (n = 349) who did not report a minimum of 2 years of secondary education from the multinomial logit analysis. (They were subsequently returned to the analysis presented in later tables.) In Table 2 we present the logit coefficient estimates on two log odds ratios: taking a second chance rather than being a leaver (column 1), and second chance rather than taking the main road (column 2). The results in column 1 show that the ethnicity variables are not statistically significant in affecting the odds that a student who failed to obtain the matriculation diploma by 1983 will take a second chance and do so by 1995. Arabs are on average more likely than Ashkenazi Jews to take a second chance, but this difference is not statistically significant. Vocational track students are less than half as likely as academic students to take a second chance (e _ 0 " = 0.37). In addition, parental education increases the odds of taking a second chance. The effects of the other social origin variables are not statistically significant in this column. Turning to column 2, we contrast second chancers against those who took the main road. Here, the effect of vocational track is again important, vocational track students are more than twice as likely to take a second chance route rather
114
than the "main road" to post-secondary education (e 087 = 2.39). This is an expected result because vocational tracks do not usually enable students to obtain a full matriculation diploma at the end of secondary education, and they must resort to second chance opportunities. Gender plays an important role in this column: men are more likely to have obtained the matriculation diploma via a second chance route. We also see strong socioeconomic effects. First, the effect of parental education is negative and much larger than its positive effect in the previous equation. Second, family size has a positive effect, indicating that students from larger families are more likely to take second chance routes as opposed to the main route. Third, standard of living has a large negative effect (e-0-46 = 0.63) on the log odds of arriving at post-secondary education via a second chance route rather than the main route. Clearly, students whose parents are more educated and who are from wealthier families are more likely to succeed in the first attempt at obtaining the diploma and are less likely to require a second chance. In terms of ethnicity, while we do see that Mizrahim and Arabs are more likely than Ashkenazim to take a second chance route rather than the main route to post-secondary education, these results are not statistically significant. In sum, vocational track and parents' education have opposite effects in the two equations. Vocational education and lower parental education reduce the odds of taking a second chance versus leaving, but they increase the odds when second chancers are compared to those on the main route. However, socioeconomic background is more effective in discriminating between second Table 2. Multinomial Logit Estimates of Educational Route.
Second Chance vs. Leavers Second Chance vs. Main Route
Independent Variables
0)
Ethnicity (reference: Ashkenazim) Sex (reference: female) Parents' Education Number of Siblings Standard of Living Secondary Track (reference: Academic) Vocational Intercept * Significant at /?<0.05. Mizrahim Arabs 0.03 0.51 0.17 0.03* 0.06 0.00 0.99* 0.17
(2)
0.36 0.31 0.36* -0.07* 0.10* -0.46* 0.87* -0.81*
115
chancers and main routers than between second chancers and leavers. In other words, although second chancers originate from better-educated families than the leavers, they more closely resemble leavers, in terms of social origin, than the users of the main road. The Overall Effect of Second Chance on Inequalities The analysis presented so far pertains to the question: who is more likely to take the back door versus the front door to the matriculation diploma and higher education? We now turn to our principal question: do second chance structures for the attainment of the matriculation diploma enhance or attenuate social inequalities? In Table 3, we present the determinants of attaining a matriculation diploma by 1983 (column 1) and by 1995 (column 2). The comparison of the two columns should show whether the existence of second chance mechanisms to obtaining the diploma have widened or narrowed the stratification of educational attainment by social origin. Those who obtained the diploma by 1983 did so largely via the front door while those who obtained it between 1983 and 1995 did so through second chance structures. If second chance equalizes the odds of obtaining the matriculation diploma across social groups, we hypothesize that the effects of the various independent variables in column 2 would be smaller than in column 1. Conversely, larger effects in column 2 would indicate that second chance structures enhance inequalities in the odds of matriculation. Comparing the estimates in the two columns shows that the negative impact of being Mizrahi versus Ashkenazi on the odds of obtaining the diploma is reduced between 1983 and 1995. The comparison of Arabs and Jews shows that although the former are less likely to obtain the diploma, when controlling for social origin, the effect of being Arab is positive rather than negative. This result has been shown by previous research (e.g. Shavit, 1990). The net advantage of Arabs is larger in 1995 than in 1983 reflecting thefindingshown in Table 2. Mainly that, net of the various control variables, Arabs are more likely than Jews to take second chances at obtaining the diploma than to be leavers. In 1995, the male disadvantage relative to females is half of what it is in 1983. This suggests that although males are at a disadvantage with respect to the odds of obtaining the diploma through traditional routes, they make-up for it by utilizing second chance routes to its attainment. By contrast to the reductions in the effects of ethnicity and gender, the effect of parental education does not decline and the effects of standard of living and number of siblings decline only slightly.
116
YOSSI SHAVIT, HANNA AYALON AND MICHAL KURLAENDER Table 3. Binary Logit Estimates of Matriculation Diploma.
Diploma Obtained By 1983 (1) Diploma Obtained By 1995 (2) -0.14 0.24 -0.30* 0.05* -0.13* 0.41* 0.24*
Ethnicity (reference: Ashkenazim) Sex (reference: female) Parents' Education Number of Siblings Standard of Living Intercept * Significant at p<0.05.
Mizrahim Arabs
-0.230.01 -0.51* 0.05* -0.15* 0.46* 0.12
-Significant at p<0.10.
We conclude that the second chance structures substantially reduce ethnic and gender inequalities in the log odds of obtaining a matriculation diploma. Second chance structures slightly reduce the effects of the economic circumstance of childhood and family size, and do not reduce the effects of parental education on the odds of obtaining a matriculation diploma. A Second Chance to Where? In this section we explore whether the use of second chance structures affects later educational and occupational attainment and inequality therein. Breen and Johnson (2000) studied alternative educational routes in Sweden and have suggested that the non-conventional routes to tertiary education leads to lower ultimate achievements than the main route.6 In Table 4 we present the determinants of post-secondary educational achievements by 1995. The dependent variable - post-secondary status - in this analysis has three categories: "university degree," including those who have obtained a university degree by 1995; "non-academic post-secondary qualification," including those who have obtained a non-academic post secondary degree; and "no post-secondary qualification," including those who did not obtain any post secondary qualification by 1995. In Table 4 we present multinomial logit estimates for the log odds of three contrasts: university degree versus no post-secondary qualification, nonacademic qualification versus no post-Secondary qualification, and university degree versus non-academic qualification. The analysis is conditional on having obtained a matriculation diploma, whether via the main road or through second
Schooling Alternatives, Inequality and Mobility in Israel Table 4. Multinomial Logit Estimates of Post-Secondary Education Conditional on Matriculation Diploma.
University vs. no Post Secondary Education Non Academic Post Secondary vs. no Post Secondary Education (1) Ethnicity (reference: Ashkenazim) Mizrahim Arabs Sex (reference: female) Parents' Education Number of Siblings Standard of Living Secondary Track (reference: academic) Vocational Route (reference: main road and halters) Second Chance Intercept Significant at p<0.05. -0.26 0.12 0.18 0.14* 0.02 0.29* -0.36* -1.01* -1.20 -0.21 -0.53 0.12 0.06* 0.09* -0.13* -0.03 -0.16 -0.85*
117
University vs. Non Academic Post Secondary Education (2)
-0.04 0.65 0.05 0.07* -0.07 0.42* -0.32 -0.85* -0.15
chance opportunities. The equation for each contrast estimates the effects of social origin, high school track and whether one took the second chance route to post-secondary educational attainment. The results presented in column 1 of the table show that the odds of second chancers obtaining a university degree are lower than those on the main road - regardless of social origin. The findings are different when we investigate the determinants of the log odds to obtain a non-academic qualification rather than no post-secondary qualification; second chance has no statistically significant effect on the log odds of this contrast. Evidently, non-academic programs are not very selective, and therefore, the effects in this column are generally small and only two estimates are statistically significant. The contrast of the two levels of post-secondary education (column 3) shows that the second chancers have higher odds than do their main road counterparts for obtaining the less prestigious certificate - the non-academic degree. In sum, whether one had obtained a matriculation diploma via a second chance route or the main road does not affect the odds of obtaining non-academic qualifications, but the route to obtaining the matriculation diploma is statistically significant in shaping the odds of attaining a university degree. In this case, the odds are considerably
118
lower for the second chancers. In addition to the effect of route on tertiary education we find the expected effects of social origin and of track in secondary school. Graduates of academic secondary education are more likely to obtain a university diploma, as are those from wealthier and more educated families. Individuals who received a non-academic post secondary qualification often came from large families. We now turn to analyze the effects of educational route on occupational attainment. Two competing hypotheses regarding the relationship between these two variables come to mind. The first is that in comparison to main roaders, second chance students may be less academically able (otherwise they could have completed the matriculation examinations on time, and would not have needed to resort to a second chance). Thus, second chancers would be less likely to have done very well in the university or to have attained a very prestigious occupation. On the other hand, second chance students may be more highly motivated, at least as adults, than those who simply followed the traditional educational track, and this quality may enhance their occupational attainment. In Table 5 we present two OLS regressions with the dependent variable socioeconomic index (SEI) - as a measure of subjects' 1995 occupation status.7 The mean of SEI in the sample is 44.58 with a standard deviation of 24.19. Route is the major independent variable in the analysis with second chance as the omitted category. We estimate two regression equations. The first includes Route with the following controls: ethnicity, gender, parental education, number of siblings, standard of living, and track while at secondary school. The results from this analysis indicate that those who followed the main road attained a much more prestigious occupation than second chancers, who, in turn, attain a much more prestigious occupation than leavers. Clearly, these differences are largely due to the ultimate qualification attained. Therefore, in the second equation we control for tertiary qualification and find the following results: As one would expect, holders of university degrees have very high SEI scores, followed by those with non-academic post-secondary qualifications. Second, net of tertiary qualification, the effects of Route are substantially reduced relative to the first equation, but maintain a similar pattern: second chancers attain occupations whose SEI scores fall in the middle between those attained by main roaders and leavers. This result is consistent with the first of the two hypotheses presented above. It would seem that those who require a second chance are usually not the strongest students and are therefore not very likely to enter the most selective university departments (e.g., the professional schools) which lead to the top of the occupational hierarchy. When compared to leavers on the other hand, second chancers are probably more able and motivated and attain better occupations.8
Schooling Alternatives,
Inequality
and Mobility in Israel
119
Table 5. OLS Regression Effects of Educational Route on Occupational SEI Scores.

Independent Variables Ethnicity (reference: Ashkenazim) Mizrahim Arabs 0.07 -1.00 0.38 0.39* -0.76* 0.69 0.40 -2.76* -0.63 0.23 -0.61* -0.57 -2.41 5.63* 6.80* -5.01* -0.55 27.07* 11.32* 40.44 1473 0.41 (1) (2)
Sex (reference: female) Parents' Education Number of Siblings Standard of Living Secondary Track (reference: academic) Vocational -4.28* Route (reference: second chance) 18.28* Main Road -1.24 Halters -11.54* Leavers -1.05 Other Post-Secondary Qualification (reference: none) University Degree Non-Academic Post Secondary Qualification 47.59 Intercept 1485 N of cases 0.34 R2 * Significant at p<0.05.
CONCLUSION
Do Second Chance Opportunities Reinforce Disadvantage or Offer Access to Opportunity? In this paper, we investigate the role of second-chance structures in the process of educational stratification. Specifically, we study the social characteristics of those students who have made use of second chance educational opportunities, relative to those who did not. We examine the extent to which the availability of second chance education alters the degree of ethnic, gender and socioeconomic inequalities in the odds of obtaining a matriculation diploma. Moreover, we compare the ultimate educational and occupational attainments of those who took the main road to those who took the route of second chance. To the best of our knowledge, this is the first study to address all three aspects of the problem at hand.
120
Previous studies on second chance education have been rather skeptical regarding the role of second chance structures in altering the outcomes of the educational stratification process. It has been argued that students from privileged social origins that had failed in the first attempt to reach certain credentials most often use these mechanisms. Therefore, second chance structures do not reduce inequality of educational opportunity between social strata. Previous Israeli studies of external schools tend to agree with this conclusion. Our results are somewhat more optimistic than are those from previous studies. First, when compared to those who took the main road to the matriculation diploma and higher education, the social origins of second chance students are clearly disadvantaged. On the other hand, among those who did not obtain the diploma via the main road, the odds of exploiting second chance opportunities are clearly related to socioeconomic origins. In other words, it is not the case that the socioeconomic and ethnic profile of second chancers is similar to that of the general population or is middle-class. The users of second chance structures tend to come from the lower middle ranges of the social hierarchy, are predominantly men rather than women, and vocational rather than academic track students. Second, the overall inequality in the odds of obtaining the matriculation diploma between men and women, and between ethnic groups, is reduced by the utilization of second chance opportunities. Inequality between economic strata is also reduced slightly, albeit not significantly. To the best of our knowledge, this is the first reporting of such a result. Third, second chancers are not likely to reach the top of either the educational or the occupational pyramids. Rather, when compared to those who had obtained the matriculation diploma via the main road, second chancers are more likely to obtain non-academic post-secondary qualifications rather than the baccalaureate degree, and reach less prestigious occupations. And yet, their occupational attainment is clearly superior to those who had not obtained a matriculation diploma at all. This finding would be interesting to explore at other types of second chance systems. For example, at community colleges in the U.S., where many enrolled students may not reach their ultimate educational goals, does additional schooling contribute to a social positioning that is superior to those with no post-secondary education, and inferior to those who complete the baccalaureate degree? Whether implicit or explicit, if the goal in expansion of the educational system through second chance structures is one of equalizing opportunities, then the success of such programs should prove to narrow the gaps in attainment that exists between different groups in the social hierarchy. Our results indicate that second chance structures do in fact contribute to the equalization
121
of opportunities. However, these changes are minimal in terms of shifts in the overall system of social stratification. The existence of such programs clearly widens access, however the desire for upward mobility via post-secondary educational or professional attainment is not realized. Our findings show that second chance structures help to maintain the stability of the stratification system. Individuals who take advantage of second chance opportunities originate from the lower-middle strata, and second chance structures assist them in keeping this position. The advantages of using the main road compared to second chance mechanisms are obvious. However, given the fact that second chancers did not use the main road, second chance opportunities, which do not seem to provide their users upward mobility, may protect them from downward mobility. Thus, these structures do contribute to the enhancement of social equality at least by preventing the enlargement of educational and occupational gaps. This leads us to the conclusion that the critical attitude toward second chance structures, prevailing in stratification research, should be moderated.
NOTES
1. And yet, some scholars (e.g. Shavit & Kraus, 1990) point out that the role of alternative education in the stratification process should not be evaluated by comparing it to mainstream education, but rather by comparing it to a hypothetical system from which it is absent. In the absence of alternative education, lower-class students would drop out in larger numbers. 2. For an in-depth review of the theoretical underpinnings of the concept of second chance education, see Inbar (1995). 3. Subjects are identified by a household identification number and by their relation to the head of the household or the 'first' person in the household. The values of this variable are spouse, son/daughter, parents, etc. Thus, we could merge the records of the sons/daughters with those of the head of the household and of the head's spouse. The merge with parental records was not possible for subjects who were not polled in the parental household (10% of the sample). Leaving the parental home at the age of 19-20 is due mainly to early marriage, which is relatively common among females of lower socioeconomic strata. Since young married women of lower socioeconomic origin may be less interested, compared to other groups, in acquiring post-secondary education, dropouts may be underrepresented in the final sample. 4. Clearly, this category represents either reporting or coding error. The vast majority of subjects in this category attended vocational secondary education where obtaining a matriculation diploma is not very common. Some people are not fully aware of the meaning of 'matriculation diploma' "Bagrut" and fail to distinguish it from a graduation diploma that is awarded to all secondary school graduates, but is of little value in the labor market or for further education. Therefore, it is likely that the errors were in the 1983 "Bagrut" response rather than in the 1995 response. By 1995, the difference may have become more apparent to respondents and the responses may have been more accurate.
122
5. Second generation Israelis (father born in Israel) are less than 3% of the sample. In 1983, 72% of the Israeli born aged 40-64 (the age group of the parents of the subjects in our study) whose fathers were born abroad were of Ashkenazi origin (ICBS, 1994). Our data show that Israeli-born parents are very similar to the Ashkenazi parents in years of schooling, standard of living, and number of children. 6. They also found a slightly stronger effect of social origin on subsequent educational attainment for those who followed non-conventional educational routes. We did not find consistent interactions between route and social origins in our data. 7. The Socioeconomic Index of Occupations originally developed by Duncan (1961) is a composite measure of the median education and income of persons employed in each occupation. The index correlates highly with the prestige of an occupation as measured in large samples. It was adapted to the Israeli case by Tyree (1981) and later by Semyonov, Lewin-Epstein and Mandel (2000). 8. Our results are consistent with research on the labor market performance of community college entrants versus those who do not attend college at all and those who attend four-year institutions (Kane & Rouse, 1999). It is also consistent with work on GED recipients who fare better in the labor market than do their high school dropout equivalents, and worse than their high school graduate counterparts (Murnane et al., 1997, 2000).
ACKNOWLEDGMENTS
This research was supported by the Israel Science Foundation grant no. 891/99 and by a grant from the Ministry of Education. The Central Bureau of Statistics under its monographs program provided the data. We thank Moshe Ben Elyahu, Ahmed Hleihel, and Yaffa Shiff of the Israeli Central Bureau of Statistics for their assistance in preparing the data file, and Yasmin Alkalay for technical support.
REFERENCES
Arum, R., & Shavit, Y. (1995). Secondary Vocational Education and the Transition from School to Work. Sociology of Education, 68(3), 187-204. Astone, N. M., Schoen, R., Ensminger, M., & Rothert, K. (2000). School Reentry in Early Adulthood: The Case of Inner-City African Americans. Sociology of Education, 73(3), 133-154. Ayalon, H., Shapira, R., & Shavit, R. (1992). A Second Chance for Higher Education: Academic Preparatory Programs in Israel. Research in Higher Education, 33(4), 497-511. Ayalon, H. (1990). The Social Impact of Non-regular Education in Israel. Comparative Education Review, 34(3), 302-313. Boesel, D., Alsalam, N., & Smith, T. M. (1998). Educational and Labor Market Performance of GED Recipients. National Library of Education, Office of Educational Research and Improvement, U.S. Department of Education: Government Printing Office.
Schooling Alternatives,
Inequality
and Mobility in Israel
123
Breen, R., & Jonsson, J. O. (2000). A Multinomial Transition Model for Analyzing Educational Careers. American Sociological Review, 65(5), 754-772. Brint, S., & Karabel, J. (1989). The Diverted Dream: Community Colleges and the Promise of Educational Opportunity in America, 1900-1985. New York: Oxford University Press. Dougherty, K. (1987). The Effects of Community Colleges: Aid or Hindrance to Socioeconomic Attainment? Sociology of Education, 60(2), 86-103. Dougherty, K. (1994). The Contradictory College: The Origins, Impacts and Futures of the Community College. New York: SUNY Press. Dougherty, K. (1972). The Community College at the Crossroads: The Need for Structural Reform. Harvard Educational Review, 61(3), 311-336. Duncan, O. D. (1961). A Socioeconomic Index for all Occupations. In: A. J. Reiss (Ed.), Occupations and Social Status (pp. 109-138). New York: The Free Press. Gamoran, A., & Mare, R. (1989). Secondary School Tracking and Educational Inequality: Compensation, Reinforcement or Neutrality. American Journal of Sociology, 94(5), 1146-1183. Grubb, W. N. (1991). The Decline of Community College Transfer Rates: Evidence from National Longitudinal Surveys. Journal of Higher Education, 62(2), 194-222. Hearn, J. (1984). The Relative Roles of Academic, Ascribed, and Socioeconomic Characteristics in College Destination. Sociology of Education, 57(1), 22-30. Hordley, I., & Lee, D. J. (1970). The 'Alternative Route' - Social Change and Opportunity in Technical Education. Sociology, 4, 23-49. Inbar, D. E. (1995). Second Chance in Education: Principles and Rituals. The Journal of General Education, 44(1), 26-42. Inbar, D. E., & Sever, R. (1989). The Importance of Making Promises: An Analysis of SecondChance Policies. Comparative Education Review, 33(2), 232-242. Inbar, D. E., & Sever, R. (1986). The Transformation of a Goal Definition: The Story of a Second Chance Evaluation Study. Studies in Educational Evaluation, 12, 177-189. Inbar, D. E., & Sever, R. (1986). Second Chance: Some Theoretical and Empirical Remarks. Issues in Education, 4(2), 121-135. Israel Central Bureau of Statistics (ICBS). (1994). Statistical Abstract of Israel 1994 (No. 45). Kane, T. J., & Rouse, C. E. (1999). The Community College: Educating Students at the Margin Between College and Work. Journal of Economic Perspectives, 13(1), 63-84. Karabel, J. (1972). Community Colleges and Social Stratification. Harvard Educational Review, 42(A), 521-562. Kerckhoff, A. C. (1974). Stratification Processes and Outcomes in England and the U.S. American Sociological Review, 39(6), 789-801. Kerckhoff, A. C , Campbell, R., & Winfield-Laird, I. (1985). Social Mobility in Great Britain and the United States. American Journal of Sociology, 91(2), 281-308. Lee, V., & Frank, K. (1990). Students' Characteristics that facilitate the Transfer from Two-year to Four-year Colleges. Sociology of Education, 63(3), 178-193. Mare, R. (1979). Social Background Composition and Educational Growth. Demography, 16(1), 55-71. Monk-Turner, E. (1998). Community College Education and its Impact on Socioeconomic Status Attainment. Mellen Studies in Education, Vol. 41. The Edwin Mellen Press. Monk-Turner, E. (1994). Economic Returns to Community and Four-Year College Education. Journal of Socio-Economics, 23. Munns, G., & McFadden, M. G. (2000). First Chance, Second Chance or Last Chance? Resistance and Response to Education. British Journal of Sociology of Education, 27(1), 59-75.
124
Murnane, R. J., Willett, J. B., & Tyler, J. H. (2000). Who Benefits from Obtaining a GED? Evidence from High School and Beyond. The Review of Economics and Statistics, 82(1), 23-37. Murnane, R. J., Willett, J. B., & Boudett, K. P. (1997). Does a GED lead to more Training, PostSecondary Education, and Military Service for School Dropouts? Industrial and Labor Relations Review, 51(1), 100-116. McFadden, M. G. (1996). 'Second Chance' Education: Accessing Opportunity or Recycling Disadvantage? International Studies in Sociology of Education, 6(1), 87-111. Oakes, J. (1985). Keeping Track: How Schools Structure Inequality. New Haven, CT: Yale University Press. Pincus, F. L. (1980). The False Promises of Community Colleges: Class, Conflict, and Vocational Education. Harvard Educational Review, 50(3), 332-361. Raffe, D. (1979). The Alternative Route Reconsidered: Part-Time Further Education in England and Wales. Sociology, 13, 47-73. Rouse, C. E. (1995). Democratization or Diversion? The Effect of Community Colleges on Educational Attainment. Journal of Business Economics and Statistics,!3(2), 217-224. Semyonov, M., Lewin-Epstein, N & Mandel, H. (2000). Update Socioeconomic Scale for Occupations in Israel. Megamot, 40(A), 701-729. Saha, L. (1985). Orientations, Vocational Training Plans, and Educational Attainment Among Urban Australian Youth. Sociology of Education, 58(A), 228-240. Shavit, Y. (1984). Curricular Tracking and Ethnicity. American Sociological Review, 49(2), 210-220. Shavit, Y. (1990). Segregation, Tracking and the Educational Attainment of Minorities: Arabs and Oriental Jews in Israel. American Sociological Review, 55(1), 115-126. Shavit, Y., & Vered Kraus. (1990). Educational Transitions in Israel: A Test of the Industrialization and Credcntialization Hypotheses. Sociology of Education, 63(1), 133-141. Tyree, A. (1981). Occupational Socioeconomic Status, Ethnicity and Sex in Israel: Considerations in Scale construction. Megamot, 22(1), 7-21. Whitaker, D., & Pascarella, E. (1994). Two-Year College Attendance and Socioeconomic Attainment: Some Additional Evidence. Journal of Higher Education, 65(2), 194-210. Yogev, A. (1997). Second Chance Education and Alternative Routes. In: L. J. Saha (Ed.), International Encyclopedia of the Sociology of Education, (pp. 469-472) New York: Pergamon.
Commentary: IS SOCIAL CAPITAL THE SELF-ESTEEM OF THE 1990s?

Commentary on Kao and Shavit et al. papers.
Maryellen Schaub and David P. Baker
INTRODUCTION
Social capital and its less studied cousin, cultural capital, are shaping up to be major theoretical concepts not only in sociology but in other social sciences as well. This makes the present volume timely and prompts the question, where is this all going? In stepping back from the more immediate empirical work and taking stock of possible roles of these two concepts for research in the sociology of education, one is struck by crucial questions. Is social capital being defined consistently? Can the dizzying array of ideas, definitions, and measures associated with social capital ever be trimmed into a coherent set of conceptually powerful and testable propositions? Can these ideas help sociology of education to better understand not only social reproduction and school effects but also macro-questions like historical change in schooling and its function in modern society? Why are there more empirical analyses claiming to use the more ambiguous notion of social capital than the less ambiguous notion of cultural capital? Full answers to these and similar questions are beyond the scope of this essay, but we pose them before placing the chapters by Kao and Shavit et al. in this wider context.
125
126
MARYELLEN SCHAUB AND DAVID P. BAKER
We chose a provocative question to title our critique because, in our judgment, social capital seems at risk of suffering an intellectual and policy fate similar to the once-reigning construct of "self-esteem." It became the concept celebre of some social sciences in the late 1970s and 1980s. Starting as a promising but under-defined notion of how an individual thinks about herself, the term quickly moved to include the interaction between the individual's psychological development and the social world. In so doing, the term took on greater and more varied dimensions until it included so many weighty notions of the modern self (e.g. self concept, self image, self efficacy, self identity, self respect), it became almost useless as a theoretical concept. At the same time that self-esteem was having its theoretical and definitional problems, the general public became enchanted with the idea and it took off. In the policy world, the notion of increasing self-esteem to solve all sorts of social problems took on a cure-all quality, and some states in the U.S. actually debated legislation declaring it as a human right to be fostered by the state. Its doomed fate as a concept was not because the idea was unimportant, indeed quite the opposite. It became loaded with too many central ideas about the self and society, and the general fascination with the idea fueled it to take on even greater dimensions, leading eventually to its ultimate collapse as a theoretical concept.
SOCIAL CAPITAL: A THEORY IN WAITING?

In reviewing the burgeoning literature on social capital, we are struck by the confused state of theory development. In reviewing recent critiques of social capital from a variety of disciplines, most agree that something theoretically important is attempting to be born, but some are wondering if it ever will (see Foley & Edwards, 1999; Bowles, 1999; Durlauf, 1999; Portes, 1998). The problem of definition, and hence theoretical usefulness, is the most pressing concern for the future of social capital in the sociology of education.1 Problems began with the earliest development of the idea. One can choose from two and one half stories about the origins of social capital as a concept. As Portes describes, the two main founding stories are about Pierre Bourdieu and James Coleman, who separately developed the same-named concept with similar ideas, but with some important and lasting differences that have led to very different operationalizations of social capital (Bourdieu, 1980, 1985; Coleman, 1988, 1990). Both of these founders of social capital launched their idea with very weighty promises and accompanying broad definitions. It is noteworthy
Commentary: Social Capital, Self-Esteem?
127
for this volume to mention that both of these scholars initially developed their notions of social capital in the context of theorizing about education, including the character of schools and accounts of student achievement. For Bourdieu, social capital becomes the third in a trilogy of capitals (after cultural and economic capitals) that 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" (1985, pp. 248-249). Further, when these three capitals are considered together, Bourdieu believes they explain no less than "the structure and dynamics of differentiated society" (Bourdieu & Wacquant, 1992, p. 119). For Coleman, social capital is "a variety of entities with two elements in common: They all consist of some aspect of social structures, and they facilitate certain action of actors whether persons or corporate actors within the structure" (Coleman, 1988, S98). To match the theoretically liberating development in economics of the concept of human capital, and even to ultimately explain human capital production, Coleman saw the idea of social capital as central to an expanded sociology of modern society. Finally, a number of other scholars in sociology and economics were using similar notions at about the same time, such as Loury's (1977, 1981) rejection of more individualistic notions of the origins of racial income inequality and their policy implications and Granovetter's (1974, 1985) analyses of networks and labor market success. The problem of definition begins when these ideas are actually put to use. For example, Coleman's notion of the variety of forms of social capital takes on what Foley and Edward's review calls "the relative incoherence of [a] laundry list" (1999, p. 144). For instance, Coleman's social capital can include: "obligations and expectations," "information potential," "norms and effective sanctions," "authority relations," "appropriable social organization," and "intentional organization," understood as "direct investment in social capital" (Coleman, 1990, pp. 306-313). Further, for Coleman, social capital inheres in the social relations of a network or organization, and as such it is not fungible, while for Bourdieu it is fungible like financial capital. This is a major point that needs to be resolved. Similarly, Coleman has been criticized for a disconnect between his theoretical vision of social capital as embedded in relationships of the collective and his empirical formulations that are attached to individual decision-making (Tilly, 1998). The theory of social capital is expanding faster than it can resolve crucial inconsistencies among its different versions; perhaps new work will remedy this problem (e.g. Lin, 2001). Political scientist Robert Putnam's (1995) popularized image of declining civic involvement in American society transformed the idea of social capital into a public term. Much like self-esteem, as a concept social capital reached
128
the popular media long before it had been fully developed theoretically and empirically. Can use of a premature notion of social capital for social policymaking be far behind (LeTendre et al., 2001)? Furthermore, Putnam's academic work puts yet another twist on the theory (Foley & Edwards, 1999). His form of social capital becomes focused on "civic culture" with an emphasis on trust, norms, and values resulting from association. In some empirical work, this has been operationalized to mean social capital as an individual's general trust in public institutions. Lastly, political scientists are currently debating whether social capital is a bundle of variables, like norms and values, that are exogenous to the political process, or more endogenous to social structure as many sociologists see it (e.g. Jackman & Miller, 1998). As social capital becomes part of sociology of education, we must address these concerns to save it from the fate of self-esteem. One way to do so is to undertake more research on what social capital really is, how we can measure it effectively, how it varies in predictable ways, and what precursors this variation might have. In their review of 45 recent empirical political and policy studies using social capital, Foley and Edwards (1999) find that about one half viewed social capital as the dependent variable being investigated. We hope that would be the case in sociology of education, but we suspect that it is not. This line of investigation needs to be pursued to strengthen the usefulness of social capital in the analysis of education. For example, what do we really mean when we say that one type of school has more social capital among its clients (parents) than another? Given that there probably can be both good and bad social capital (there is still confusion in the conceptual literature over this point), and networks can have high and low amounts of resources, what do different qualities of social capital mean for central questions about schooling? Can current definitions and operationalizations of social capital help us analyze variation in networks of children, teachers, or parents in ways that illuminate important school outcomes? It strikes us that social capital as a concept for sociology of education is lacking appropriate empirical description that is necessary before the concept itself can be applied as an explanatory concept. Deep description leads to clearer measures and clearer hypotheses about how social capital may play a role in schooling and learning. This problem has hampered sociology of education research in the past, as with the imbalance between extensive analysis on curricular tracking as an independent variable and its much more limited work on tracking as a dependent variable (Baker, 1994). We should not repeat the mistake with social capital. Lastly, we ask whether social capital can help us think about the historical development of schooling and its role in society. Sociologists of education have
129
made significant progress in developing sociological accounts of the growth of formal schooling, the selective rise and decline of private schooling, and institutional isomorphism across education organizations. These macro arguments are now starting to help us put our more micro- and meso-level analyses into fuller theoretical perspective. As of yet, it is unclear what role, if any, the notion of social capital will play in this story.
FOCUSING ON POSSIBLE ELEMENTS OF SOCIAL CAPITAL

The chapters by Grace Kao and Shavit et al. do not directly deal with social capital. Instead, they focus on more traditional themes of educational status attainment. On their own, both of these chapters are fine examples of empirical sociology of education that bring data to central questions in a new way, and hence are well worth the read on each specific topic. Additionally, the two papers offer a context to illustrate some of our concerns about whether or not social capital will add to our understanding of educational stratification. So with our apologies to the authors in advance, we hold up these chapters to examine issues about the usefulness of new conceptions of capital in education research well beyond the original intentions of the authors. Grace Kao's chapter offers a much-needed empirical description of differing levels of aspirations for college held by different groups of parents in U.S. By focusing on race/ethnic/immigrant status differences, she reports many interesting things. Chief among these is that within socioeconomic levels, African American, Hispanic, and Asian parents have higher educational aspirations for their high school children than do white parents. Further, immigrant parents have higher educational aspirations than native-born parents. Kao goes on to model these basic trends in several informative ways that shed light on the multivariate relationship among these variables and parents' aspirations for their children. Relevant to our discussion of social capital is Kao's stated motivation for examining these trends. Citing the fact that recent studies of status attainment have expanded to look beyond socioeconomic status - i.e. human and economic capital - and found that "parents provide a multitude of educational resources, which often are not directly associated with socioeconomic status" (Kao in this volume, pp. 87-106) - i.e. social capital (and perhaps cultural capital) - she intends to do the same for parental college aspirations. Her findings raise some provocative questions when viewed through the lens of social capital. Since Kao finds that non-white and immigrant families have higher aspirations than
130
whites and native-born families controlling for socioeconomic status, does a social capital explanation make sense here? Or could it have nothing to do with resources derived from a network, but instead be a combination of lower family financial resources to pass on to children, perceived discrimination of less privileged non-whites, and a perception of schooling as a route to the middleclass that are driving aspirations of these parents up? In a way, Kao's findings turn the notion of social capital on its head. Isn't it difficult to assume that less privileged nonwhite families obtain non-financial resources from other similar parents that would lead them to having higher aspirations than white parents and their social networks? If, at the very least, social capital inheres in the social relations of a network, what do we need to know about the qualities of networks of parents to understand Kao's interestingfindings?The kinds of results that Kao reports are not unique among studies of how ethnic and minority groups use education to navigate the status hierarchy within the American society (e.g. Collins, 1979). But unless social capital theory can account for variation across networks, it is hard to understand how the theory works in each particular case (for example, a network with high social closure but low collective resources is assumed to yield useful resources for network members). Without deeper description and more theoretical development, the image of social capital may not provide reasonable interpretations of the educational experiences of less privileged minorities in developed societies. Shavit, Ayalon, and Kurlaender's chapter provides an interesting examination of the education attainment consequences of an alternative route to higher education matriculation examinations among secondary school students in Israel. Specifically, Shavit and his colleagues are interested in whether or not alternative second-chance examinations and preparation programs decrease educational stratification by socioeconomic and ethnic status. They find three processes. First, second-chance mechanisms are used by students from traditionally disadvantaged groups. Second, use of a second chance slightly reduces economic and ethnic inequalities. Lastly, second chance takers tend not to end up with high educational and occupational attainment as adults. Beyond these enlightening findings, the chapter illustrates the institutional power of formal schooling to generated alternative forms of credentialing and how these might influence educational status attainment. Seen in this light, the chapter illustrates the challenge to social capital theory to incorporate various aspects of new educational innovations and their impact on educational stratification. As formal schooling is increasingly institutionalized and legitimated as the main inter-generational mobility mechanism in modern society, it can produce an array of augmenting structures for actors to "play the educational credentialing game." Sociologists of education have examined everything from alternative
131
private schooling to shadow education (Baker et al., 2001) and alternative examinations (Smith, 2000) as innovations of formal schooling that interact with educational stratification. Shavit et al.'s chapter adds to this growing literature, and in the context of the current volume raises the question, can social capital theory help us to explain the origins of these kinds of innovations and their role in stratification? What would a theory of social capital aimed at educational innovation look like? Most applications of social capital operate with very micro networks, but what about networks of larger collectives such as political organizations and multinational agencies and so forth, which are often at the core of implementing new educational structures? Could a more specified theory of nonfinancial resources stemming from these kinds of networks be empirically useful? It is not clear to us that in its current state of theoretical development social capital can help us understand major questions of how institutional and organizational change influence educational stratification. The image of social capital has stimulated considerable research in sociology of education. Given that the centrality of the components behind the bundle of ideas floating under the banner of social capital - such as collective resources, fungible nonfinancial resources, an expanded notion of capital, civic actions, and so forth - clearly holds great promise for sociological theory. But the risk of runaway popularity of a notion of social capital moving beyond solid theoretical and empirical work threatens its full development into a theory that can enhance sociological analysis of education. The challenge before sociology of education that social capital as a theory presents is clear. We hope that it will be met in the near future.
NOTE
1. Most of our comments are aimed at social capital since the problems we raise are the most pronounced for this kind of capital. To a lesser degree, similar points can be made about the future use of cultural capital.
REFERENCES
Baker, D. (1994). In Comparative Isolation: Why Comparative Research has So Little Influence on American Sociology of Education? In: A. Pallas (Ed.), Research in Sociology of Education and Socialization (Vol. 10, pp. 53-70). Greenwich, CN: JAI Press. Baker, D., Akiba, M., LeTendre, G. K., & Wiseman, A. W. (2001). Worldwide Shadow Education: Outside School Learning, Institutional Quality of Schooling, and Cross-national Mathematics Achievement. Educational, Evaluation and Policy Analysis, 23(1), 1-17. Bourdieu, P. (1980). Le trios etats du capital culturel. Actes Rech. Sci. Soc, 30, 3-6.
132
Bourdicu, P. (1985). The Forms of Capital. In: J. Richardson (Ed.), Handbook for Theory and Research for the Sociology of Education. New York: Greenwood Press. Bourdieu, P., & Wacquant, L. (1992). An Introduction to Reflexive Sociology. Chicago: University of Chicago Press. Bowles, S. (1999). Social Capital and Community Governance. Focus, 20, 1-5. Coleman, J. (1988). Social Capital in the Creation of Human Capital. American Journal of Sociology, 94, S95-121. Coleman, J. (1990). Foundations of Social Theory. Cambridge: Belknap Press of Harvard University Press. Collins, R. (1979). The Credential Society. New York: Academic Press. Durlauf, S. (1999). The Case Against Social Capital. Focus, 20, 1-5. Foley, M., & Edward, B. (1999). Is it Time to Disinvest in Social Capital? Journal of Public Policy, 19, 141-173. Granovetter, M. (1974). Getting a Job: A Study of Contacts and Careers. Cambridge, MA: Harvard University Press. Granovetter, M. (1985). Economic Action, Social Structure, and Embeddcdness. American Journal of Sociology, 91, 481-510. Jackman R., & Miller, R. (1998). Social Capital and Politics Annual Review of Political Science, 1, 47-73. LeTendre, G. K., Akiba, M., Goesling, B Wiseman, A. W., & Baker, D. P. (Spring 2001). The Policy Trap: National Educational Policy and the Third International Math and Science Study. The International Journal of Educational Policy, Research and Practice. Lin, N. (2001). Social Capital: A Theory of Social Structure and Action. Cambridge University Press. Loury, G. C. (1977). A Dynamic Theory of Racial Income Differences. In: P. A. Wallace (Ed.), Women, Minorities and Employment Discrimination. AM La Mond. Lexington MA: Heath. Loury, G. C. (1981). Intergenerational Transfers and the Distribution of Earnings. Econometrica, 49, 843-867. Portes, A. (1998). Social Capital: Its Origins and Applications to Modern Sociology. Annual Review of Sociology, 24, 1-24. Putnam, R. (1995). Bowling Alone: America's Declining Social Capital. Journal of Democracy, 6, 65-78. Smith, T. (2000). Who Values the GED? Factors Underlying the Supply and Demand of High School Equivalencv Credential. Dissertation, Penn State University. Tilly, C. (1998). Durable Inequality. Berkley and Los Angeles:University of California Press.
GETTING AHEAD IN KENYA: SOCIAL CAPITAL, SHADOW EDUCATION, AND ACHIEVEMENT

Claudia Buchmann
ABSTRACT
This study of Kenya addresses two questions: Do social and cultural capital matterfor school achievement, net of family background factors? If so, what aspects matter most? The paper examines shadow education, which consists of tutoring and exam preparation classes, as a form of cultural capital utilized by some families to advance children's likelihood of educational success. With data for 506 households, it determines who is most likely to participate in shadow education and assesses the consequences of shadow education and other measures offamily social and cultural capital for two educational outcomes: grade repetition and academic performance. Results indicate that boys and urban children are more likely to participate in shadow schooling which, in turn, is related to lower incidence of grade repetition and higher academic performance. Other measures of cultural capital, such as language spoken in the home and mother's reading habits, reduce the likelihood of grade repetition.
133
134
CLAUDIA BUCHMANN
INTRODUCTION
In developing countries, as in industrialized ones, most families hope to see their children succeed in school but some families have more resources to improve their children's likelihood of academic success than others. Many of the ways in which families influence children's educational trajectories are well established. Educated parents often make more informed educational decisions for their children and can assist them in other ways that boost their academic progress. Wealthy parents are usually better able to ensure that their children receive high quality schooling than poor parents. It is also increasingly clear that resources beyond family financial and human capital matter for children's educational outcomes. Parents' cultural capital, conceptualized either as highstatus cultural knowledge and preferences (Bourdieu, 1977; DiMaggio, 1982; Wong, 1998) or cognitive and linguistic skills (Farkas, 1996; De Graaf, De Graaf & Kraaykamp, 2000), has been found to impact children's educational outcomes. Likewise, social capital, embedded in the relationships between children and their parents as well as those between the family and community, has come to be viewed as an important intervening mechanism in the relationship between family background and educational attainment and achievement (Coleman, 1988; Coleman & Hoffer, 1987). Much of the research on social and cultural capital as they relate to educational attainment and achievement has focused on Western, industrialized societies. Far fewer studies have investigated the impact of such resources on children's school performance and outcomes in non-Western or lessindustrialized contexts.1 Instead, educational research in developing countries has focused primarily on socioeconomic status and family structure as determinants of educational outcomes (see Buchmann & Hannum (2001) for a review). The failure to examine social and cultural capital in such contexts is unfortunate. Greater attention to social and cultural capital in studies of schooling in less-industrialized societies could be informative on two fronts. First, such research could help determine whether and how the importance of such resources in the transmission of status is conditioned by social context and varies across societies. The substantial political, economic, and sociocultural differences found in developing countries can serve to richly inform our knowledge of cross-societal variations in the relationship between family background and educational performance. Second, if differences in social and cultural capital are fundamental determinants of unequal educational outcomes, failure to analyze their effects will lead to an incomplete understanding of educational stratification and, in turn, poorly-formulated attempts to improve equality of educational opportunity in developing regions.
Getting Ahead in Kenya
135
With these broad goals in mind, this paper addresses two questions in the developing country of Kenya: Do family social and cultural capital play a role in adolescents' school achievement, net of family socioeconomic status? If so, what aspects matter most for getting ahead in Kenya? Specifically, I examine the role of shadow education as an organizational feature of Kenyan education that provides cultural capital to some students and is utilized by some families to advance their children's likelihood of educational success. Shadow education consists of nonformal, extra-educational activities that aid children in preparing for national exams and improving their academic performance. With interview data for more than 500 Kenyan households, I investigate the determinants of shadow education, in order to establish which students are most likely to have access to this resource. I then assess the consequences of shadow education as well as measures of family social and cultural capital for two educational outcomes: grade repetition and educational performance. The results demonstrate that shadow schooling is a valuable educational resource that is quite unevenly distributed among Kenyan children. Those enrolled in school who participate in shadow education have a lower incidence of grade repetition and higher academic performance than those who do not. Aspects of social and cultural capital also prove to be important. For example, children with many siblings perform less well in school, likely because of a diffusion of social capital in the household. Family cultural capital, measured as the language spoken in the home and mother's reading habits, reduces the likelihood of grade repetition, net of other family background factors.
SOCIAL AND CULTURAL CAPITAL: RELEVANT CONCEPTS IN DEVELOPING SOCIETIES?

In many developing societies, formal schooling is a relatively recent institution; it is often an import of colonialism that has been in existence for fifty years or less. The recent development of formal education, along with the limited resources that most Third World governments devote to education, are reflected in the low enrollment and completion rates, even at the primary level, in many developing countries. Under such conditions, we might expect family wealth and the local availability of schools to be the primary factors determining children's educational enrollment and attainment. Upon closer examination, however, several features of developing societies and their educational systems suggest that aspects of family social and cultural capital could be important determinants of patterns of children's educational
136
CLAUDIA BUCHMANN
participation and achievement in these contexts. Low levels of economic development combined with a weak position in the world system have stunted the growth of wage employment in most developing countries. High unemployment rates have led to a degree of competition for educational credentials seldom seen in industrialized countries. Moreover, in post-colonial societies, educational systems retain many of the features of the British, French or German systems that were imported during colonialism, especially a reliance on national school-leaving examinations in order to determine which children progress to higher levels of education. The highly-competitive nature of educational and employment systems in many developing countries has important implications for the strategies that families use to help their children succeed. It is not enough for families provide financial resources for school enrollment; they must also find ways to ensure that their children's academic performance meets the standards necessary to belong to the select group that continues on to higher levels of schooling. In such cases, it is reasonable to expect that families utilize various cultural and social resources in order to help their children persist as long as possible in the competitive educational system. Social Capital Social capital, broadly defined, inheres in the relations between people (Coleman, 1988). "Trust, obligations and expectations, norms, and shared information are all examples of social capital because they are resources that arise from the social relationships of individuals who share membership in a common social structure" (Carbonaro, 1998, p. 296). In terms of children's educational progress, researchers have focused on the role of social capital within the family and between the family and external others, such as schools and communities. The relationships among parents in a community or parents' relationships with community institutions comprise social capital that is relevant to children's development. For example, Coleman and Hoffer (1987) attributed differences between the educational performance of students in public versus Catholic schools in the United States to differences in the social capital of the communities surrounding the schools, maintaining that Catholic school communities have greater social capital through their shared norms, expectations and information. Within the family, social capital is evident in the relations between children and their parents as well as other members of the household and is quite distinct from family human capital. Social capital is one route through which human capital can have an impact on children's educational progress. Thus, parental involvement in their children's schooling - captured by the frequency of parentchild communication about progress in school, parents' expectations for their
137
children, and time spent assisting children with homework and other educational tasks - should be influential for children's educational outcomes. To the extent that other family members (such as older siblings and other adult relatives) take part in these activities, they are also relevant.2 Researchers must be careful to conceptualize social capital in ways that are sensitive to the contexts they are studying. For example, aspects of family structure are related to social capital within the family, since social capital is generated through the "density and consistency of educationally-focused relationships that exist among parents, children and schools" (Teachman, Paasch & Carver, 1997, p. 1344). Studies in industrialized countries demonstrate that the family's ability to foster educationally-focused relationships between parents and children depends on the number of adults and children in the household. One-parent families and families with many children have less social capital than two-parent families with fewer children (Astone & McLanahan, 1991), since parental attention and material resources are diluted with each additional child in the household (Coleman, 1988; Blake, 1989; Downey, 1995). Different family structures in less-developed societies may lead to different relationships between family structure and educational outcomes. For example, in some African societies single-female headship appears to be associated with greater educational opportunities for children. In a study of adolescent girls in South Africa, Fuller and Liang (1999) reported that father absence served to decrease girls' risk of leaving school. Similarly, Lloyd and Blanc (1996) analyzed the effects of headship in seven African countries and found that children in female-headed households were consistently more likely to be enrolled in school and to have completed grade four than children in households headed by men. These findings may be linked to stronger extended kinship structures in Africa that serve to condition the effects of single parenthood on children. For example, Lloyd and Blanc (1996) also noted that extended family networks in Africa enable children with academic promise to move to households of "patron" family members, who help them gain access to high quality schools. Extended family systems can also provide resources to mediate the otherwise negative effect of sibship size on children's schooling. In Kenya, I found no effect of sibship size on children's probability of enrollment in prior research (Buchmann, 2000), while Montgomery and Lloyd (1997) found no impact of excess fertility (fertility departing from stated family-size preferences) on educational attainment. Siblings may even contribute to household resources in some contexts. In rural Botswana, the number of seven to fourteen year-old children in the household was positively related to educational enrollment and attainment (Chernichovsky, 1985), while sibship size was positively associated with years of schooling in Kenya (Gomes, 1984).
138
CLAUDIA BUCHMANN
These differences in family structure must be acknowledged in any attempt to measure the impact of social capital in the family on children's educational attainment and performance. For example, the number of "non-parent" adults in the household may be an important measure of social capital in societies where nuclear families are not the norm. Grandparents, adult siblings, and other relatives residing in the household are potential sources of support that may serve as social capital for children, whether they reside in single-parent or twoparent households. Chernichovsky (1985) found that the presence of grandparents in the household was positively related to children's school enrollment because they could contribute labor and resources to the household that might otherwise be borne by children. Similarly, Lloyd and Gage-Brandon (1994) noted that in Ghana older siblings contribute resources to the household that may benefit later-born children. Cultural Capital Following Bourdieu (1984), research in industrialized countries has usually conceptualized cultural capital as participation in highbrow cultural activities such as attending concerts, reading literature, and visiting art museums. Several studies have found that cultural capital, whether measured as students' or parents' participation in and preferences for such activities,3 has significant positive effects on educational attainment (DiMaggio, 1982; DiMaggio & Mohr, 1985; De Graaf, 1986). Of course, the cultural codes that are deemed valuable should vary from society to society (De Graaf et ah, 2000, p. 93). Because societies differ in the institutionalization of high culture, it is important to consider how cultural capital is determined by national differences in educational structures and other societal characteristics. For example, it appears that Bourdieu's original conception of cultural capital is more appropriate for the case of his home country, France, than for the U.S. or Great Britain. In fact, some researchers have argued that an emphasis on highbrow cultural activities misses other aspects of cultural capital that should be more relevant to educational success for some groups and in some societies (Farkas, 1996; Lareau & Horvat, 1999; De Graaf et al., 2000). Especially among low-status or poor populations, a conceptualization of cultural capital that focuses on the reading habits and linguistics skills of parents or other family members may be more relevant. Parents can transmit linguistic and cognitive skills to children through their own reading behaviors and by helping children become familiar with reading. Some research has found parental reading habits to be beneficial to children's educational performance (Farkas, 1996; Crook, 1997). One recent study of educational attainment in the Netherlands found that parental reading habits were more important for children
139
from poor origins than for those from high socioeconomic backgrounds (De Graaf et alM 2000). Similarly, in developing countries, educational success is less likely to be predicted by parents' participation in Western, highbrow cultural activities, and more likely to be based on parental linguistic skills and cognitive behaviors. Other considerations figure into the task of constructing valid measures of cultural capital in these societies. Due to the limited availability and the high cost of books in developing countries, even literate people are less likely to read books than more accessible materials such as newspapers and magazines. Leisure reading of novels and literature is probably less common than reading for information and knowledge, except perhaps among the urban middle class. Attempts to construct appropriate measures of reading habits or preferences as indicators of cultural capital (as opposed to socioeconomic status) must account for these realities. In many developing countries, language skill may be an especially important form of cultural capital. In post-colonial societies in particular, the language of school instruction often is not the language spoken in the homes of most children. In Kenya, for example, the nation's two official languages, English and Swahili, are used in school. Swahili is the preferred language in primary schools, English is used in secondary schools and universities. But the great majority of Kenyans speak one of 59 languages other than Swahili or English as their mother tongue.4 Experience speaking either of these languages at home before starting school may help children make the transition to school and more generally aid their academic performance. Thus, speaking Swahili or English at home regularly may provide children with valuable cultural capital and savvy parents may go to special effort to speak Swahili with their children in order to prepare them for school. Finally, shadow education is a feature of Kenyan education that may provide some children with cultural capital. Shadow education generally takes the form of tutoring or after-school classes that mimic or "shadow" more formal schooling processes (Baker et al., 2000; see also, Bray, 1999). Shadow education has flourished in many societies and is likely related to educational performance. Recent research by David Baker and associates (2000) finds considerable cross-national variation in the prevalence and goals of shadow education. In many countries it serves primarily a remedial role for lowperforming students, while in some societies it is used to enhance learning. To this point there has been very little empirical research on shadow education as it relates to academic achievement in developing countries.5 Yet, anecdotal evidence suggests that shadow education has become a common phenomenon in developing countries, especially those that utilize central examination
140
CLAUDIA BUCHMANN
systems. Most studies of shadow education to date have focused on East Asia and Europe. Stevenson and Baker (1992) found shadow education to be an important predictor of high exam performance and acceptance to university in Japan. In Greece, Katsillis and Rubinson (1990) noted that students utilize tutors to prepare for national exams. Perhaps it is surprising that shadow education exists at all in developing regions. Stevenson and Baker (1992) predicted that shadow education should be most prevalent in societies where there are tight linkages between the labor market and the educational system. This is rarely the case in developing countries where shadow education has flourished even students with high educational credentials face great challenges finding employment and there are few guaranteed paths to high status occupations. But a closer look at the situation in Kenya suggests that it is precisely for these reasons - the extreme competition for educational credentials and employment - that shadow education has come to be seen as a valuable route for improving one's chances of getting ahead. Moreover, as Baker et al. (2000) suggest, in developing countries with less than full enrollment and low levels of government funding for education, shadow education may develop "as a kind of market reaction to underdevelopment" (p. 24). The following section examines these aspects of Kenyan society in detail.
EDUCATIONAL MOBILITY IN KENYA: A SERIES OF CONTESTS

According to Aliu Fafunwa, former Nigerian Minister of Education, "African countries are running examinations rather than educational systems" (Daily Nation, June 30, 1994). This comment is applicable in Kenya, where the most important determinant of a student's academic progress is passage of the highly-competitive exams at the end of primary and secondary school. The first of these contests, the Kenyan Certificate of Primary Education (KCPE), is administered at the end of the eighth year of primary education and is used to select students for secondary schools. The score on the KCPE determines the quality of secondary school the child can attend;6 the highestscoring students gain access to the most prestigious national schools. Despite a great expansion of secondary schools in the nearly four decades since independence, many Kenyan children do not have the chance to pursue education beyond primary school. The nation's 3,000 public and private secondary schools typically admit only 40 to 50% of the more than 400,000 primary school students who take the KCPE annually. Thus, approximately half of all students completing primary school are pushed out of the race for
141
secondary schooling because there is no room for them (Daily Nation, January 4, 1996). In addition to an inadequate supply of secondary schools, problems in the labor market are to blame for the highly-competitive nature of the educational system. The formal wage sector has grown slowly since independence and remains small compared to the population. Over 80% of the total labor force is engaged in agriculture and pastoralism (Republic of Kenya, 1993); only about 14% is employed in the formal sector while an additional 3.5% works in the informal sector (Hughes, 1991). As the labor force grows at an average rate of 3.6% annually, there is intense competition in the wage labor sector (Republic of Kenya, 1991a). Moreover, as a result of two decades of economic stagnation, there are even fewer jobs for the ever-growing number of secondary-school and university graduates. In 1986, urban unemployment reached 16%. Over 55% of the unemployed was under 25 years of age and the unemployment rate for women (24%) was double that for men (12%) (Republic of Kenya, 1991b, p. 20). Therefore, while education has become a necessity for formal sector employment, it is not a guaranteed path to occupational mobility. Grade Repetition and Shadow Education The stark competition for educational credentials and, ultimately, employment has led to two common educational outcomes: high rates of grade repetition and shadow education. A survey of more than 500 households in three regions of Kenya conducted in 1995, described below, provides an indication of the prevalence of these practices. Of enrolled adolescents ages 13-18 in the survey (N = 526), 66% had repeated a grade at least once and nearly one-fourth had repeated twice or more. Given the importance of the national examinations at the end of primary and secondary school, it is in the interest of students and schools alike to see that students do not proceed to the next grade if they have not mastered the knowledge and skills of the current grade.7 Repeating more than once is generally a sign that a student is struggling academically and in danger of dropping out. Students who fall behind their peers may grow frustrated and lose interest in school, especially if they are asked to weather the experience of repeating a grade more than once. As one survey respondent who attributed her daughter's dropout to repetition explained: "She left school because she felt she was too big for standard (grade) seven and was angry that she was asked to repeat again." The impact of grade repetition in Kenya is reflected further in the inflated average age of students at each level of education. More than 45% of sixteen year-olds and nearly 40% of seventeen year-olds in the survey were enrolled in primary school, despite the fact that according
142
CLAUDIA BUCHMANN
to the government, this age group should be mid-way through secondary school. While the expected age for beginning secondary school is fourteen, less than 10% of fourteen year-olds and less than 20% of fifteen year-olds were actually enrolled in secondary school. Given the risks associated with frequent grade repetition and the importance placed on examination performance, shadow education hasflourishedin Kenya. Shadow education includes private and school-based tutoring and exam preparation classes. In the past decade, a virtual cottage industry has grown around examinations. Many schools offer extra classes and tutoring after school or on weekends for an additional cost. Some parents hire private tutors. Tutoring appears to be most common in the later years of primary school when students are preparing for the national primary school exam. The phenomenon has spread rapidly in cities, especially Nairobi. It is less common in rural regions where children often must walk greater distances to school or are expected to help with farm work and may have less discretionary time for extra educational activities. Also, since most rural schools do not have electricity, evening schoolbased tutoring is not a practical option. These urban-rural variations in access to shadow education may provide urban youth with an advantage in the series of contests that comprise the Kenyan educational system. In addition to regional differences in the availability of shadow education, the cost of tutoring likely results in unequal participation in shadow education. Whether school-based or home-based, tutoring is an additional educational expense for families. According to the survey, tutoring expenses averaged 180 shillings (4.3 U.S. dollars) per child per month,8 but actual costs vary greatly. The fee for students attending short after-school sessions may be as little as 30 shillings per month, while private tutors may charge as much as 1000 shillings monthly. Costs also vary according to the amount of tutoring received. Among youth in the survey, time spent in shadow education activities ranged from four to 24 hours per week and averaged ten hours weekly. Children in lower grades might attend one-hour tutoring sessions after school, four days a week. Students in standard seven or eight who are preparing for the KCPE may spend as much as three hours each weekday and all day Saturday in private tutoring and examination preparation classes. Finally, girls may be underrepresented in shadow education activities. Unquestionably, girls in Kenya have made significant educational gains in recent decades. Other research using data from the same survey found no gender differences in the likelihood of enrollment for adolescents 13-18 years of age (Buchmann, 2000). This suggests that Kenyan parents recognize the importance of educating both sons and daughters. But lingering gender stereotypes regarding job prospects and gender biases in children's expected contributions
143
to housework may mean that parents are less willing to provide additional educational resources to their daughters, especially in cases where family resources are severely limited (i.e. in poor families or those with many children). Results from the survey reflect strong gender-role stereotypes in Kenya. Forty percent of respondents rank math as the most important school subject for boys, but one-third rank home economics as the most important subject for girls. The survey also indicates that girls do more housework (4.22 household tasks on average) than boys (3.13), and that girls in rural households owning land perform the highest average number (5.49) of household tasks (Buchmann, 2000, p. 1360). The greater expectations for girls to contribute to housework may be especially detrimental to their chances of taking part in shadow education. Motherhood also may interfere with secondary school completion and higher education for females. At the time of the survey, the median age of first birth was 19 in Kenya (National Council for Population and Development, 1994). These norms and sex stereotypes likely impact female participation in "extra" educational activities such as shadow schooling and are partly responsible for the limited participation of girls at higher levels of education.9 In sum, several features of Kenyan society suggest that educational mobility, while not a guaranteed route to employment, is a highly desirable outcome that requires the investment of multiple family resources. Some of these resources, such as the social capital embedded in extended kinship networks, are rooted in traditional social systems that have existed for centuries; others, such as participation in shadow schooling, are much more recent developments. The analysis that follows seeks to understand how families utilize these resources to help their children get ahead in Kenya today.
DATA AND METHODS

In order to understand adolescents' schooling experiences in Kenya, the author and a team of interviewers surveyed nearly 600 Kenyan households in 1995. The mother in each household was asked a wide range of questions about the household and children's schooling. Among other issues, the survey examined the educational experiences - including enrollment, grade repetition and participation in shadow education - of all youth ages 13-18. Three geographically-distinct districts were surveyed: Nairobi, Kwale, and Murang'a. Nairobi is Kenya's capital and largest city; an estimated 1.74 million people of all ethnic and socioeconomic backgrounds reside in a densely-populated 180 square miles of Nairobi District (Republic of Kenya, 1994). Murang'a is an agriculturallyfertile district in the central highlands of Kenya. Kwale is a poor rural district on the coast of the Indian Ocean. Within each of these districts, two or three
144
CLAUDIA BUCHMANN
representative areas were chosen. In Nairobi, where houses are numbered, households were randomly selected from a sampling frame of all addresses. Households plots are not numbered in rural areas, so all households in the area were sampled and those with school-age children were interviewed. The final sample, while not representative of the entire country, offers a diverse crosssection of the Kenyan population and captures the variations expected to influence children's educational experiences. In order to establish which Kenyan adolescents are most likely to have access to shadow schooling, a logistic regression analysis examines the determinants of shadow education, a dummy variable coded one if the student participates in informal educational activities including private tutoring at home or school, or classes outside of school hours. Then, ordered logistic regression is used to analyze the determinants of grade repetition, measured as the total number of times (0, 1, 2, 3 or more) the student has repeated a grade. A second ordered logistic regression analysis examines the determinants of educational performance, which is the mother's report of the student's academic performance in the past year, based on a five-category scale, ranging from 1 (failing) to 5 (very good). While parental report is less desirable than a direct schoolbased measure of academic performance, the survey queried parents about the source of their knowledge regarding their child's academic performance and the great majority cited a recent school report card as the source of this information. Thus, it is reasonable to presume that this measure reflects a school-based assessment of educational performance, but parental bias cannot be ruled out. Prior studies lacking school-based measures of academic performance have also relied on parental report of student performance (e.g. Lipman & Offord, 1997). Ordered logistic regression is used for the analyses of grade repetition and academic performance because these dependent variables are ordinal. They are comprised of categories that can be ranked from low to high, but the distances between adjacent categories are unknown. Ordinary least squares (OLS) regression relies on the assumption that the intervals between adjacent categories are equal and may provide misleading results in the case of such ordered dependent variables (Long, 1997). The analyses are conducted for enrolled children, ages 13-18 (N =526). Missing data reduce the sample to 506 cases. Descriptive statistics and definitions for all dependent and independent variables are reported in Table 1. Independent Variables All models include the Age and Sex of the adolescent, the household's Financial Situation, and Parent's Education. Age and sex are included to examine whether participation in shadow schooling or educational outcomes are significantly
145
Table 1. Variables Used in Analysis.

Name Description Mean S.D.
Educational Outcomes: Shadow Education Dummy variable coded 1 if child has participated in non-formal educational activities including private or group tutoring or after school classes in past three years. Grade Repetition Number of times (0, 1, 2, 3 or more) child has repeated a grade in school. School Performance Child's school performance in past year. Coded as: 1 = failing, 2 = poor, 3 = adequate, 4 = good, 5 = very good. Independent Variables: Family Social Capital Number of Adults Total number of adults in the household (ranges from 1-9). Number of Children Total number of children in the household (ranges from 1-17). Proportion of Child's ordinal position in the sibship. Children Older Calculated as number of older siblings/total number of children. Homework Help Coded 1 if child gets help with homework from a family member. Family Cultural Capital Home Language Coded 1 if family usually speaks English or Swahili at home. Mother's Reading Coded 1 if mother has read a newspaper in Habits the past month. Other Age of Child Sex of Child Parent's Education
0.34
0.47
0.97 3.61
0.89 0.86
3.27 5.68 0.30
1.57 2.37 0.25
0.46
0.50
0.26 0.33
0.44 0.47
15.00 Child's age ranging between 13 and 18 years. Coded 1 if child is female. 0.51 Highest level of education attained 2.19 by either parent. Coded as: 0 = none 1 = standard 1-4 (lower primary) 2 = standard 5-7 (upper primary) 3 = form 1-4 (lower secondary) 4 = form 5 or more (upper secondary or university).
1.07 0.50 1.09
146 Table 1.
Name Financial Situation Description
CLAUDIA BUCHMANN Continued.

Mean 2.18 S.D. 1.09
Urban High Economic Status
Mother's perceived financial well-being of the household in the past five years. It closely approximates income and is coded as: 1 = have had a very hard time getting by 2 = with careful planning and help from others, have been able to meet some needs 3 = have been able to fulfill most needs but cannot plan for the future 4 = have been able to fulfill needs and saved some money or made an investment for the future. Coded 1 if household resides in urban area. Coded 1 if household has windows.
0.27 0.29
0.45 0.45
different for boys and girls or vary by age. Since in developing countries, people often do not know or cannot calculate their income, respondents were asked about household wealth and financial well-being in addition to income. The financial situation variable reports the mother's perception of the household financial situation in the past five years, ranging from one to four. It is highly correlated with household income and has fewer missing cases. Parent's education is measured as the highest level of education of either parent. In addition to these variables, the first model of each analysis includes a variable for high economic status, in order to examine whether the wealthiest households in the sample differ significantly in their ability to acquire shadow education for their children. It is a dummy variable indicating whether the house has glass windows, as field observations and cross-tabulations of survey results indicate that the presence of glass windows is a good proxy measure for high socio-economic status, especially in rural areas. In order to preserve degrees of freedom, it is kept in remaining models only if it is initially significant. Family social capital: Three measures of family structure are included as independent variables order to test whether the structure of the family influences educational outcomes, presumably by affecting the amount of social capital available for each child in the household. Number of children indicates the number of people 18 years-old or younger in the household. Number of adults consists of all people older than 18 residing in the household, including parents, older siblings, grandparents and other relatives. As noted above, in societies where households frequently are comprised of extend kinship groups, the number of adults in the household may be an important predictor of
147
children's school success, since adults can provide economic resources and social support for children that facilitate their educational achievement. Additionally, older siblings residing in the household may be a valuable source of social capital for younger children. Especially when parents have little or no experience with formal schooling, older siblings may be better able to help with Table 2. Frequency Distributions of Categorical Variables Used in Analysis (N = 506).
Variable Grade Repetition Category 0 1 2 3+ 5 4 3 2 1 (very good) (good) (adequate) (poor) (failing) Percent 33.8 42.5 16.4 7.3 100.0 11.8 47.9 33.1 4.1 3.1 100.0 0.8 42.1 22.3 17.0 7.3 6.5 4.0 100.0 0.6 2.6 8.5 16.4 24.7 23.1 11.1 6.7 6.3 100.0 1.6 24.1 74.3 100.0
School Performance
Number of Adults
1 2 3 4 5 6 7+ 1 2 3 4 5 6 7 8 9+ English Swahili Other
Number of Children
Home Language
148
CLAUDIA BUCHMANN
homework or define appropriate behavior and motivation for educational success. The proportion of children older than the target child is included in the analyses to examine this possibility. Table 2 presents the response distributions for select categorical variables used in the analysis. The distributions for the number of adults and number of children indicate that large families are common in Kenya. More than half of children live in households with three or more adults; only four households in the sample have just a single adult present. Households have an average of 5.7 children under the age of 18. It is important to note that measures of family structure only indirectly reflect social capital processes. The finding of a positive relationship between number of adults in the household and academic performance would be consistent with social capital arguments, but cannot demonstrate definitively that social capital is the main intervening mechanism between family structure and educational outcomes. In contrast, a fourth measure of social capital, homework help, directly indicates the presence of an educationally-focused interpersonal relationship between family members. It is a dummy variable that is coded 1 if the student receives help with homework from a family member. In most cases, help is received from a parent; older siblings are the second most common response to the question, "if the child receives help with his/her homework at home, who is the main person that helps him/her?"10 Cultural capital: Two measures of cultural capital are included as independent variables in the analyses of grade repetition and academic performance. First, home language is a dummy variable coded 1 if the respondent reported that either English or Swahili was the main language spoken in the household, especially with children. As explained above, such language skills should be valuable in easing a child's transition to school. Children who reside in households where English or Swahili are not spoken may also have a harder time integrating school experiences into their home lives. If parents do not speak the language of school instruction, they may be less involved in their children's education and less able to help children with homework. Second, mother's reading habits is a dummy variable indicating whether the mother has read a newspaper in the past month. Because Kenyan newspapers are printed in English or Swahili, this variable reflects the mother's active reading abilities in the languages of central importance to children's schooling. It goes a step beyond measuring literacy or language ability, however, to measure actual reading behavior. Additionally, shadow education, used as a dependent variable above, is included as an independent variable in these analyses in order to examine whether shadow education serves to improve academic performance or reduce the likelihood of grade repetition.
149
RESULTS
For each analysis, the first model includes individual and family background variables including the high economic status variable; then the three family structure variables are added in the second equation. In order to preserve degrees of freedom, family structure variables are kept in remaining models only if they have a statistically significant effect on the dependent variable. In the analysis of shadow education, a variable indicating urban residence is added in the final equation in order to examine whether there are significant urban-rural differences in access to shadow education. The analyses of grade repetition and educational performance also test the impact of social and cultural capital measures as well as shadow education. Who Participates in Shadow Education? Table 3 examines the determinants of shadow education. Model one includes adolescents' age and sex, as well as parent's education, household financial situation, and the variable for high economic status. Age is not a significant predictor of receiving tutoring but sex is negative and significant, indicating that families may be less willing to invest in additional educational expenses like shadow education for daughters. Parent's educational background is a significant positive predictor of shadow education. Finally, the strong positive effect of the high economic status variable indicates that wealthier families are more able to afford this additional educational expense for their children. Models two and three examine the effects of three measures of family structure: number of adults, number of children, and the proportion of older siblings. In model two, the number of adults is significant, indicating that the greater the number of adults in the household the more likely it is that the child participates in shadow education. This finding supports the idea that adults may contribute either financial resources, that enable the family to afford shadow education, or human capital, that leaves children free to pursue extra educational activities. In this model, the previously significant gender effect narrowly misses significance (p = 0.052). Model three includes an interaction term for the number of children and the sex of child in order to see whether girls in families with many children are more disadvantaged than other children, but the variable is not significant. Model four assesses the effect of urban residence. As expected, residence in Nairobi significantly increases a child's odds of participating in shadow education. While urban children have no greater odds of enrollment than rural
150 Table 3.
CLAUDIA BUCHMANN Coefficients from Logistic Regression of Participation in Shadow Education on Selected Independent Variables.
Model 1 -0.005 (0.064) -0.315* (0.187) 0.203* (0.115) 0.073 (0.094) 0.568** (0.230) Model 2 -0.023 (0.065) -0.308 (0.189) 0.262* (0.116) 0.075 (0.096) 0.493* (0.234) 0.154* (0.079) Model 3 0.004 (0.065) 0.055 (0.426) 0.250* (0.115) 0.072 (0.095) 0.530* (0.233) Model 4 -0.013 (0.065) -0.316* (0.190) 0.129 (0.120) 0.074 (0.097) 0.288 (0.241) 0.121* (0.064)
Independent Variable Age of Child Sex of Child (Female = 1) Parent's Education Financial Situation High Economic Status Number of Adults Proportion of Children Older Number of Children
-0.026 (0.042)
Number of Children x Sex of Child Urban Constant LRX2 P -1.227 (0.974) 22.35 0.00 -1.429 (1.003) 4.54 0.10
0.535 (0.427) 0.047 (0.044) -0.065 (0.063) 0.794** (0.255) -1.490 (1.014) 14.78 0.00
-1.881 (1.063) 2.78 0.43
N = 506. Coefficients are unstandardized and robust standard errors are in parentheses. *p < 0.05, one-tailed test, ** p < 0.01, one-tailed test. LRx2 is based on comparison to most previous nested model, for model 1, it is based on comparison to baseline model.
children (Buchmann, 2000), they are more likely than rural children to take part in shadow education. Since high economic status and parent's education are reduced to nonsignificance in this model, it appears that urban residence is more important than either of these aspects of family background in determining whether a child receives tutoring. Thus, in terms of shadow education, some Kenyan children have distinct advantages over others. Boys and children in urban areas are more likely to participate in tutoring and other shadow education activities. When it comes time for national exams, girls and rural youth are competing on an uneven playing field.
151
Table 4. Coefficients from Ordered Logistic Regression of Grade Repetition on Selected Independent Variables.
Independent Variable Child Receives Shadow Education Home Language Mother's Reading Habits Homework Help Number of Adults Proportion of Children Older Number of Children Number of Children x Sex of Child Age of Child Sex of Child (Female = 1 ) Parent's Education Financial Situation High Economic Status Threshold 0/1 Threshold 1/2 Threshold 2/3 or More -0.062 (0.094) 0.510 (0.499) -0.053 (0.065) 0.132** (0.048) 0.262** (0.056) -0.739 (0.343) -0.215** (0.088) -0.445 (0.091) 1 2 3 -0.301* (0.182) -0.915** (0.196) -0.647** (0.242) -0.309 (0.191) 4 5 6
0.246** (0.050) 0.025 (0.162) -0.147 (0.098) -0.028 (0.091) -0.425 (0.267) 2.485 (0.783) 4.428 (0.805) 5.836 (0.819)
-0.054 (0.064) 0.128** (0.048) 0.244** (0.051) -0.732* (0.340) -0.205* (0.091) -0.034 (0.092)
-0.080 -0.070 (0.064) (0.057) 0.140** 0.145** (0.049) (0.048) 0.260** 0.243** (0.051) (0.051) -0.738* -0.815** (0.340) (0.349) -0.175* -0.093 (0.088) (0.103) 0.009 -0.017 (0.090) (0.097)
-0.056 (0.061) 0.121** (0.047) 0.230** (0.052) -0.663* (0.340) -0.207** (0.090) -0.029 (0.094)
2.299 (0.884) 4.242 (0.899) 5.657 (0.912)
2.012 (0.856) 3.957 (0.870) 5.372 (0.887)
2.068 (0.866) 4.073 (0.884) 5.519 (0.902)
2.114 (0.841) 4.089 (0.855) 5.505 (0.870)
1.766 (0.873) 3.717 (0.887) 5.138 (0.906)
N = 506, except for models 5 (N = 504) and 6 (N = 503). Coefficients are unstandardized and robust standard errors are in parentheses. * p < 0.05, one-tailed test, ** p < 0.01, one-tailed test.
Determinants of Grade Repetition and Academic Performance Table 4 examines the impact of shadow schooling and the measures of family social and cultural capital on grade repetition. The first model indicates that the chances of repeating a grade increase with age. No other variables are significant. Model two includes the measures of family structure and an interaction
152
CLAUDIA BUCHMANN
term to see whether the number of children in the household differentially impacts girls and boys. Only the interaction term is significant, indicating that for girls the number of children in the household increases the frequency of repeating a grade." Model three includes shadow education as an independent variable in order to determine whether participation in extra school activities reduces a students' chances of grade repetition, net of other variables. Indeed, students who participate in shadow education activities are significantly less likely to repeat grades. Models four and five test the two additional cultural capital measures. Adolescents from homes where English and Swahili are commonly spoken are less prone to grade repetition, suggesting that language skills developed in the home are important for buffering students from academic problems in school. Students' whose mothers have read a newspaper in the past month are also less prone to repeating grades. A mother's active use of reading skills influences her children's academic behaviors, perhaps because she can more effectively support and facilitate their education. Model six tests the final measure of social capital; homework help. Whether or not students get help on their homework from a family member has no significant impact on grade repetition. Note that in all models, parent's education is negatively related to grade repetition, although this relationship is not significant in two of the six equations.12 Importantly, home language and mother's reading habits impact the likelihood of grade repetition net of parent's education, indicating that these forms of cultural capital can serve as filters through which the benefits of parent's education are transmitted to children.13 Finally, Table 5 examines the determinants of academic performance. In model one, only age is significantly related to academic performance. Older children are less likely to perform well, likely because the level of competition becomes greater in later years of schooling. From Table 4 it is also clear that the likelihood of grade repetition increases with age, so the effect of age may also reflect a negative relationship between repetition and academic performance. The non-significant effect of age when grade repetition is added to the model and the strong negative effect of grade repetition in model eight further support this idea. Of the family structure variables included in model two, only the number of children in the household is significant; it is negatively related to academic performance, indicating that youth with many siblings are less likely to have high academic achievement. This result is in line with the expectation that the attention that parents can provide to any one child is diluted with each additional child in the household. This finding is especially interesting in light of prior research that finds no negative impact of sibship size on school enrollment or
153
Table 5. Coefficients from Ordered Logistic Regression of Academic Performance on Selected Independent Variables.
Independent Variable Child Receives Shadow Educationi Home Language Mother's Reading Habits Homework Help Grade Repetition Number of Adults Proportion of Children Older Number of Children -0.079 (0.092) 0.603 (0.467) 1 2 3 4 0.463** (0.182) -0.067 (0.195) 0.295 (0.247) 0.275 (0.195) -0.887** (0.121) 5 6 7 8
-0.074* -0.062 -0.069* -0.064* -0.059* -0.060* -0.058* (0.037) (0.050) (0.032) (0.032) (0.031) (0.030) (0.031) -0.001 Number of Children x Sex of Child (0.063) Age of Child -0.135** -0.156** -0.134** -0.136** -0.135** -0.133** -0.123** -0.033 (0.050) (0.053) (0.050) (0.050) (0.050) (0.050) (0.051) (0.092) 0.092 0.122 0.117 0.119 Sex of Child 0.148 0.124 0.124 0.130 (Female = 1 ) (0.177) (0.178) (0.411) (0.178) (0.177) (0.179) (0.180) (0.178) Parent's -0.022 -0.018 -0.007 -0.039 -0.003 -0.072 -0.017 -0.071 Education (0.098) (0.088) (0.088) (0.088) (0.089) (0.106) (0.087) (0.085) Financial 0.026 0.007 0.024 0.034 0.039 0.040 0.034 0.037 Situation (0.093) (0.092) (0.093) (0.093) (0.094) (0.100) (0.095) (0.091) High Economic 0.160 Status (0.219) Threshold Failing/ -5.370 -6.060 -5.737 -5.720 -5.748 -5.806 -5.482 -5.506 Poor (0.867) (0.941) (0.933) (0.897) (0.899) (0.901) (0.924) (0.945) Threshold Poor/ -4.474 -5.165 -4.841 -4.822 -4.852 -4.908 -4.584 -4.572 (0.836) (0.905) (0.896) (0.869) (0.870) (0.872) (0.895) (0.907) Adequate Threshold Adequate/-2.284 -2.961 -2.643 -2.612 -2.656 -2.703 -2.382 -2.151 (0.836) (0.874) (0.865) (0.839) (0.840) (0.841) (0.864) (0.885) Good 0.139 -0.522 -0.211 -0.155 -0.222 -0.276 Threshold Good/ 0.071 -0.513 (0.804) (0.863) (0.863) (0.835) (0.835) (0.834) (0.870) (0.897) Very Good
N = 489. Coefficients are unstandardized and robust standard errors are in parentheses. *p < 0.05, one-tailed test, ** p < 0.01, one-tailed test.
154
CLAUDIA BUCHMANN
attainment in several developing countries, including Kenya. Considering these findings, we might speculate that large family size need not be detrimental for children's school entry and persistence, but that children's educational achievement does suffer in families with many children. Further research that explicitly examines the relationship between family size and educational achievement is necessary to test this speculation. Model three includes the interaction term for number of children and the sex of child in order to examine whether there is a gender-specific impact of sibship size and finds no significant effect of this variable. Models four through eight include the remaining cultural and social capital measures and grade repetition, as independent variables. Of the three cultural capital variables (shadow education, home language, and mother's reading habits), only shadow education is a significant determinant of academic performance. Also noteworthy and as expected, the frequency of grade repetition is negatively related to academic performance; the more times a student repeats a grade, the lower his or her academic performance. In this final model, the age variable ceases to be significant, suggesting that the significant effect of age in prior models mainly reflects the inverse relationship between grade repetition and academic performance. The language spoken within the home, mother's reading habits, and help with homework from family members have no impact on academic performance. While home language and mother's reading habits are valuable forms of cultural capital for buffering youth from negative educational outcomes, such as grade repetition, apparently they are less important for boosting academic achievement. Finally, analyses that included interaction terms with gender and home language, mother's reading habits, and homework help (not shown) revealed no significant effects, suggesting that these aspects of family social and cultural capital do not differentially influence girls' and boys' educational outcomes.
CONCLUSIONS
One limitation of the current study is the cross-sectional nature of the data. Without longitudinal data on adolescents' educational histories, we cannot determine the precise causal relationships between the variables. For example, while shadow education likely boosts achievement, it is also possible that parents chose to send their most academically proficient children to shadow schooling and thus, these factors are co-determined. It is also possible that more motivated families acquire shadow schooling for their children and that the effects of shadow education reflect, in part, unmeasured aspects of family motivation.
155
These caveats notwithstanding, the results suggest that shadow education is a particularly valuable educational resource for getting ahead in Kenya's highly competitive educational system. Students who are involved in tutoring or examination preparation classes are less likely to repeat grades and more likely to have higher academic performance. In light of the importance of shadow education for these educational outcomes, the unequal patterns of participation in shadow education are all the more disturbing. The advantage that boys and urban youth have in terms of access to shadow schooling may mean that even as gender and urban-rural disparities in enrollment and attainment are addressed, girls and rural youth will under perform relative to their counterparts on national examinations. Failure to recognize disparities in shadow education as a potential source of disparities in examination performance could lead to misguided assumptions regarding the academic abilities of different groups of students and ill-informed solutions to the problem. Interestingly, these results regarding shadow education in Kenya closely correspond to those reported by Stevenson and Baker (1992) in their study of shadow education in Japan. In Japan, as in Kenya, the educational system is highly competitive, national exams determine educational advancement, and shadow education has flourished. Stevenson and Baker (1992) find that the likelihood of participating in shadow education is greater for males and children living in urban areas. Their analysis also controls for parent's educational background and family financial status. Together, these findings from Kenya and Japan suggest that shadow education plays an important role in determining educational stratification in some societies. Future comparative research on shadow education could enhance our understanding of how advantages are transferred across generations through such informal educational activities. Moreover, failure to examine shadow education as a route to educational mobility in societies where such tutoring and examination preparation strategies are used to enhance formal education will result in an incomplete understanding of the processes by which families seek to increase their children's educational mobility. For example, there are indications that wealthy families in the United States increasingly are purchasing exam preparation courses and private tutors for their children in hopes of boosting their scores on standardized achievement tests such as the SAT and ACT (Schwartz, 1999). These tests are the closest thing to a national examination in the United States and they are regarded carefully by university admission officers. Yet the phenomenon remains relatively unstudied by sociologists of education. Beyond shadow schooling, other forms of family social and cultural capital matter for Kenyan children's educational experiences. Children with many siblings perform less well in school, likely because of a diffusion of social
156
CLAUDIA BUCHMANN
capital across many children. Measures of cultural capital, such as language spoken in the home and mother's reading habits, reduce the likelihood of grade repetition, net of other family background factors. These findings indicate some of the ways that non-economic family resources may be critical to children's school success in the developing world. They also demonstrate that parents with less abundant financial capital at their disposal may be able to use social and cultural resources effectively to provide educational benefits to their children or to counteract financial capital deficits. As Coleman argued, "if the human capital possessed by parents is not complemented by social capital embodied in family relations, it is irrelevant to the child's educational growth" (1988, p. SI 10). This point appears to be as applicable to the case of Kenya as it is to more industrialized countries.
ACKNOWLEDGMENTS
This research was supported by fellowships from the Spencer Foundation and the Fulbright-International Institute of Education Program. In Kenya, the Office of the President granted research clearance and the Bureau of Educational Research at Kenyatta University and the Department of Sociology at the University of Nairobi provided research support for the study. The survey was conducted with the help of John Njoka, John Kamau Gatare, Lucy Maina Wangui, and Pennina Kihika.
NOTES
1. Wong's (1998) study of socialist Czechoslovakia is one important exception. 2. The current study is limited to the focus on social capital within the family, since data on aspects of social capital within the community or between families and schools are not available in the survey of Kenyan households. In Kenya, like many developing countries, demand for education exceeds the state's ability to supply public schools and communities often play a central role in building and managing local schools. Thus, the examination of how community-level social capital is related to educational outcomes in developing countries is a worthy goal for future research. 3. Parental cultural behaviors are probably better measures of cultural capital than students' cultural interests since it is difficult to determine whether students cultural interests are causally prior to their academic performance (De Graaf, 1986). 4. Swahili is the mother tongue of only an estimated 131,000 (0.6 %) of Kenya's 21.5 million people (Grimes, 2000). 5. In one of the few previous empirical analyses of the phenomenon of shadow schooling in a developing society, Hua (1996) found that nearly three-fourths of all grade 8 students in Egypt attend either private lessons or school tutoring. 6. Similarly, the score on the Kenyan Certificate of Secondary Education (KCSE) determines a student's options to pursue post-secondary education.
157
7. Students' performance on national examinations has implications for schools, since each year the government compiles a national ranking of all primary and secondary schools in the country, on the basis of their students' examination results. These rankings are published on the front page of Kenya's national newspaper, the Daily Nation, and are the source of much discussion among parents, students, and the media. 8. In 1995, the time of the survey, 42 Kenyan shillings was equivalent to one U.S. dollar. Income per capita was approximately 370 U.S. Dollars annually. Using these figures, one year of tutoring for one child is roughly equivalent to one-seventh of the average household income. 9. Girls comprise 44% of secondary school students and roughly 30% of university students (Republic of Kenya, 1991b). 10. The amount of assistance parents can provide on homework may depend on their educational background, although the interaction term for parent's education and homework help was not significant (analysis not shown). Other direct measures of parental involvement in children's schooling would be useful. Prior research has found that time parents spend talking to children about school, parents' expectations for their children, and other forms of involvement in educational activities are important forms of family social capital for children's outcomes, but such measures are not available for this study. 11. In other analyses, family structure variables were entered without the interaction term and had no significant effect. For the sake of brevity, they are presented in Table 4 along with the interaction term. 12. In order to explore further the relationship between parent's education and the dependent variables, a dummy variable for parent's education (1= at least some secondary) was included in place of the ordered categorical variable presented in the tables. Results did not differ from those reported in Tables 3-5. 13. There is some debate as to whether social and capital serve to boost the effects of high socioeconomic background or function more as buffer against the detrimental effects of low socioeconomic status (see De Graaf et al., 2000). In other words, the effects of social and cultural resources may vary by socioeconomic status. To examine this possibility, interaction terms for parental education and financial situation with the variables home language, mother's reading habits, and homework help were included in the models presented in Table 4 (analyses not shown). None of these interaction terms were significant, suggesting that the effects of these social and cultural resources apply across the range of socioeconomic backgrounds.
REFERENCES
Astone, N. M., & McLanahan, S. S. (1991). Family Structure, Parental Practices and High School Completion. American Sociological Review, 56, 309-320. Baker, D. P., Akiba, M., LeTendre, G., & Wiseman, A. W. (2000). Worldwide Shadow Education: Outside School Learning, Institutional Quality, and Cross-National Mathematics Achievement. Pennsylvania State University, mimco. Blake, J. (1989). Family Size and Achievement. Los Angeles: University of California Press. Bourdieu, P. (1977). Cultural Reproduction and Social Reproduction. In: J. Karabel & A. H. Halsey (Eds), Power and Ideology in Education (pp. 487-511). New York: Oxford University Press.
158
CLAUDIA BUCHMANN
Bourdieu, P. (1984). Distinction: A Social Critique of the Judgement of Taste. Cambridge: Cambridge University Press. Bray, M. (1999). The Shadow Education System: Private Tutoring and Its Implications for Planners. Paris: UNESCO, International Institute for Education Planning. Buchmann, C. (2000). Family Structure, Parental Perceptions and Child Labor in Kenya: What Factors Determine Who is Enrolled in School? Social Forces, 78, 1349-1379. Buchmann, C , & Hannum, E. (2001). Education and Stratification in Developing Countries: A Review of Theories and Research. Annual Review of Sociology, 27, 77-102. Carbonaro, W. J. (1998). A Little Help from my Friend's Parents: Intergenerational Closure and Educational Outcomes. Sociology of Education, 71, 295-313. Chernichovsky, D. (1985). Socioeconomic and Demographic Aspects of School Enrollment and Attendance in Rural Botswana. Economic Development and Cultural Change, 33, 319-332. Coleman, J. S. (1988). Social Capital in the Creation of Human Capital. American Journal of Sociology, 94, S95-S120. Coleman, J. S., & Hoffer, T. (1987). Public and Private High Schools: The Impact of Communities. New York: Basic Books. Crook, C. J. (1997). Cultural Practices and Socioeconomic Attainment: The Australian Experience. Westport, CT: Greenwood Press. Daily Nation, June 30, 1994. Nairobi, Kenya. Daily Nation, January 4, 1996. Nairobi, Kenya. De Graaf, N. D., De Graaf, P. M., & Kraaykamp, G. (2000). Parental Cultural Capital and Educational Attainment in the Netherlands: A Refinement of the Cultural Capital Perspective. Sociology of Education, 73, 92-111. De Graaf, P. M. (1986). The Impact of Financial and Cultural Resources on Educational Attainment in the Netherlands. Sociology of Education, 59, 237-246. DiMaggio, P. (1982). Cultural Capital and Social Success: The Impact of Status Culture Participation on the Grades of High School Students. American Sociological Review, 47, 189-201. DiMaggio, P., & Mohr, J. (1985). Cultural Capital, Educational Attainment and Marital Selection. American Journal of Sociology, 90, 1231-1261. Downey, D. (1995). When Bigger is Not Better Family Size, Parental Resources, and Children's Educational Performance. American Sociological Review, 60, 746-761. Farkas, G. (1996). Human Capital or Cultural Capital? Ethnicity and Poverty Groups in an Urban School District. New York: Aldine de Gruyter. Fuller, B., & Liang, X. (1999). Which Girls Stay in School? The Influence of Family Economy, Social Demands, and Ethnicity in South Africa. In: C. H. Bledsoe, J. Casterlinc, J. JohnsonKuhn & J. Haaga (Eds), Critical Perspectives on Schooling and Fertility in the Developing World (pp. 181-215). Washington D.C.: National Academy Press. Gomes, M. (1984). Family Size and Educational Attainment in Kenya. Population and Development Review, 10, 647-660. Grimes, B. (2000). Ethnologue: Languages of the World. SIL International. Hua, H. (1996). Which Students are Likely to Participate in Private Lessons and School Tutoring in Egypt? Unpublished doctoral dissertation, Havard University, Graduate School of Education. Hughes, R. (1991). Examining the Roots of Educational Demand: The Case Supporting Rural Agrarian Development. World Development, 19, 213-223. Katsillis, J., & Rubinson, R. (1990). Cultural Capital, Student Achievement and Educational Reproduction in Greece. American Sociological Review, 55, 270-279.
159
Lareau, A., & Horvat, E. M. (1999). Moments of Social Inclusion and Exclusion: Race, Class, and Cultural Capital in Family-School Relationships. Sociology of Education, 72, 37-53. Lipman, E. L., & Offord, D. R. (1997). Psychosocial Morbidity Among Poor Children in Ontario. In: G. J. Duncan & J. Brooks-Gunn (Eds), Consequences of Growing Up Poor (pp. 239-287). New York: Russell Sage Foundation. Lloyd, C. B., & Blanc, A. K. (1996). Children's Schooling in Sub-Saharan Africa: The Role of Fathers, Mothers and Others. Population and Development Review, 22, 265-298. Lloyd, C. B., & Gage-Brandon, A. J. (1994). High Fertility and Children's Schooling in Ghana: Sex Differences in Parental Contributions and Educational Outcomes. Population Studies, 48, 293-306. Long, J. S. (1997). Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage. Montgomery, M. R., & Lloyd, C. B. (1997). Excess Fertility, Unintended Births and Children's Schooling. Policy Research Division Working Paper, No. 100. New York: Population Council. National Council for Population and Development and Central Bureau of Statistics (1994). Kenya Demographic and Health Survey 1993. Macro International. Republic of Kenya (1991a). Rural Labor Force Survey. Nairobi: Government Printer. Republic of Kenya (1991b). Employment and Earnings in the Modem Sector. Nairobi: Government Printer. Republic of Kenya (1993). Economic Survey. Nairobi: Government Printer. Republic of Kenya (1994). Kenya Population Census, 1989. Volume 1. Nairobi: Government Printer. Schwartz, T. (1999). The Test Under Stress. New York Times Magazine, January 10, 30-35. Stevenson, D. L, & Baker, D. P. (1992). Shadow Education and Allocation in Formal Schooling: Transition to University in Japan. American Journal of Sociology, 97, 1639-1657. Teachman, J. D Paasch, K., & Carver, K. (1997). Social Capital and the Generation of Human Capital. Social Forces, 75, 1343-1359. Wong, R. S. (1998). Multidimensional Influences of Family Environment in Education: The Case of Socialist Czechoslovakia. Sociology of Education, 71, 1-22.
Commentary: CULTURAL AND SOCIAL CAPITAL IN EDUCATIONAL RESEARCH

Commentary on the Buchmann paper.
Raymond Sin-Kwok Wong

INTRODUCTION
The past ten years have seen a surge of research on how non-standard forms of capital affect various sociological phenomena, including educational inequalities, labor market outcomes, and voluntary association in community (e.g. Bian, 1997; Bian & Ang, 1997; Granovetter, 1973, 1985; Putnam, 1993, 2000; Wong, 1998). In studying whether and how two less tangible forms of parental capital - cultural and social capital - may play an important role in structuring educational inequalities in a developing society, Claudia Buchmann's article offers an important extension of current research. While the relevance of these two forms of capital (especially social capital) is still open to debate (Morgan & S0rensen, 1999; Carbonaro, 1999; Hallinan & Kubitschek, 1999), they have been incorporated into a unified neo-capital theory that combines various theoretical traditions such as Marxist class theory, neo-classical microeconomic theory, social network analysis, and social reproduction theory (Lin, 2000). In sociology of education, the rising interest in cultural and social capital reflects a desire to go beyond conventional measures of socioeconomic background to gain a better understanding of the quality of social environments that parents provide to their offsprings, as past research shows a weak relation between conventional indicators and the quality of family
161
162
RAYMOND SIN-KWOK WONG
environments (Bradley & Caldwell, 1984). The weak relationship suggests that besides tangible material resources, the family provides more elusive forms of resources that can foster cultural and social skills, motivations, and commitment to learning. To explore and identify such non-material resources and their roles in the reproduction of educational inequality, the concepts of cultural and social capital have proven to be quite useful. Yet, as Buchmann points out, most research has focused on western, industrialized societies. Exploring the significance of social and cultural capital and school performance in a developing African society, Buchmann's study makes a notable contribution. Its extension of the concepts to a new and different context helps us to assess not only their general usefulness, but also whether and how the measures of cultural and social capital are sensitive to local economic and cultural circumstances and may need to be adjusted accordingly.
CULTURAL CAPITAL
Buchmann argues that cultural and social capital are useful concepts for understanding educational outcomes in a non-western developing society such as Kenya, but their operationalization and situated indicators must be modified to reflect the local context. The biggest modification is in the measurement of cultural capital. Considering Kenya's non-industrial history, Buchmann does not think it is appropriate to follow the conventional definition of cultural capital as a set of institutionalized high status cultural signals (attitudes, preferences, formal knowledge, behavior, goods and credentials) used to demonstrate cultural inclusion or exclusion. With Kenya's concentration of poor families, she contends, a focus on the reading habits and linguistic skills of parents (or other family members), rather than highbrow cultural activities as constituents of cultural capital, is more fitting. This extension of the conceptualization of cultural capital is certainly in line with the theoretical arguments advanced by Bourdieu (1977) and Bourdieu and Passeron (1977). The central tenet of Bourdieu's argument is that apparently neutral academic standards are in reality laden with specific cultural values and skills, typically held by the dominant classes, which children acquire at home through parent-led cultural activities. In other words, cultural capital is variably embedded in the differing quality of home environments and the time that parents spend with their children. Family lifestyles and consumption patterns constitute a critical source of children's cultural formation, which in turn represent valuable educational resources that aid children's learning motivation and performance inside schools. The classic examples of such capital include what
Commentary: Cultural and Social Capital
163
Buchmann calls highbrow activities like regular visits to theaters, concerts, galleries, and libraries, or access to books and classical music. It is entirely reasonable to expand this list to include other cultural or takenfor-granted activities that contribute to academic success and which are differentially distributed in society. Especially for societies under different socioeconomic and cultural conditions, this wider list would include parental reading habits and linguistic skills, as Buchmann argues in her study of schools, children and parents. To ignore highbrow cultural activities altogether is, however, a different matter. Whether and the extent to which these activities are truly irrelevant in Kenya and other developing countries is debatable and better addressed empirically. For the case of Kenya, a former British colony, it seems rather counterintuitive to dismiss the continuing significance of highbrow western cultural practices as status symbols for upper- and middle-class Kenyans. Also, given the multiplicity of spoken dialects and the small percentage (less than one percent) of Kenyans whose mother tongue is Swahili, any cultural activities involving either English or Swahili would likely constitute cultural capital that gets rewarded in the school system. In short, the measure of cultural capital Buchmann used could be refined without contradicting either her argument or the conventional understanding of cultural capital. Such a refinement would enable us to further differentiate the strength and quality of cultural capital, not just its presence. This would yield more information about the effects of cultural capital on children's school achievement. In her adjustment of the measure of cultural capital to the specific social context of Kenya, Buchmann also includes shadow education as one influential host of cultural capital that is often neglected in the study of academic achievement. Regrettably, her effort to discuss it in terms of cultural capital is sketchy and tentative, and its theoretical locus in the study becomes ambiguous. On the one hand, why shadow education is an aspect of cultural capital remains unclear. On the other hand, there are signs from the results that shadow education is instead more closely related to social capital. Shadow education is measured as a dichotomous outcome, with a value of one if the child participates in a range of informal educational activities that include private tutoring at home or school and classes after school and on weekends, or zero otherwise. In addition to studying the effect of shadow education on educational outcomes, Buchmann also analyzes the determinants of its participation. Unfortunately, the explanatory power of the models (her Table 3) is weak. But the finding suggests that participation in shadow education is almost exclusively an urban phenomenon. Indeed, once urban residence is controlled in model 4, the importance of parental education and socioeconomic-status disappears and none of the other variables are statistically significant at the
164
0.05 level (two-tailed test). Given these signs of shadow education being limited largely to urban areas, it would be illuminating to explore whether participation in shadow education stems more from availability, rather than representing an exogenous form of cultural capital. On this question of availability, it is important to differentiate between private shadow education at home and after-school programs, which are treated as the same in Buchmann's view of shadow education as cultural capital. Because of the monetary costs, the former should be tied more closely to parental socioeconomic background. For the latter, the issue involves whether the after-school programs are more readily available in certain types of schools, whether they are offered to all or only selected students, and how students get sorted into different schools in the first place. In clarifying whether and how participation in shadow education is determined by availability or other selective mechanisms, these questions enable a better assessment of the exact role that shadow education may play in educational stratification, and the precise mechanisms at play. To explore these questions, however, would mean reorienting the conceptualization of shadow education from cultural capital to social capital.
SOCIAL CAPITAL
In sharp distinction from cultural capital, social capital is derived from social relations between individuals rather than particular values, attitudes, or knowledge that individuals acquire. According to Bourdieu (1983, p. 248), 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, membership in a group." For social capital to be operational, however, the existence of network ties is not sufficient. The quality of the relationships matters as well; only relationships of a particular type - trusting and positive - will work as social capital (Paxton, 1999, p. 92). So, if parents do not utilize their social relations effectively, the nominal existence of social ties may be irrelevant to their children's stratification outcomes. More than a mediating factor between socioeconomic background and unequal outcomes, social capital facilitates and enhances the conversion of other forms of capital into children's long term human capital. As Coleman (1988) notes: "If human capital possessed by parents is not complemented by social capital embodied in family relations, it is irrelevant to the child's educational growth" (Coleman, 1988, SI 10). That is, unless parents use their economic and human capital effectively as resources in their parental roles, the mere presence of such capital will have little effect on their children's attainment.
165
An under-specified mobility step is required. Conversely, with effective utilization of social capital, parents having modest or little economic and human capital may still be able to positively influence their children's educational outcomes and social behavior, for instance, by getting them into shadow education. Needless to say, not all social relations and structures are important resources for educational success. It is the schooling-focused relationships that count, and these include not only relations between children and parents, but also relations inside community organizations that are useful for the child's social or cognitive development (Coleman, 1990, p. 300). Buchmann's operationalization of social capital largely follows the leads of earlier works (Lichter, Corn well & Eggeben, 1993; Parcel & Menaghan, 1993, 1994), focusing on social capital within the family. Her measure of family social capital includes four components: (1) number of children; (2) number of adults; (3) proportion of children older than the target child; and (4) whether the child received homework help from a family member. Except for the last component, this measure simply indexes the family structure and contains little information about the strength or quality of parent-child relationships. Given the theoretical importance of the quality of relationships noted above, it is not entirely surprising to find that the various components of this measure of social capital have rather weak effects on schooling success. If the study includes other measures that index directly the nature of the relationship between parents (and other adults in the household) and the child, for example, the degree of parental involvement in the child's academic and extra-curricular activities, the effects may become clearer. More importantly, this measure of social capital neglects relationships embedded in the community and through which parents can mobilize their organizational resources to enhance their child's socialization and facilitate educational opportunities. Past research has found that parental social networks and their involvements with teachers, students, and other parents have important implications for educational outcomes in the United States (Farkas, Grobe, Sheehan & Shuan, 1990; Stanton-Salazar & Dornbusch, 1995). The same is true of parental membership in the Communist party in several countries under socialism (Wong, 2001b). With the majority of Kenyans still living in rural areas and/or closely-knit communities, the types of kinship ties within the extended family and networks with other families in the community could be important determinants of social capital, outside the family's relationships with other social institutions. From a comparative perspective, then, further consideration of social capital that is sustained outside the family, but in forms that parents can utilize to enhance their children's academic success, could enrich Buchmann's study and our understanding of how social capital works in different settings.
166
COMMON ISSUES
The omission of social linkages outside the family from the measures of social capital is rather common. Besides this problem of operationalization, the concept of social capital itself has been criticized. Some have argued that its definition is ambiguous and imprecise, and its use in empirical research is sometimes based on circular and tautological reasoning, with the concept being defined in terms of its effects rather than its a priori characteristics (Durlauf, 1999; Portes & Landolt, 1996). Critics also have called attention to the negative influences of some social capital (such as kids who join urban gangs) on school achievement - underlining the tautological nature of only defining social capital in terms of its functionality in enhancing academic performance. This suggests a need for greater theoretical clarification and careful operationalization before social capital can be convincingly identified as a bundle of determinants of school achievement. The same is true of cultural capital. In particular, a number of empirical issues need to be addressed. First, it must be stressed that neither cultural nor social capital is simply possessed; both quality and quantity are significant. To adequately register the gradation in quality and quantity, a multiple indicators approach to index its variation is needed. Such an approach has the additional advantage of allowing researchers to study the measurement issue directly, a task so far largely ignored by sociologists (see Paxton, 1999 for an exception). Researchers often use a single proxy or cluster of variables to represent the concepts, then study their relationships with other variables without making correction for attenuation due to measurement error. The construction of latent variables of cultural and social capital from multiple indicators would provide a better representation of the underlying theoretical constructs and the true structural relationships between constructs. Second, if cultural and social capital theoretically play more than a mediating role between socioeconomic factors and stratification outcomes, their nonmediating role has to be demonstrated empirically. So far, there have been few systematic attempts to study the relationships among the various forms of capital and whether they each exert an independent influence on schooling outcomes (Wong, 1998 is an exception). Furthermore, there is no empirical evidence that the putative effects of cultural and social capital are not due to some unobserved family characteristics rooted in socioeconomic background. These concerns can be systematically addressed by integrating sibling resemblance research into the study of cultural and social capital. The combined approach will provide a better construction of the unobserved common family environment, and enable an
167
investigation of the relative influence of various forms of parental capital on children's educational attainment. The discussion of social capital and its effects on stratification must also register the possibility of changes in effects over time. Social relationships that provide positive resources in the past may cease to be beneficial because of macrostructural changes. This can be illustrated by findings from my study of business firms and business entrepreneurs in Hong Kong (Wong, 2001a). It has long been accepted that extensive familial and kinship ties constitute valuable social capital for Chinese entrepreneurs, as Chinese firms are basically family organizations that rely on family members or relatives to run the business. My study, however, finds that this is no longer the case. By the late 1990s, less than one-third of Hong Kongfirmsemploy any family members or relatives. Perhaps more importantly, the studyfindsthat business entrepreneurs hold more negative than positive views toward the practice of hiring kin members. This suggests that with changes in social relations linked to westernization, what used to be valuable forms of social capital for economic activities lose their force in the new environment. Such temporal changes in the effects of particular social relations on stratification outcomes serve to underline the importance of limiting the identification of social capital to specific social contexts, a point that Buchmann rightly emphasizes.
CONCLUSION
In sum, Claudia Buchmann's study reminds us of the challenges facing educational researchers as they seek to understand the factors and processes that shape and reinforce the unequal distribution of opportunities, and how families may actively mediate and offer resources for their children's schooling experiences. Her findings on the significance of social and cultural capital in education achievement show, once again, that conventional measures of socioeconomic background are insufficient proxies for the types of resources that families can utilize, and an exclusive focus on these measures would elude the importance of non-material resources as bases of inequality. Her study demonstrates the importance of adjusting our conceptual tools to local conditions, as the structures and processes of inequality are always conditioned by specific economic and cultural contexts. To gain new insights into the function and significance of cultural and social capital in educational stratification, then, research within widely differing societies would be extremely valuable. Like any new and evolving idea, cultural and social capital need further refinement in conceptualization and development of rigorous measurement. Notably, the relationship between cultural and social
168
RAYMOND SIN-KWOK W O N G
capital should be explored, a point to which the ambiguous status of shadow education in Buchmann's study attests. After all, social capital is culturallybased and cultural capital is socially-constructed. With such outstanding issues, the long-run strength of these concepts will depend on our ability to identify the conditions under which they are forceful and our efforts to compare their influence over time and across various contexts.
REFERENCES
Bian, Y. (1997). Bringing Strong Ties Back in: Indirect Connection, Bridges, and Job Search in China. American Sociological Review, 62, 366-385. Bian, Y., & Ang, S. (1997). Guanxi Networks and Job Mobility in China and Singapore. Social Forces, 75, 981-1006. Bourdieu, P. (1977). Cultural Reproduction and Social Reproduction. In: J. Karabel & A. H. Halsey (Eds), Power and Ideology in Education (pp. 487-511). New York: Oxford University Press. Bourdieu, P. (1983). Forms of Capital. In: J. G. Richardson (Ed.), Handbook of Theory and Research for the Sociology of Education (pp. 241-258). New York: Greenwood Press. Bourdieu, P., & Passerson, J.-C. (1977). Reproduction in Education, Society, and Culture. Beverly Hills, CA: Sage. Bradley, R. H., & Caldwell, B. M. (1984). The HOME Inventory and Family Demographics. Developmental Psychology, 20, 315-320. Carbonaro, W. J. (1999). Opening the Debate: On Closure and Schooling Outcomes. American Sociological Review, 64, 682-686. Coleman, J. S. (1988). Social Capital in the Creation of Human Capital. American Journal of Sociology, 94, S95-S120. Coleman, J. S. (1990). Foundations of Social Theory. Cambridge, MA: Bellknap Press. Durlauf, S. N. (1999). The Case Against Social Capital. Focus, 20(3), 1-5. Farkas, G., Grobe, R. P., Sheehan, D., & Shuan, Y. (1990). Cultural Resources and School Success: Gender, Ethnicity, and Poverty Groups within an Urban District. American Sociological Review, 55, 127-142. Granovetter, M. (1973). The Strength of Weak Ties. American Journal of Sociology, 78, 1360-1380. Granovetter, M. (1985). Economic Action and Social Structure: The Problem of Embcddedness. American Journal of Sociology, 91, 481-510. Hallinan, M. T., & Kubitschek, W. N. (1999). Conceptualizing and Measuring School Social Networks. American Sociological Review, 64, 687-693. Lichter, D. T., Cornwell, G. T., & Eggeben, D. J. (1993). Harvesting Human Capital: Family Structure and Education Among Rural South. Rural Sociology, 58, 53-75. Lin, N. (2000). Social Capital: A Theory of Structure and Action. Cambridge: Cambridge University Press. Morgan, S. L., & S0rensen, A. B. (1999). Parental Networks, Social Closure, and Mathematical Learning: A Test of Coleman's Social Capital Explanation of School Effects. American Sociological Review, 64, 661-681. Parcel, T. L., & Menaghan, E. G. (1993). Family Social Capital and Children's Behavior Problems. Social Psychology Quarterly, 56, 120-135.
Commentary:
Cultural and Social
Capital
169
Parcel, T. L., & Menaghan, E. G. (1994). Early Parental Work, Family Social Capital, and Early Childhood Outcomes. American Journal of Sociology, 99, 972-1009. Paxton, P. (1999). Is Social Capital Declining in the United States? A Multiple Indicator Assessment. American Journal of Sociology, 105, 88-127. Portes, A., & Landolt, P. (1996). Unsolved Mysteries: The Tocqueville Files II. 77/e American Prospect, 26(May-June), 18-21. Putnam, R. D. (1993). Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press. Putnam, R. D. (2000). Bowling Alone: The Collapse and Revival of American Community. New York: Simon and Schuster. Stanton-Salazar, R. D., & Dornbusch, S. M. (1995). Social Capital and the Reproduction of Inequality: Information Networks Among Mexican-Origin High School Students. Sociology of Education, 68, 116-135. Wong, R. S. (1998). Multidimensional Influences of Family Environment in Education: The Case of Socialist Czechoslovakia. Sociology of Education, 71, 1-22. Wong, R. S. (2001a). Made in Hong Kong: A Case Study of Chinese Business Organizations and Business Culture. Paper presented at the International Conference on Ethnic Chinese Business and Culture in Global and Local Contexts. February 15-17,2001 in Taipei, Taiwan. Wong, R. S. (2001b). Egalitarianism Versus Social Reproduction: Changes in Educational Stratification in Five Eastern European Countries. Paper presented at the Research Committee on Social Stratification and Mobility (RC28). Mannheim, Germany, April 26-28.
CONCLUSIONS: CROSS-CULTURAL VIEWS OF SOCIAL CAPITAL, INSTITUTIONS, AND STRATIFICATION

Emily Hannum and Bruce Fuller
INTRODUCTION
Sociological interest in institutions and networks and the information, values and attitudes they afford long predates the development of social capital as a concept. However, an important contribution of social capital theory has been to bring concern with these issues to researchers working in a status attainment framework of individual actors and to the policy community (Sandefur & Laumann, 1998, p. 496; Lin, 1999; Warner, 1999; Schaub & Baker, this volume). In these contexts, the expanding breadth of the concept and the lack of clarity about distinctions between social capital, its sources, and the benefits it confers remain significant concerns (see Portes, 1998 and commentaries by Wong, Schaub and Baker, and Fernandez Kelly in this volume). In current usage, social capital is indeed broad, spanning "structural" elements, including social organization, roles, rules, and networks, and "cognitive" elements, including information, norms, values and attitudes (Warner, 1999, p. 375).
171
172
EMILY HANNUM AND BRUCE FULLER
The empirical papers in this volume illuminate the impact of social capital, in a diverse set of guises, early in the status attainment process. Collectively, the papers trace the impact of educational organization, classroom processes, and family attitudes and practices on students' educational achievement and attainment. Three of the papers - Buchmann, Bankston and Zhou, and Goyette and Conchas - draw directly on Coleman's (1988) concept of social capital, investigating how it functions across cultural settings in families, schools, or other types of social institutions. The remaining two papers - Shavit et al. and Kao - contribute cross-cultural perspectives on the impact of organizational features of the education system and parental attitudes and practices on children's schooling outcomes. In their alternative framing of concepts often vaguely or explicitly subsumed within the social capital framework, these two papers demonstrate an important challenge to social capital theorists: to better specify the distinctions from, and links to, older, overlapping frameworks in the sociology of education. Definitional debates aside, the papers in this volume affirm the role of social institutions and networks in conditioning children's educational progress. The papers suggest that significant advancements can be made in understanding processes of social stratification by identifying elements of social capital that support children's education, and the institutional structures and less formal contexts in which social capital can flourish. At the same time, the papers pose challenges to this task. They illustrate the full complexity of differentiating social capital explanations for schooling outcomes from structural, economic, or cultural explanations. Further, they stress that the social supports that matter for education vary across societies and cultures, and even across socio-cultural groups within a given educational setting. In this concluding chapter, we briefly review key results from each paper, stressing those that link to the broader themes of the volume. We discuss first the three papers that emphasize family practices as they relate to children's schooling, namely Buchmann, Kao, and Bankston and Zhou, then consider the two papers that emphasize school structures and practices: Shavit et al. and Goyette and Conchas. We close with a commentary on fruitful directions for the development of social capital theory suggested by these international and cross-cultural perspectives.
SOCIAL CAPITAL, INSTITUTIONS AND SCHOOLING: KEY RESULTS

The first paper, by Claudia Buchmann, emphasizes family economic, social and cultural resources and how these translate into advantages and disadvantages
Conclusions: Cross-Cultural Views of Social Capital
173
for students in Kenya. Buchmann's study, based on a survey conducted in 1995, illustrates some of the ways that non-economic family resources may be critical to children's school success in the developing world. Buchmann starts with Coleman's (1988) definition of family social capital as relations between children and parents or other adults in the household. This definition includes "structure" and "process" elements: the physical presence of adults in the household and the attention that they give to the child. Buchmann employs Teachman, Paasch and Carver's (1997, p. 1344) specification of "process": social capital is generated through the "density and consistency of educationally-focused relationships that exist among parents, children and schools". She conceptualizes cultural capital as cognitive and linguistic skills rather than competence in high status culture, with the justification that for "low-status or poor populations, a conceptualization of cultural capital that focuses on the reading habits and linguistics skills of parents or other family members may be more relevant" (p. 138). Buchmann finds that children with many siblings perform less well in school, and that language spoken in the home and mother's reading habits reduce the likelihood of grade repetition, net of other family background factors. Buchmann also shows that children who participate in shadow schooling enjoy significantly greater achievement and reduced repetition rates. Her critical discussion of family social and cultural capital in an impoverished developing setting contributes an important new perspective to a field dominated by research in western, industrialized settings. Buchmann's attention to shadow schooling in this setting is similarly novel, as much of the research in this field has occurred in highly industrialized East Asian nations where educational gaps between advantaged and disadvantaged children are relatively narrow. Her results suggest the need to consider this institution in other less developed settings where educational inequalities are extreme. Buchmann argues convincingly for the significance of shadow education as a contributor to educational stratification in a global context, given the wellestablished market for private tutoring in East Asia and the rising market for corporate private tutoring in the U.S. and elsewhere. Like Buchmann, Kao focuses on family attitudes and practices that support children's education. Capitalizing on longitudinal data from the National Educational Longitudinal Survey in the United States, she examines minority parents' aspirations and savings behavior and their implications for minority students' subsequent aspirations. She finds that minority parents have extremely high aspirations for their children and that these aspirations exert considerable influence over youth aspirations and their maintenance over time. Of close interest to the larger theme of this volume, she also finds that Asian
174
American parents are better able to support their children's aspirations financially because their savings patterns are more aggressive than other minority parents' are. Kao grounds her work in a status attainment tradition rather than a social capital framework. In its non-social capital framing, her analysis illustrates a common critique of social capital theory: that it "repackages" a number of elements of more established theoretical frameworks. For example, the aspirations of parents and of children, long staple elements of the status attainment tradition, could also be viewed as indicators of Warner's (1999) "cognitive" elements of social capital or the "educationally-focused relationships" among parents and children described in the Buchmann paper. Conceptually, though not empirically, a link can also be made between social capital theory and the savings behavior results. It is likely that information about college costs, or group norms and expectations about education, differ across the ethnic groups in ways that favor Asian Americans. Both the informational and norm-setting functions of social networks are often cited as important elements of social capital. Kao's results thus contribute two insights to the theme of this volume: first, parents' attitudes and ways of supporting education are important mechanisms for the development of children's human capital net of parental human capital and material resources. Second, the specific supports employed differ significantly across socio-cultural groups within the U.S. Bankston and Zhou drive home both the importance of social capital and its different functioning across social and cultural groups. They test a specific element of Coleman's definition of social capital: intergenerational closure. Closure refers to the idea that parents who maintain close supervision over their own children through connections to the adults present in the lives of their children are able to impose beneficial, consistent norms and standards. Bankston and Zhou argue that the closure aspect of social capital does not apply well for immigrant families in the United States, since migration disrupts family structure. Analyzing data from the National Longitudinal Study of Adolescent Health, Bankston and Zhou focus on involvement in religious organizations and the impact that this involvement has on grades. They conclude that grades of immigrant children can be explained by involvement in ethnic religious organizations rather than by parental involvement in social networks. This paper makes an important contribution in focusing attention on religious participation, often ignored in status attainment research and in sociology of education, as a social resource that supports children's educational progress. Most significant in the context of this volume's themes, Bankston and Zhou, like Buchmann, investigate family supports for schooling that are made possible by the presence of institutions external to the family: churches in the first case,
175
and shadow education as an organizational feature of Kenyan education in the second. The final two empirical papers in this volume focus directly on institutions and the social processes within them. Adopting a systemic perspective, Shavit et al. trace the consequences of educational "second chance" structures for students' eventual educational and occupational outcomes. The paper capitalizes on unusual longitudinal data from merged Israeli census files that allow tracking of stratification outcomes for the nation. The paper concludes that educational inequalities between socio-economic groups, between men and women, and between ethnic groups were reduced by the presence of second chance structures. Occupational outcomes for those taking advantage of second chance opportunities were not as good as for those opting for what the authors call "the main road" (p. Ill), but outcomes were better than those for dropouts. While acknowledging that larger occupational stratification systems are maintained, the authors conclude that second chance structures provide an important service to disadvantaged groups: protection from downward mobility. Shavit and his colleagues make the essential point that educational organization is an important backdrop to individual educational outcomes, and to social stratification. In recent years, a significant policy debate about the relationship between investments in schools and student outcomes has emerged, stimulating numerous empirical studies in the field of economics (Betts, 1996; Burtless, 1996; Card & Krueger, 1992; Hanushek, 1995; Kremer, 1995). Few contributors to this debate have considered the more sociological question of whether the organization of schooling matters for student outcomes. By quantifying the consequences of educational organization for both the economic outcomes of schooling and social disparities in these outcomes, Shavit and his colleagues demonstrate the unique perspective that sociologists can contribute to this important policy debate. The question of how educational organizations create or ameliorate stratification is the point at which social capital theory connects to large-scale stratification processes. Sociologists have long argued the importance of organizationally defined activities of school personnel in relation to the sponsorship and advancement of students (Cicourel & Kitsuse, 1977). In the case of Israel, second chance structures presumably afford socio-economic protection by providing new skills, but also by reconnecting individuals to school-based networks and sponsorship - the weak ties that Ferndndez Kelly emphasizes as critical for educational and economic success. Goyette and Conchas explore the nature of those ties in their paper. Specifically, Goyette and Conchas consider the consequences of social relationships in schools for study habits of Vietnamese- and Mexican-American students. In an innovative, mixed-method design, they complement statistical
176
analysis of data from the National Educational Longitudinal Survey with analyses of qualitative interview transcripts from two years of ethnographic fieldwork in a California high school. Goyette and Conchas show that teachers and schools construct meanings about ethnicity that translate to unequal access to the school resources that aid in achievement and attainment. Social relationships in schools, specifically positive peer associations and supportive relationships with teachers, explain a large portion of the difference in time spent studying between Vietnamese- and Mexican-American students. The Goyette and Conchas paper is significant because it shows that student behaviors - often used to explain cross-group differences in other achievement or attainment outcomes - are themselves a function of school experience. Further, Goyette and Conchas make the important point that differences in social relations in the classroom translate to an uneven quality of school experience even for students with access to the same economic resources and human capital, in the same schools and classrooms. The question that links the Shavit et al. and Goyette and Conchas papers and ties both to the broader theme of this volume is, are social networks and relations in schools an important link between institutional structures and educational stratification? The full set of relationships linking educational organization, the nature of social relations that emerge within schools, and social inequality, remains empirically unexplored.
FUTURE DIRECTIONS FOR SOCIAL CAPITAL

In the introduction to this chapter, we referred to debates that have emerged about the definition and usefulness of social capital as a concept. The studies showcased in this volume affirm the importance of various dimensions of social capital - whether named as such or not - in educational achievement and educational stratification. At the same time, collectively, the papers offer several insights that are relevant to a fundamental criticism of social capital: its blurred and expanding definition. They also suggest fruitful directions for research that will further develop the concept of social capital as it relates to children's schooling. We close by offering thoughts on the implications of the crosscultural studies contained in this volume for social capital theory. First, the papers illuminate an important critique of research in the social capital framework: that it has failed to establish boundaries that clearly demarcate social resources from alternative notions of material and cultural resources. The papers in this volume offer concrete examples of the intensive linkages between social, cultural and economic resources that complicate development of clean boundaries. Kao's paper illustrates different savings practices across
177
ethnic groups with similar socio-economic backgrounds, suggesting an important mechanism by which differences in family attitudes toward education translate eventually into different economic resources to support schooling. Conversely, Buchmann demonstrates that purchased tutoring serves as a sort of "marketized" cultural capital - an advantage designed specifically to help children connect to learning in the formal education sector that is available only to families withfinancialresources. Wong's commentary on Buchmann further complicates the interpretation of shadow schooling, arguing that it could be conceptualized as a form of social capital. The intertwining nature of cultural, social and economic resources illustrated in these examples demonstrates that the apparent nature of the resource - social, economic, or cultural - depends to some degree on the timing of data collection and the predilection of the researcher. As Wong observes, social capital is culturally-based and cultural capital is socially-constructed (this volume, p. 168); both are often tied to family economic resources that may or may not be fully captured in standard measures of socio-economic status. Convincing empirical contributions to the further development of social capital theory will require each of these kinds of resources - social, economic, and cultural - to be considered together with the others. A similar case can be made for considering the relationship between social capital and older theoretical traditions emphasizing social institutions and organizations. Indeed, the intertwining of social structures and social networks was recognized in both of the major "founding stories" of social capital theory, in Bourdieu's notion of social capital as "more or less institutionalized relationships" and Coleman's notion of social capital as social structures that facilitate certain types of action for the individuals within those structures (Schaub & Baker, this volume, p. 127). In education, as Fernandez Kelly forcefully argues, the "weak ties" that aid success are entwined with the extra-familial institutions to which children have access (this volume, pp. 73-83). Taking examples from this volume, Goyette and Conchas emphasize social relationships inside schools, Bankston and Zhou, religious participation, and Buchmann, participation in after school tutoring, as important factors in children's academic performance. While good relationships with teachers, religious participation, and participation in tutoring can be seen as manifestations of the social or cultural capital of students, each is also highly contingent upon the presence of certain types of social institutions. On the other side, Shavit et al. attest to the significant benefits alternative educational structures confer on disadvantaged youth. Here, the social capital engendered by alternative educational institutions is likely to play a role. While the papers taken collectively point to the importance of considering linkages between social institutions and social capital, they are able to provide
178
little empirical evidence on the nature of this relationship. Theoretically speaking, further development of social capital as a framework requires greater conceptual clarity about how educational (and other) institutions fit into the ever-expanding definition of the concept. Specifically, are institutions themselves an element of social capital, a source of social capital, or a competing explanation for educational advantages or disadvantages conferred by social capital? To begin to address this question, empirical studies of schooling grounded in the social capital framework need to adopt measurement strategies that provide better descriptions of the institutions of interest and distinguish access from participation. Investigating social relationships and networks without fully considering the social institutions that enable those relationships, along with other benefits, renders causal interpretations difficult. Full attention to the structural underpinnings of social capital in educational organizations would also increase the practical relevance of social capital research, as organizational structures, much more than networks or relationships, are subject to policy intervention. Finally, the most important insight to arise from these papers derives directly from their cross-cultural nature. The papers bear out hypotheses posited in recent reviews of social capital and social network theory that unequal access to social capital across groups plays a significant role in the process of social stratification (Hofferth, Boisjoly & Duncan, 1999; Lin, 1999). Collectively, the papers also suggest a complicating refinement to this hypothesis: not only do different groups have different levels of access to social capital, they employ different kinds of social capital. Except at a broad level, the factors that constitute social capital vary across social and cultural groups and across societies. This insight offers a caveat to the task of developing a standard, theoretically defensible definition of social capital for use in sociology of education. Standard definitions are needed for theory building and for illuminating crossgroup disparities in access to the same social supports for schooling. Yet, the danger suggested by this volume is that use of a narrow, standard definition, true as it might be to the theoretical origins of the concept, risks producing seriously misleading results in research that crosses socio-cultural lines. This volume challenges researchers to recognize the culturally-circumscribed nature of the social capital in children's lives, and to acknowledge the complexity this fact presents for the task of developing reasonable definitions, particularly operational definitions. Lareau and Horvat (1999) have argued that operational definitions of social capital should be based on how these resources are valued in the specific context under investigation - such as the school experience (p. 42). The papers in this volume take this argument a step further: operational
Conclusions:
Cross-Cultural
Views of Social
Capital
179
definitions need to recognize that what is valued in the school experience varies across societies and cultures, and even across socio-cultural groups within given educational settings. Further development of the social capital framework requires the development of operational definitions that are sensitive to - and encompass - the varying forms of social capital that serve different sociocultural groups.
REFERENCES
Betts, J. (1996). Does School Quality Matter? Evidence from the National Longitudinal Survey of Youth. The Review of Economics and Statistics, 77(2), 231-250. Burtless, G. (Ed.) (1996). Does Money Matter?: The Effect of School Resources on Student Achievement and Adult Success. Washington, D.C.: Brookings Institution Press. Card, D., & Krueger, A. (1992). Does School Quality Matter? Returns to Education and the Characteristics of Public Schools in the United States. Journal of Political Economy, 100, 1-40. Cicourel, A., & Kitsuse, J. (1977). The School as a Mechanism of Social Differentiation. In: J. Karabel & A. Halsey (Eds), Power and Ideology in Education (pp. 283-292). New York: Oxford University Press. Coleman, J. (1988). Social Capital in the Creation of Human Capital. Supplement to American Journal of Sociology, 94, S95-120. Hanushek, E. (1995). Interpreting Recent Research on Schooling in Developing Countries. World Bank Research Observer, 10, 247-254. Hofferth, S., Boisjoly, J., & Duncan, G. (1999). The Development of Social Capital. Rationality and Society, 77(1), 78-110. Kremer, M. (1995). Research on Schooling: What We Know and What We Don't (a Comment on Hanushek). World Bank Research Observer, 10, 247-254. Lareau, A., & Horvat, E. (1999). Moments of Social Inclusion and Exclusion: Race, Class, and Cultural Capital in Family-School Relationships. Sociology of Education, 72, 37-53. Lin, N. (1999). Social Networks and Status Attainment. Annual Review of Sociology, 25, 467-487. Portes, A. (1998). Social Capital: Its Origins and Applications in Modern Sociology. Annual Review of Sociology, 24, 1-24. Sandefur, R., & Laumann, E. (1998). A Paradigm for Social Capital. Rationality and Society, 70(4), 481-501. Teachman, J., Paasch, K., & Carver, K. (1997). Social Capital and the Generation of Human Capital. Social Forces, 75, 1343-1359. Warner, M. (1999). Social Capital Construction and the Role of the Local State. Rural Sociology, 64(3), 373-393.
ABOUT THE AUTHORS
Bruce Fuller is a professor of education and public policy at the University of California, Berkeley. His current work focuses on the dilemmas around the de-centering of public aims and institutions within the worlds of child care, family welfare, and school reform. Prior to becoming a full-time teacher and scholar, he worked for a state legislature, a governor, then as a sociologist at the World Bank. He taught at the Harvard graduate school of education before returning home to the San Francisco Bay Area. His most recent books are (with Susan D. Holloway), Through My Own Eyes: Single Mothers and the Cultures of Poverty (Harvard, 1997). Inside Charter Schools: The Paradox of Radical Decentralization (Harvard, 2000), and Government Confronts Culture (Taylor & Francis, 1999). Emily Hannum is assistant professor of sociology at the University of Pennsylvania. Her research focuses on education, poverty, and social inequality, particularly in China. Recent publications include "Ethnic Differences in Basic Education in Reform-Era Rural China" (forthcoming, Demography) and "Education and Stratification in Developing Countries: A Review of Theories and Empirical Research" (with Claudia Buchmann, Annual Review of Sociology, 2001). Currently, she is working on a project sponsored by the Spencer Foundation that investigates factors in the family, school and community that support rural children's education and healthy development in Northwest China. Hanna Ayalon is associate professor at the School of Education and Department of Sociology and Anthropology, Tel Aviv University. Her areas of interest include sociology of education and quantitative research methods. Her current research focuses on reforms and inequality in Israeli secondary education and the expansion of higher education. David P. Baker, professor of education and sociology at the Pennsylvania State University, currently directs the university's Institute for Policy Research and 181
182
ABOUT THE AUTHORS
Evaluation. His scholarly interests focus on the institutionalization of formal education in modern society. Carl L. Bankston III conducts research on issues of immigration, social capital, sociology of education, and sociology of religion. His work has appeared in a wide variety of journals, including Social Forces, Sociology of Education, Journal of Marriage and the Family, and International Migration Review. He is co-author (with Min Zhou) of Growing Up American: How Vietnamese Children Adapt to Life in the United States (Russell Sage Foundation, 1998), co-author (with Min Zhou) of Straddling Different Social Worlds: The Experience of Vietnamese Refugee Children in the United States (ERIC Clearinghouse on Urban Education, 2000), and co-author (with Stephen J. Caldas) of A Troubled Dream: The Promise and Failure of School Desegregation in Louisiana (Vanderbilt University Press, 2001). Claudia Buchmann is an assistant professor of sociology at Duke University. Her research focuses on comparative education and stratification in developing countries with an emphasis on educational inequality in Africa. Currently, she is investigating the sources of racial differences in educational enrollment and attainment in South Africa at the end of apartheid. Recent publications include: "Family Structure, Parental Perceptions and Child Labour in Kenya: What Factors Determine Who is Enrolled in School?" (Social Forces, 2000) and "Education and Stratification in Developing Countries: A Review of Theories and Research," with Emily Hannum, (Annual Review of Sociology, 2001). At Duke University, Professor Buchmann teaches courses on the sociology of education, comparative historical methods, and globalization. Gilberto Q. Conchas is an assistant professor of education at Harvard Graduate School of Education. His research focuses on socio-cultural processes within the school context that structure variations in educational opportunity for lowincome Latino, Asian American and African American youth. Recent publications include "How Context Mediates Policy: The Implementation of Single Gender Public Schooling in California" (with A. Datnow and L. Hubbard) (Teacher's College Record, 2001) and "Structuring Failure and Success: Understanding the Variability in Latino School Engagement" (Hai-vard Educational Review, 2001). Kimberley Goyette is an assistant professor of sociology at Temple University. Her research interests are stratification, education, race and ethnicity, and immigration. Currently, she is working on a research project funded by a
Abouth the Authors
183
National Academy of Education/ Spencer Foundation postdoctoral fellowship that investigates the mismatch between educational expectations and attainment across race and social class. Grace Kao is assistant professor of sociology and Asian American studies at University of Pennsylvania. She received her Ph.D. from the University of Chicago, and is a former National Academy of Education/Spencer Foundation Post-Doctoral Fellow. Currently, with support from NICHD, she is examining interracial romance and friendship among adolescents, with Kara Joyner. Patricia Fernandez Kelly holds a joint position in the Office of Population Research and the Department of Sociology at Princetown University. She has conducted extensive research on economic internationalization, women's employment in export-processing zones and migration. Her pioneering research focused on Mexico's maquiladra program. With filmmaker Lorraine Gray she co-produced the Emmy Award winning documentary, The Global Assembly Line, which portrays the effects of international investment on working families in the Phillipines, Mexico and the United States. She is co-author of For We Are Sold, I and my People: Women and Industry in Mexico's Frontier and the editor with June Nash of Women, Men and the International Division of Labor. Michal Kurlaender is a doctoral candidate in education policy at Harvard University Graduate School of Education. Her research interests include educational reform efforts. She works as a researcher for The Civil Rights Project at Harvard University. Maryellen Schaub is completing her Ph.D. in sociology at the Pennsylvania State University, where she is also an instructor in the sociology of education. Her scholarly interests are historical changes in parenting, young children's cognitive development, and schooling. Yossi Shavit is currently head of the Department of Sociology and Anthropology at Tel Aviv University and Secretary of the ISA's Research Committee on Social Stratification (RC28). He is interested primarily in processes of educational stratification, and is conducting research on the consequences of educational expansion and differentiation for equality of educational opportunity in Israel and other countries. Raymond Sin-Kwok Wong is associate professor of sociology at the University of California-Santa Barbara. His main areas of interests are social stratification,
184
ABOUT THE AUTHORS
sociology of education, and statistical methods. His recent publications include "Multidimensional Association Models: A Multilinear Approach" (Sociological Methods & Research, 2001) and "Multidimensional Influences of Family Environment in Educational Attainment: The Case of Socialist Czechoslovakia" (Sociology of Education, 1998). Min Zhou's research focuses on immigration and immigrant adaption, race and ethnicity, and urban sociology. Her work has been published in numerous books and journals, including the American Sociological Review, the Annual Review of Sociology, Social Forces, and the Annals of the American Academy of Political and Social Science. She is author of Chinatown: The Socioeconomic Potential of an Urban Enclave (Temple University Press, 1992), co-author (with Carl Bankston) of Growing Up American: How Vietnamese Children Adapt to Life in the United States (ERIC Clearing house on Urban Education, 2000), and co-editor (with James V. Gatewood) of Contemporary Asian America: A MultiDisciplinary Reader (NYU Press, 2000).
INDEX
achievement 1-2, 5-6, 9, 11, 13, 15-16, 19-21, 28, 31-32, 34, 42-43, 47-49, 59, 61-62, 64-67, 72, 85, 90, 127, 133-134, 136, 139, 147, 152, 154-155, 163, 167, 172-173, 176 Africa 111, 137 aspirations 8, 10, 23, 31, 78-79, 85-102, 108, 129-130, 173-174 college 21, 23, 31, 43, 46, 49, 52-61, 65, 71-72, 85, 87-102, 106-107, 122, 129, 174 cross-cultural 171-172, 176, 178 cultural capital 4, 8, 11, 88, 125, 129, 131, 133-135, 138-139, 145, 148-149, 151, 152, 154-156, 162-164, 166-168, 173, 177 cultural resources 156-157, 172, 176 culture 3, 6, 45, 74, 76-77, 79, 82, 128, 138, 173 differences in achievement 11 economic sociology 75 education 1, 4, 8, 14, 16, 19, 21, 24 27, 29, 31, 33-34, 39, 42-47, 49, 56, 77-79, 85-96, 100-102, 106-122, 125-131, 135-136, 139-141, 143-144, 146, 148-150, 152, 155-157, 161, 163, 167, 172-175, 177-178 educational attainment 34, 85-86, 100, 105-107, 115, 117, 122, 134, 137-138, 167 educational inequality 106-107, 109, 161-162, 173, 175 educational opportunity 105, 120, 134 educational organizations 5, 175, 178 educational stratification 119-120, 129-131, 134, 155, 164, 167, 173, 176 educational structures 106, 131, 138, 177 ethnicity 15, 21, 23, 25-26, 28, 38, 46, 61-66, 79-80, 85, 87, 89-90, 96, 111-119, 176 family environment 161, 166 family structure 23, 25, 50-51, 57, 134, 137-138, 146, 148-149, 151-152, 157, 165, 174 gender 3, 21, 80, 110-112, 114-116, 118-119, 142-143, 149, 154-155 grade repetition 133, 135, 141-145, 147-149, 151-154, 156, 173 immigrant adaptation 16, 80 immigrants 13-15, 17-23, 25-26, 28-32, 43, 47, 67, 73, 77-81 immigration 3, 13, 25, 28, 31, 46^17, 57-60, 71-72, 90 inequality 4, 10, 14, 48, 67, 75, 105-107, 110, 116, 120, 127, 167, 176 international 156, 172 Israel 2, 8, 17, 105, 108, 110-111, 122, 130, 175 Kenya 2, 8, 133, 135, 137, 139-143, 148, 154, 155-157, 162-163, 173 language 3, 8, 39, 80, 133, 135, 139, 145, 147-148, 151-154, 156,-157, 173 matriculation 105, 108-111, 113-121, 130 minorities 30, 46, 48-49, 66-67, 77, 89, 130 185
186 mobility 6, 8, 10, 18-19, 31, 77, 89-90, 105-107, 121, 130, 140-141, 143, 155, 165, 175 network ties 19, 31-34, 164 networks 5-9, 11, 15, 17, 19, 21, 25, 32-33, 42, 47, 67, 74, 76, 87, 127-128, 130-131, 137, 143, 165, 171-172, 175, 178 occupational attainment 105, 110, 116, 118, 120, 130 parent-child relationships 17, 23, 26, 28, 101, 165 parental behavior 96 parental involvement 13, 20, 25-26, 28, 33, 136, 157, 165, 174 race 4, 21, 23, 25-26, 28, 38, 49-50, 58, 65-66, 80, 85, 87, 89-90, 92, 95-96, 99, 101-102, 129, 140 religion 18, 22, 28-29, 39, 80 religiosity 80, 81 savings 85-90, 94, 99, 100-102, 173-174, 176 school achievement 2, 8, 17, 19, 30, 32, 41, 82, 133, 135, 163, 166
INDEX school effects 125 second chance 105-111, 113-121, 130, 175 secondary education 88, 95, 108-111, 113-114, 117-118, 120-121, 156 shadow education 8, 131, 133, 135, 139-146, 148-155, 163-165, 168, 173, 175 shadow schooling 133, 135, 143-144, 154-156, 173, 177 social capital 1-18, 25, 28, 31-34, 41-45, 47-62, 65-67, 71, 73-78, 82, 125-131, 133-138, 143, 145-148, 152, 154-157, 161-168, 171-179 social networks 2, 5-8, 13-16, 18, 20, 23-26, 28, 33-34, 42, 76-77, 82, 130, 165, 174, 176-177 social relationships 136, 167, 175-178 social resources 136, 176 social stratification 121, 172, 175, 178 status attainment 86, 88-89, 129-130, 171-172, 174 strong ties 74, 79 tracking 128, 175 tutoring 133, 139, 142, 144, 149-150, 155-157, 163, 173, 177 weak ties 74-75, 79, 81-82, 175, 177

Bruce Fuller, Emily Hannum Schooling and Social Capital in Diverse Cultures, Volume 13 Research in Sociology of Education 2002

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bruce Fuller, Emily Hannum Schooling and Social Capital in Diverse Cultures, Volume 13 Research in Sociology of Education 2002

Uploaded by

Copyright:

Available Formats

INTRODUCTION: SCAFFOLDS FOR ACHIEVEMENT?

INSTITUTIONAL FOUNDATIONS OF SOCIAL CAPITAL

GRASPING SOCIAL CAPITAL, UNDERSTANDING ITS SUDDEN RISE

BRUCE FULLER AND EMILY HANNUM

Introduction: Scaffolds for Achievement?

BRUCE FULLER AND EMILY HANNUM

REINVENTING SOCIAL CAPITAL

Introduction: Scaffolds for Achievement?

BRUCE FULLER AND EMILY HANNUM

BRINGING SOCIAL CAPITAL TO ISSUES OF SCHOOLING

Introduction: Scaffolds for Achievement?

BRUCE FULLER AND EMILY HANNUM

HOW DOES SOCIAL CAPITAL ADVANCE CHILDREN'S LEARNING?

Introduction: Scaffolds for Achievement?

BRUCE FULLER AND EMILY HANNUM

VALUE-ADDED BY SOCIAL CAPITAL? DRAWING FROM CLOSE CONCEPTUAL COUSINS

Introduction: Scaffolds for Achievement?

BRUCE FULLER AND EMILY HANNUM

SOCIAL CAPITAL AND IMMIGRANT CHILDREN'S ACHIEVEMENT

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant Children's Achievement

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant Children's Achievement

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant Children's Achievement

CARL L. BANKSTON III AND MIN ZHOU

DATA AND METHODS

Social Capital and Immigrant Children's Achievement

Social Capital and Immigrant Children's Achievement

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant Children's Achievement

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant Children's Achievement

5.625** (0.425) 3.045 0.116

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant Children's Achievement

Table 4. Unstandardized Coefficients of Predictors of Average Grades. (Standard Errors in Parentheses)

3.368** (0.087) 0.704 0.066

2.773** (0.095) 0.679 0.134

2.185** (0.120) 0.671 0.154

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant Children's Achievement

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant Children's Achievement

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant

CARL L. BANKSTON III AND MIN ZHOU

Social Capital and Immigrant

CARL L. BANKSTON III AND MIN ZHOU

APPENDIX: VARIABLES IN THE ANALYSIS

Social Capital and Immigrant Children's Achievement

KIMBERLY A. GOYETTE AND GILBERTO Q. CONCHAS

Family and Non-Family Roots of Social Capital

KIMBERLY A. GOYETTE AND GILBERTO Q. CONCHAS

DATA AND METHODS

Family and Non-Family Roots of Social Capital

DIFFERENCES IN STUDENTS' STUDYING BEHAVIOR

KIMBERLY A. GOYETTE AND GILBERTO Q. CONCHAS

Demographic and Socioeconomic Characteristics by Ethnicity, NELS 1988-1990 Panel.

67.8 15.5 0.4 16.3 0.44 0.93

13.0 35.4 38.8 12.8 -0.71 0.68

8.0 22.9 27.6 41.5