When Are Racial Disparities in Education the Result of Racial Discrimination? A Social Science Perspective
Susan Notes: The question is crucial: If public schools do not address educational disparities, then who or what institution will? As Roslyn observes, public education is the only institution that can touch the lives of all students in this nation. We can't continue to ignore this.
In this article I seek to answer the question, "When are racial disparities in education the result of racial discrimination?" To answer it I synthesize the social science research on racially correlated disparities in education. My review draws from the sociology, anthropology, political science, psychology, history, and education literatures. I organize explanations into six categories: biological determinism, social structure, school organization and opportunities to learn, family background, culture, and the state. I arrive at three answers. The first is a definition: Racial discrimination in education arises from actions of institutions or individual state actors, their attitudes and ideologies, or processes that systematically treat students from different racial/ethnic groups disparately or inequitably. The second answer is that while distinguishing racial discrimination from disparities may be an interesting intellectual, legal, and statistical challenge, the conclusion probably is less meaningful than social scientists and policy makers might hope. The third answer follows from the first two. I propose the following reformulation of the original question: "When are racial disparities in education not due to discrimination?" I argue that the reformulated question is more likely to bring solutions to the race gap than the original one. Even if we conclude that discrimination does not cause racial disparities in education, we should not conclude that schools have no role in addressing them. If public schools do not address educational disparities, then who or what institution will?
Globalization, immigration, and the post-Fordist information-based economy are transforming 21st century America. The hallmarks of globalization-the flow of capital, information, and people across political boundaries-are sure to enrich, challenge, and complicate contemporary society. Public education already is experiencing many of the challenges associated with these trends. Schools and communities more ethnically and racially diverse than ever before. Racial and ethnic disparities in educational are not new to America's schools. Recent demographic, political, and economic shifts have the potential to exacerbate existing inequalities. In addition, social class and gender divisions complicate the emerging racial gaps in educational outcomes.
Identifying when disparities are the result of discrimination is difficult. Historically, racial discrimination was integral to public education, the intentional result of discriminatory laws and practices. Jim Crow education was designed, implemented, and upheld by the state (Anderson, 1988; Kluger, 1977; Walker, 1999; Walters, 2001; Watkins, 2001).2 Since the middle of the last century, however, laws, court rulings, and policies of the state-along with many heroic efforts by private citizens-have eliminated the formal legal architecture of educational discrimination. During the past three decades legal segregation has been outlawed, although de facto resegregation is on the increase (Clodfelter, Ladd, & Vigor, 2002; Orfield & Gordon, 2001; Orfield & Yun, 1999; Yun & Reardon, 2002). Importantly, literacy and median years of schooling are comparable among blacks and whites, multicultural curricula are used widely; and overtly racist material has been eliminated. Gamoran (2001) points out that any comparison of American education today with that of 1901 reveals overt racial discrimination and disparities in school outcomes have been reduced dramatically. Nevertheless, racially correlated disparities continue.
This article has two tasks: first, to synthesize the extant social science research on racially correlated disparities in education to better understand their structural and cultural antecedents because doing so helps clarify when such disparities are the results of discrimination and, second, to identify the key points in students' educational trajectories at which discrimination is likely to occur. My review of the social science literature draws from the sociology, anthropology, political science, psychology, history, and education literatures.3 Although I attend to student-level factors, I heed Coleman's (1994) admonition to scholars and make the explanatory focus of this article the social system, not the individual.
I begin by theorizing race and racial identity because they are central to any discussion of discrimination in education and are crucial to the social dynamics in racially plural societies like America. In the next section I catalog current racial disparities in education and propose how we might conceptualize when disparities are due to discrimination. In the third section I synthesize the substantive and theoretical social science literature accounting for racial disparities and discrimination in education. I conclude with a series of answers to the question suggested in the title, what constitutes racial discrimination in education? The final answer reconceptualizes the core problem itself: Rather than considering when disparities reflect discrimination, I propose that the more appropriate question is "When are racial disparities in education not discrimination?" I announce this reformulation now in the hope that it will guide the reading of this article.
CONSTRUCTING RACE AND ETHNICITY
Race and racial identity are central to this article, but these subjects are too immense to be fully theorized here. I consider race and ethnicity to be historically contingent social constructs intimately entwined with identity; they are meaningful and consequential in a sociopolitical and historical context (Waters, 1990). Recent theories of race and ethnicity, such as racial formation theory (Omi & Winant, 1986) and critical race theory (Delgado & Stefancic, 2001) require us to consider the process-oriented and relational character of racial meanings and identities.
RACE AND ETHNICITY
Social scientists generally agree that races are socially constructed in loose relation to perceived phenotypical differences among humans. Ethnicity relates to national ancestry and signifies the cultural, linguistic, and historical differences among groups. Key aspects of ethnicity are the beliefs on the part of people who identify with an ethnic group that they descend from common ancestors, share a common culture with coethnics, and choose to identify with that ethnic group (Waters, 1990).
A group's history of contact with the United States, the conditions of its incorporation into American society, and the contemporary politics of an ethnic group's country of origin influence the way an ethnic group constructs its identity, the nature of the education historically provided, and the barriers or bridges to education currently afforded its children. For example, the U.S. government extends special educational privileges to most Cuban American children because of their status as political refugees from a communist state. Haitians, however, are defined as economic refugees and thus do not enjoy these privileges (Schmidt, 2001; Van Hook, 2002).
Racial self-identity has two complementary aspects: how others identify a person's race and how the person constructs her or his own racial identity, in part as a reaction to others' behavior. Both aspects are relevant to racial discrimination in education. A student's racial or ethnic self-identity need not be consistent with others' construction of that identity. On the other hand, racial dynamics in the United States are concerned less with how an individual constructs his or her racial identity than with how others perceive, construct, and label it (Fordham, 1996; McCarthy & Crichlow, 1993). To capture the dynamic interplay between culture, structure, and agency in producing discrimination in education, one must consider the contributions of all three to racial identity.
O'Connor (2001, p. 159) observes that, to make sense of the ways that culture, structure, and human agency affect achievement outcomes, we must understand how people are positioned differently in the social world and how in that world, individuals have multiple social identities. She observes that students experience their own structured and cultured social identity both by interpreting and performing their identities (reflection) and by registering and reacting to others' responses to them (refraction). To paraphrase Cousins (2002), racial identity is not simply about being black, biracial, white, Latino, Asian, or Native American in a white world but about being so in a white, media- and technology-saturated, capitalist- and information-driven globalizing world of relations between youths and adults, boys and girls, and men and women.
To truly fulfill the article's mission, would require me to include comparative research on students who are Asian, Latino, biracial, and Native American (including American Indians, Inuits, Hawaiians, and other Pacific Islanders), as well as black and white. Even panethnic labels such as Asian or Latino mask important ethnic distinctions within these larger categories (Lopez & Espiritu, 1990). Any comprehensive treatment of discrimination in education would include issues of generation, language, and immigration status as well, but these topics are well beyond the scope of this article. Therefore, despite problematizing the concepts of race and ethnicity, and arguing for concepts that are socially constructed, historical, relational, and variable, I yield to the practicalities of space and mission and contradict key aspects of the above arguments. That is, I focus primarily on black-white racial disparities, and when I label students, I use the five static categories commonly employed by the federal government and most school systems: Asian Americans, blacks, Hispanics, Native Americans, and whites.
RACIAL DISPARITIES AND DISCRIMINATION IN EDUCATION
During the 1970s and 1980s, the racial gap in educational outcomes narrowed (Grissmer, Flanagan, & Williamson, 1998). Today, however, racially correlated disparities in K–12 education are present in grades, test scores, retention and dropout rates, graduation rates, identification for special education and gifted programs, extracurricular and cocurricular involvement, and discipline rates. The task of determining when educational disparities are caused by racial discrimination is complicated by the close association of race with social class.4 Sorting out race effects from class effects is extremely difficult, both conceptually and methodologically. Reviews of recent large-scale studies suggest that at most, socioeconomic background explains 33% of the racial gap in education (Hedges & Nowell, 1998, 1999; Jencks & Phillips, 1998). This leaves a large portion of the variance in racial disparities in education that must be explained by factors other than social class differences among students of different races.
According to results obtained by the National Assessment of Educational Progress (NAEP), white, black, and Latino students in all assessment years (ages 9, 13, 17) show gains in mathematics between the mid-1970s and the late 1990s. Whites and blacks show gains in science for 9 and 13 year old students, and Latino students show gains in science across all assessment years. Among blacks and Latinos, overall gains in reading appear at each age; for whites, reading gains are evident for 9- and 13-year-old students (Campbell, Hombo, & Mazzeo, 2000; Hedges & Nowell, 1998, 1999). The gains, however, conceal an important story. According to Michael Nettles, minority children have mastered the basics but not higher level skills (Hoff, 2000).
NAEP results indicate that white students score higher in reading, mathematics, and science than do blacks and Latinos. According to Hedges and Nowell (1998, 1999) other national surveys indicate racial gaps comparable in size to the current differences found in NAEP: The black-white differences among 17-year olds range from 0.7 SD in reading to approximately 0.9 SD in mathematics. NAEP results suggest that racial gaps in test scores held steady among 9- and 17-year-old students but declined among 13-year-olds (Campbell et al., 2000). In short, across all three age cohorts in all three subjects, the smallest test gaps occurred in the 1980s; although the gaps of the 1990s were larger than in the previous decade, they were smaller than in the 1970s (Campbell et al., 2000; Gamoran, 2001; Hedges & Nowell, 1999).
SAT results reflect similar patterns. The College Board (2001) reports that SAT verbal scores are highest among whites (528) and lowest among blacks (434); SAT mathematics scores are highest among Asians (565, whites average 530) and lowest among blacks (426).
Other educational indicators show similar differences by race. Using Current Population Surveys, Hauser, Simmons, and Pager (2000) report that dropout rates are the lowest among whites and highest among Latinos. Since the 1970s, rates have declined among whites and blacks but not among Latinos. Blacks are more likely than whites to repeat a grade (Campbell et al. 2000) and to be placed in special education programs (Oswald, Coutinho, & Best, 2000), especially in school systems operating under court orders to desegregate (Eitle, 2002). Blacks, Latinos, and Native Americans disproportionately are found in lower tracks (Hallinan, 1998; Lucas, 1999; Lucas & Berends, 2002; Mickelson, 2001; Oakes, 1985, 1994; Oakes, Muir, & Joseph, 2000; Welner, 2001) where curricula and instructional practices are weaker. Minorities attend schools where they have access to fewer advanced placement classes than whites (Pachón, Federman, & Castillo, in press). Blacks, Latinos, and Native Americans are more likely to learn in schools with fewer material and teacher resources, a weaker academic press, and greater concentrations of poor, homeless, limited English-speaking, and immigrant students (Kahlenberg, 2001; Lee, Burkam, & LoGerfo, 2001; Natriello, McDill, & Pallas, 1990; Van Hook, 2002).
IDENTIFYING WHEN DISPARITIES RESULT FROM DISCRIMINATION
As shown previously, the task of describing racial disparities is fairly straightforward. Determining when they are caused by racial discrimination is far more complex, however. We can all agree on simple instances of discrimination-a racist teacher or a dual school system-but it is more challenging to identify complex cases because they result from the cumulative effects of institutions' and people's actions conditioned by structure and culture and framed by history.
Under Jim Crow education, disparities in opportunities to learn and in outcomes were caused by official racial discrimination against blacks, Native Americans and (in some states) Asians and Latinos. Today, however, one can argue that racially correlated disparities may or may not be due to racial discrimination. For example, racial gaps in test scores are not necessarily evidence of discrimination-unless there are systematic racial differences in opportunities to learn (OTL) materials on the tests and in access to licensed, experienced educators. Similarly, not all racial disparities in track placement are necessarily discriminatory. Although decisions in track placement decisions lie primarily with educators, the students and their parents also are involved in the choices to varying degrees. Yet if comparably able students of different races are placed systematically in different tracks because educators teach, advise, or schedule blacks differently than whites, we may conclude that discrimination is involved. The existence of systematic racial disparities over time suggests the school system engages in institutional discriminatory practices.
HOW SOCIAL SCIENCE ACCOUNTS FOR THE RACIAL GAP IN EDUCATION
In this article I focus on the not-so-simple manifestations of racial discrimination and on the complex, often ambiguous processes that generate them. To successfully distinguish when the race gap is more than a disparity and is discrimination, it is necessary to survey the explanations offered by social scientists for the gap. I organize explanations for the race gap in educational outcomes into six major categories: biological determinism, social structure, school structures and opportunities to learn, family background, culture, and the state.
Periodically during the 20th century, theories that the racial gap is due to inherited, fixed racial differences in cognitive abilities circulate in the popular and scientific press (Jensen, 1969; Terman, 1923). Herrnstein and Murray (1994) resurrected this argument with The Bell Curve, a book claiming that racial and social class differences in educational achievement, attainment, wealth, and poverty are the result of immutable genetic differences in cognitive ability. They recommended an end to what they regard as futile social welfare and educational programs designed to eliminate racial and class gaps in outcomes.
Decades ago, anthropologists established that the relationship between genetic diversity and racial/ethnic diversity is weak; diversity in the former does not map easily onto the latter because genetic variation within socially constructed racial groups is greater than between those groups (Marks, 2002b). Moreover, genetic variations among humans do not predict patterns of behavioral diversity (Marks, 2002c). The scholarship of biologists (Gould, 1981), molecular anthropologists (Marks, 2002a), cognitive psychologists (Gardner, 1993; Myerson, Frank, Rains, & Schnitzler, 1998) and sociologists (Fisher et al. 1996; Sorensen & Hallinan, 1986) has discredited the putative scientific basis of biological determinist explanations for racial differences in academic performance.
SOCIAL STRUCTURAL EXPLANATIONS
Enduring inequalities in outcomes tarnish the significant accomplishments of American educational institutions. The United States is not unique in this regard: Blossfeld and Shavit (1993) examined 13 industrialized nations and found that the relative likelihood of graduating from secondary school and entering higher education for people from different socioeconomic backgrounds has remained essentially the same over the past several decades despite an enormous expansion of educational capacity in all 13 nations. In the Netherlands and Sweden, where overall social inequality declined, so did the effects of social background on educational attainment. In view of the relationship between race and social class, it is therefore necessary to consider the connection between social stratification and the racial gap in educational outcomes.
During the past few decades, racial inequality in educational outcomes has declined. But can we anticipate that the race gap will eventually disappear? From a social stratification perspective, the answer is it depends, because approximately one third of the racial gap is estimated to be due to social class. Thus, even if racial discrimination in education were to be obliterated, a racial gap still would exist because blacks (and most other minorities) are poorer than whites.
Social scientists offer a number of theories to explain the resilience and durability of class differences in educational outcomes. One of the most provocative is reproduction theory. Over 25 years ago, Bowles and Gintis (1976) argued that in a capitalist society, schools are designed to reproduce the class system by providing unequal education to children according to their class of origin. Because the social relations of school (and the home) correspond to the social relations of production, schools foster capitalist structures of production and reproduce class inequities. Reproduction theory has been criticized for undertheorizing race and gender, for failing to show explicitly how schooling reproduces class inequality, and as overdeterministic. Whatever its theoretical shortcomings, reproduction theory offers a powerful analysis of the differential opportunities to learn that exist along class lines.
Resistance theory (Giroux, 1981; MacFarland, 2000; MacLeod, 1986; Willis, 1977) incorporate human agency into the processes by which schools reproduce class inequality. Resistance theorists argue that youths respond to the disjuncture they see between (on the one hand) promises of mobility and social transformation through educational success and (on the other) a stratified political economy in which class inequalities are daily realities. Adolescents' partial understanding of this disjuncture results in conscious challenges to schooling. By rejecting educational credentials, however, resistant students foreclose any possibilities of upward mobility through education. Like reproduction theory, resistance theory undertheorizes the role of race and gender in social and educational inequality.
Maximally Maintained Inequality
Class inequality in education has not declined despite the increase in educational access during the last century because dominant social groups have preserved their advantaged positions. Once disadvantaged groups attain the educational credentials previously held by the dominant groups, the credential requirements for higher status jobs are raised. Social and educational inequalities are maintained because privileged groups are able to protect their advantages until all group members reach a given status (such as high school graduation or a BA degree). Then the credential requirements for status maintenance and mobility are ratcheted up to the next level, where the elites are better able to meet the new standards. Raftery and Hout (1993) theorize that these dynamics operate as a maximally maintained inequality (MMI) system.
Effectively Maintained Inequality
Lucas's (2001) theory of effectively maintained inequality (EMI) refines MMI. Lucas notes that the effects of social background operate in at least two ways: they influence who completes a given level of education if completion is not universal at that level, and they influence the kind of education people will receive within levels of education that are nearly universal (such as high school). Curricular tracking is a central mechanism through which EMI operates. For example, EMI is the key reason why more highly privileged parents vigorously fight tracking reforms (Wells & Serna, 1996). The social background factors that determine whether students will continue schooling also effect where in the stratified curriculum they will be schooled. As they make the transitions from elementary school through higher education, their location in the stratified curriculum at any given point has implications for their curricular trajectories during the remainder of their educational careers. In these ways, Lucas concludes, advantages of social background effectively maintain educational privilege for children of the advantaged (pp. 1678–1681).
All four models describe how social structure intersects with schooling to generate racial discrimination in educational outcomes. This is the case because social reproduction and mobility depend on not only the acquisition of educational credentials, both in terms of skills and certification, but also socialization into the relations of production. As I show in the following sections, discrimination influences racial differences in the opportunities to acquire these educational credentials.
SCHOOL STRUCTURE AND RACIALLY CORRELATED OPPORTUNITIES TO LEARN
As Lee (2002) observes, we cannot hold schools responsible for the racial and class disparities in school readiness that are evident as soon as kindergartners walk through the classroom doors. The educational system is responsible, however, for the initial disparities by race that grow rather than diminish with each year children attend school.
American school systems (almost 16,000) are highly stratified on precisely the dimensions that are most related to student achievement. The ways schools are stratified is documented in numerous scientific and popular articles and books. School organization is a key contributing factor to educational stratification. Possibly the most powerful recent popular account of this inequality appears in Kozol's (1992) Savage Inequalities in which the author describes the breadth and depth of racial and class differences in schooling across the United States. In this section I focus on the relationship between the racial gap and three aspects of school organization: resources, racial composition, and tracking.
Whether money matters for school outcomes is a longstanding debate dating back at least to the Coleman Report's finding that funding is not closely related to achievement (Coleman et al., 1966). Although skeptics remain unconvinced (Hanushek, 1994, 1996, 1997), a growing body of research establishes that money does matter and that where and how the money is spent is also extremely important (Ferguson, 1998a, 1998b; Greenwald, Hedges, & Laineet, 1994; Hedges, Laine, & Greenwald, 1994a, 1994b; Weglinsky, 1997).5
Not only bricks and books but also human resources, such as high-quality, experienced, credentialed teachers instructing in their area of expertise in small classes (Darling-Hammond, 2000; Ingersoll, 1999; Lankford, Loeb, & Wyckoff, 2002), are directly related to money. Other resources are indirectly related to funding levels, such as stable peers, active involvement by parents, motivated peers who value achievement and who share knowledge with classmates, and a school climate imbued with high expectations. These are associated with the racial and SES composition of communities (Kahlenberg, 2001).6
Students who attend resource-poor schools are disproportionately members of minority groups (Duncan & Brooks-Gunn, 1997; Lee, Burkam, and LoGerfo 2001; Payne & Biddle, 1999). Given the system of public school financing, which depends largely on property taxes, and in view of the racial segregation in public and private housing markets (Powell, Kearney, & Kay, 2001), it is not surprising to find racial (and class) differences in school financial resources and in the opportunities to learn that they purchase.
The key here is that blacks and other minorities are less likely than whites to have equitable access to these critical resources both within and between school systems. The racial discriminatory practices that generate and allocate resources inequitably to schools contribute to the racial gap in outcomes.
State Racial Composition
The original social science rationale for school desegregation rests largely on claims that desegregation improves black youths' access to the higher-quality education more often provided to whites. Almost five decades after the epochal Brown v. Board of Education (1954) decision, there is little argument about the positive long-term effects of desegregation on minority students' status attainment, racial attitudes, and other life course indicators (Armor, 1995; Braddock & McPartland, 1988; Crain & Mahard, 1978; Hawley, 2002; Wells & Crain, 1994). Although the Coleman Report (Coleman et al., 1966) showed that academic outcomes were better for blacks who attended desegregated schools, until recently questions remained about the positive short-term effects of desegregation on minority youths' achievement (Armor, 1995; Cook, 1984; St. John, 1975).
A growing body of recent research identifies the benefits of desegregation and the harms of segregation. Grissmer, Kirby, Berends, and Williamson (1994) concluded, on the basis of comparisons of NAEP scores over time, that the significant increases in academic achievement of black students in some states and not in others were due, in part, to desegregation during the 1970s and1980s. Hallinan's (1998) synthesis of the social science evidence identifies multiple positive effects of diversity on students' outcomes. A group of eminent social scientists filed an amicus curiae brief supporting desegregation in the 1992 Freeman v. Pitts case (School Desegregation: A Social Science Statement, 1991). These researchers argued that when schools employ practices that enhance equality of opportunity, including the elimination of tracking and ability grouping, desegregation produces clear (albeit modest) academic benefits for black students and does no harm to white students.
Other recent empirical research offers further evidence of the harm of segregation (Bankston & Caldas, 1996) and positive academic outcomes from desegregated schooling (Brown, 1999). My own survey research reveals that the longer blacks and whites learned in desegregated schools and classrooms, the better are their academic outcomes (Mickelson, 2001, 2002a). I also found that many of the potential gains of desegregation were subverted by resegregation through tracking, even in "desegregated" schools.
Hawley (2002) synthesized the extant empirical literature on the effects of diversity on educational quality for the U.S. Department of Justice. He concluded that "students who have the opportunities to learn in schools that are populated by students from different races and ethnicities can have an education that is superior to that of students who do not have this opportunity" (p. 1).
Even as America's schools resegregate (Orfield & Eaton, 1996; Orfield & Yun, 1999; Yun & Reardon, 2002), several recent federal court decisions have concluded that contemporary manifestations of school segregation are not evidence of dual systems (Board of Educ. of Oklahoma v. Dowell, 1991; Capacchione v. Charlotte-Mecklenburg Schools, 1999; Freeman v. Pitts, 1992). However, judicial standards of evidence are often quite different from social science standards of evidence (Ryan, 2002), and, in my view, racially discriminatory practices by the school boards generate de facto resegregation. These actions include drawing school boundaries in ways that maximize racial homogeneity in schools, siting of new schools in white suburbs rather than midway between black and white communities, permitting greater numbers of advanced placement courses to be offered in middle-class white schools, or allowing the better teachers unfettered freedom to move to schools with less challenging, middle-class students. Therefore any race differences in outcomes that can be traced to resegregation-differential access to better teachers, safer schools, more rigorous academic climates-are evidence of racial discrimination by a school system.
As Hallinan and Sorensen (1977) observed, students cannot learn what they are not taught, no matter how highly motivated and how capable they are. Numerous studies indicates that student in higher tracks-even less academically able ones-learn more because they are exposed to broader curricula and better teaching (Braddock & Dawkins, 1993; Darity, Castellino, & Tyson, 2001; Finley, 1984; Gamoran & Mare, 1989; Hallinan, 2001; Mehan, Villanueva, Hubbard, & Lintz, 1996; Oakes et al., 2000; Slavin, 1990). Research also shows that a critical component of the racial gap in achievement is the relative absence of black students in higher-level courses and their disproportionate enrollment in lower-level ones (Oakes, 1993, 1994a; Jencks & Phillips, 1998; Lucas, 1999; Mickelson, 2001; Oakes, 1990, 1993, 1994a 1994b; Wheelock, 1992).7
Tracking and ability grouping begin very early in children's school careers (Alexander, Entwisle, & Lettgers, 1998; Kornhaber, 1997) and have consequences that follow students throughout the course of their education. The effects of ability grouping and tracking are cumulative: young students who possess similar social backgrounds and cognitive abilities but who learn in different tracks become more and more academically dissimilar each year they spend in school.
Central to tracking theory is the notion that track placement processes are objective, technical, and rational rather than subjective or arbitrary (Kulik & Kulik, 1982, 1987). Any flaws in the execution of tracking are merely glitches in implementation, not system design (Hallinan, 1994). Pallas and Natriello (1999) investigated the process of tracking. They dispel any remaining assumptions we may hold that tracking decisions are systematically objective and rational. They show that arbitrary and idiosyncratic placement decisions are common, and can be related to seat availability in classes or students' ascribed characteristics. A growing body of research (Baker & Stevenson, 1986; Kornhaber, 1997; Lareau, 1987; Lareau & Horvat, 2001; Lee, Smith, & Croninger, 1997; Lucas, 1999; Oakes, 1985; Useem, 1992; Wells & Serna, 1996) documents how privileged parents use their superior financial resources, knowledge, and social networks to ensure that their children are placed into the top academic trajectories. That schools and school personnel respond favorably to exercises of these race and class advantages is part of the problem.
The relationship between desegregation and tracking often is discussed in terms of first- and second-generation segregation (Meier, Stewert, & England, 1989; Welner & Oakes, 1996). First-generation segregation generally involves the racial composition of schools within a single district; second-generation segregation involves the racially correlated allocation of educational opportunities within schools, and typically is accomplished by tracking. Because tracking can undermine the potential benefits of policies such as busing which are designed to eliminate first- generation segregation, some courts have ruled since 1967 that it is unconstitutional for school districts to use tracking and ability grouping specifically to circumvent desegregation at the school level (Hobson v. Hansen, 1967; People Who Care v. Rockford Board of Education, 1994). Even so, researchers find tracking is used to maintain an unofficial white track within a desegregated school (Kornhaber, 1997; Mickelson, 2001; Welner, 2001). For example, Eitle's (2002) nationwide study reports higher levels of racially identifiable tracking of blacks into special education in districts under a court-ordered desegregation plan than in districts not under such orders. Once again, in accord with my definition of discrimination, I conclude that racially correlated ability grouping and tracking practices result in racially discriminatory educational outcomes.
Explanations for racial disparities in school outcomes that focus on family background fall into two broad categories. The first concerns characteristics of families, such as the number of children, the marriage status of parents, the number of adults in the household, wealth and income, and the adults' educational attainment. All of these indicators correlate with race. The second category concerns social class dynamics. These dynamics are deeply connected with the ways in which families interact with schools, how parents socialize their children for schooling, and how parents participate in their children's education.
Parents from different ethnic and social class backgrounds possess different levels and types of the resources critical for their children's education (Garcia, 2001; Guo & Van Wey, 1999; Hidalgo, McDowell, & Siddle, 1990; Irvine, 1990; Phillips et al., 1998; Roscigno, 1998, 2000; Suter, 2000; Yonazawa 1997). Many social scientists find it useful to consider families' resources as their human, financial, social, and cultural capital. In my view, differences in families' capital resources contribute to racially discriminatory outcomes because of the ways in which school practices and school employees fail grant privileges to elite families who have them and fail families who lack them.
Financial and Human Capital
Lee and her colleagues (Lee, Burkam, & LoGerfo, 2002) used a nationally representative early childhood data set to demonstrate the race and class inequalities in children's school readiness are evident when they first walk through the kindergarten classroom door. The association between family background and schooling was well known even when the Coleman Report (Coleman et al., 1966) reaffirmed its importance. Families' financial capital influences how well children eat, their health, and what kinds of books, if any, are present in their homes. In addition to the necessities of life, money purchases access to the best developmental preschools, tutors, computers, psychologists who test for giftedness, and homes with quiet bedrooms for doing homework. If one parent earns enough to support the family, the other parent is free to volunteer in the child's school (this assumes a two parent family). Parents' own human capital influences student outcomes as well: for example, the more education parents have, the better able they are to help with their children's homework. When schools are neutral with regard to (or passively accept) students' differences in parental resources, parents' resources (or lack there of) interact with school organizational structures and practices in ways that contribute to discrimination in education.
Social capital refers to student and parental knowledge about how to ask and whom to ask in schools, networks of parents who share information on how to customize their children's educational careers, and shared confidence and trust in school personnel (Baker & Stevenson, 1986; Bourdieu & Passeron, 1977; Carbonaro, 1998; de Graaf, 1986; Hilliard, 1989; Kerckhoff, 1996; Lareau & Shumar, 1996; Ma, 1999; Muller, 1998).8 Because of the ways in which public schools are organized, staffed, and operate, the social capital of educated middle-class white families is more conducive to school success than is the social capital of less highly educated ethnic minority and/or poorer families.9 Bryk, Lee, and Holland (1993) found that one reason for the success of Catholic schools is the social capital shared by the students' families. Catholic schools' sense of community, trust, common mission, and shared responsibility for all children contribute to every student's success.
Bourdieu's (1977) theory of cultural capital bridges cultural and structural dimensions of family background. It illuminates the (dis)continuity between the material and the symbolic elements of school and family cultures. Bourdieu viewed cultural capital as a mechanism for transferring class advantage from one generation to another. According to Lareau and Horvat (1999), Bourdieu's major insight into educational inequality is that, because schools embody dominant cultural forms (expressed in dialect, demeanor, tastes, sensibilities, stock of elite knowledge), students who possess more of the valued cultural capital fare better than do their otherwise comparable peers who possess less of this capital. A great deal of research demonstrates the relationships between possession of cultural capital and positive school outcomes (DiMaggio, 1982; Dumais, 2002; Farkas, 1996; Lamont & Lareau, 1988; Lareau, 1987; Lareau & Horvat, 1999).
Lareau and Horvat (1999) show that the value of cultural capital depends heavily on the particular social setting, on parents' and students' skill at activating their cultural capital resources, and on the reciprocal, negotiated process by which social actors in schools respond to the activation and accord it legitimacy. Lareau and Horvat argue that these factors create moments of inclusion or moments of exclusion for families. They show how the historical legacy of racial discrimination-in conjunction with school structures and operations-makes it more difficult for black parents, independent of their social class, to activate their cultural capital on behalf of their children.
Cultural Deprivation and Cultural Difference
Cultural deprivation theories (Lewis, 1966; Loury, 1985) propose that school failures stem from values and norms that are ill suited for school success (e.g., present verses future orientation). The insidious underlying assumption is that students from nonmainstream (and by implication, inferior) cultures are deprived. Policies based on cultural deprivation theory seek to change the culture of students and their families so that their behaviors better suit the demands of the school. Cultural deprivation theories form the basis for much of the original compensatory education movement but have been largely replaced with more benign theories of cultural difference (Valentine, 1969).
Without implying cultural superiority of mainstream culture, cultural difference theorists argue that racial variations in educational outcomes exist because cultural practices and values of certain racial and ethnic groups are more conducive to educational success (Bernstein, 1977; Bialystok & Hakuta, 1994; Delpit, 1996; Heath, 1983). For example, Caplan, Choy, and Whitmore (1991) identify a cohesive family and hard work as the "core" values that form the basis of Southeast Asian refugees' beliefs; these have enabled their youths to achieve well only a few years after their arrival in the United States. It is not clear, however, that core educational values of minorities who enjoy less success in school, such as blacks and Latinos, are essentially different from those of ethnic minorities who are more successful (for a somewhat different view, see Steinberg, Brown, & Dornbusch, 1992).
Cultural difference theorists also argue that school problems are linked to the failure of white-controlled educational institutions to incorporate nondominant cultures into school culture, curricula, pedagogy, or structures (Hallinan, 2001). Much of the multicultural education movement is a response to this critique (Banks, 2003).
Oppositional Cultural Framework
On the basis of ethnographic data he collected after conducting fieldwork in Stockton, California and elsewhere, Ogbu (1991) and Ogbu and Simons (1998) propose the existence of an oppositional cultural framework that influences minority achievement. They argued that many black students, as members of an involuntary minority group, often reject educational achievement as an avenue for success. They perceive labor market discrimination as a relatively permanent barrier that cannot be scaled through education, and they develop alternative frameworks for "making it." Cultural and language differences from whites become markers of a collective identity as an oppressed people. This perspective is expressed in black students' pointed disdain for the alleged link between education and opportunity. Based on their collective history of discrimination, and the perception that schools are majority-controlled institutions, activities associated with school success (speaking standard English, carrying books, doing homework, studying for tests) come to be viewed as compromises of black social identity and group solidarity. In this way, academic achievement comes to be associated with "acting white" (Fordham & Ogbu, 1986; Ogbu & Simons, 1998).
Ethnographic and survey research by other scholars both challenges and confirms the presence and effects of an oppositional cultural framework among blacks (Ainsworth-Darnell & Downey, 1998; Carter, 2002; Cook & Ludwig, 1998; Downey & Ainsworth-Darnell, 2002; Farkas, Lleras, & Maczuga, 2002; Ferguson, 1998a, 1998b, 2001; Mickelson, 2002b; Tyson, 2002). Whether or not black students-and other nonimmigrant minority youths-possess or reject oppositional cultural norms hostile to school success and embrace academic achievement remains a subject of intense debate.
Abstract and Concrete Attitudes
Cultural values are reflected in people's attitudes. In earlier research I demonstrate that attitudes toward education and the future take two forms (Mickelson, 1989, 1990). The first form, abstract attitudes, is based on the dominant American ideology that holds that education is the solution to most individual and social problems. All adolescents value education in the abstract, but these values do not predict differences in academic achievement. Achievement is affected by students' perceptions of the opportunity structure that awaits them. Concrete attitudes embody these perceptions which fluctuate according to the race, ethnicity, and social class forces that shape individuals' and groups' experiences in the American opportunity structure. Concrete, not abstract, attitudes predict achievement among all adolescents. By examining both abstract and concrete attitudes toward education in relation to high school achievement, I show how widespread consensus in educational values can exist across race and class lines while at the same time racial and class differences exist in actual school behavior and achievement.
A number of critics (Ainsworth-Darnell & Downey, 1998; Carter, 2002; Downey & Ainsworth-Darnell, 2002; Ferguson, 1998a, 1998b) incorrectly conflate concepts of abstract and concrete attitudes with Ogbu's (1991) notion of an oppositional cultural framework. Concrete attitudes operationalize a related but earlier construct advanced by Ogbu (1978), namely the job ceiling. In that earlier work he argued that children of castelike minorities (those who were incorporated into American society against their will) face barriers to job success despite possession of educational credentials. Children witness the labor market discrimination faced by their parents, siblings, and community members, and decide that putting effort into education may not pay off. My abstract and concrete attitude constructs test Ogbu's "job ceiling" hypothesis by examining how the effects of truncated opportunities associated with the racially stratified labor market shape minority adolescents' attitudes toward education and how those attitudes influence their achievement.
Steele (Steele, 1997; Steele & Aronson, 1998) examines how the cultural stereotype of black intellectual inferiority affects the academic performance of the most accomplished and most capable black students. He notes that this stereotype is widespread and has become part of every black child's stock of knowledge. Although awareness is not acceptance, Steele believes that the stereotype, nevertheless, impairs academic performance among the brightest and most academically motivated blacks. Using experimental data, he demonstrates that if academically able black students are cued about race before engaging in an intellectual task, their performance will be lower than that of comparable blacks who did not receive such a cue. Given the salience of achievement to high performers, the limited number of blacks in most university classrooms, and the pernicious cultural stereotype of black intellectual inferiority, Steele argues that "stereotype threat" inhibits blacks' academic performance because the most able black students are reluctant to engage fully in intellectual challenges lest they validate the stereotype through trying and failing. Instead their anxiety unconsciously leads them to disengage, and disengagement undermines performance. I return to subjects of stereotype threat, oppositional cultural frameworks, and racial discrimination in education later in the article.
THE STATE, SOCIAL POLICY, AND RACIAL GAPS IN SCHOOL OUTCOMES
As I have argued elsewhere (Mickelson, 2000), state actors make policy decisions that generate, perpetuate, or ameliorate conditions and structures responsible for racial disparities in education. Walters (2001, p. 35) argues that that throughout the United States, from the establishment of the common school in the 19th century until the present, racial inequality in educational opportunities has been an explicit policy of the state. She notes that not only do elites have greater access to the state, through which they influence policies, but also that certain policies continue to allow elites (primarily middle class whites) to activate their private resources so as to stave off the intended effects of contemporary equity-minded educational reforms. For example, reliance on local property taxes as the main source of school finance and the sanctity of local school district boundaries were critical to establishing inequality within and between communities, and to maintaining stratified schooling after certain educational policies shifted toward greater racial equality of educational opportunity in the 1950s.
According to Walters (2001), the social and political processes that produce racially segregated housing and a racially stratified labor market also generate racially segregated schools and racial inequality in the distribution of school resources. State policies of concentrating public housing for low-income, largely minority families in central cities (as opposed to scattered-site public housing or mixed income communities) affect the racial composition of schools. State policies of establishing or permitting resource inequalities within and between districts exacerbate educational disadvantages facing the minority and poor children who are concentrated in these resource-poor schools. Such policies compound neighborhood disadvantages with school disadvantages (Massey & Denton, 1993; Natriello et al., 1990; Oliver & Shapiro, 1995; Powell et al., 2001; Walters, 2001; Wilson, 1987).
State actors site schools and draw attendance zones that assist or hinder desegregation; they design and operate systems of ability grouping and tracking; they operate schools and school systems so as to permit middle class white parents to activate their race and class privileges. Outside the schools, state actors create and then sustain racially separate suburban residential neighborhoods, and their policies concentrate and isolate poor people of color in underserved public housing. State actors then employ residential proximity as the guiding principle for school attendance zones; they generate reforms-such as high-stakes testing-whose harsh accountability outcomes affect whites, blacks and Latinos disparately (cf. Haney, 2000; Heubert, 2000a, 2000b; McNeil, 2000; Valenzuela, 1998), in part because these state actors often fail to ensure equitable distribution of opportunities to learn the materials covered on the tests. Inequities in funding exist largely because state actors rely on property taxes to fund schools even though this method permits strikingly inequalities in resources, and hence, in opportunities to learn, based on race and class.
The policies and practices of school choice illustrate why the state shares responsible for creating and/or ameliorating racial gaps in achievement discussed previously. The state (members of legislatures, school boards, and other governing bodies) creates the enabling legislation, rules, and regulations that permit school choice-in all its manifestations. The literature on choice, race, and achievement is voluminous and contradictory (Fuller & Elmore, 1996; Godwin & Kermerer, 2002; Henig, 1994; Peterson & Hassell, 1998). According to advocates, the evidence shows that parental choice (vouchers, charters, and magnets) leads to greater gains for minority students; critics cite studies showing the opposite (Scott, in press). Aside from the unresolved issue regarding the effects of choice on individual students' outcomes, other important questions remain: What are the effects of choice on school systems? Does choice exacerbate or ameliorate racial inequalities in opportunities to learn within school systems? How does choice affect students whose parents do not actively choose a school or a program?
Choice theorists such as Chubb and Moe (1991) anticipate that choice will create positive consequences for schools and school systems. They hypothesize that good public schools ultimately will survive market forces unleashed by choice. To date, however, there is no consistent evidence that public or private school choice improves the achievement of students in choice schools, nor that it improves educational practice in schools not selected. Evidence from Chile (McEwan & Carnoy, 2000) and New Zealand (Fiske & Ladd, 2000) suggests that nationwide choice systems create disadvantages for students from ethnic minority and poor families. Perhaps market-based school reforms like choice will prove neutral with respect to the race gaps in school outcomes. Preliminary evidence does not sustain this position; rather, early evidence suggests the opposite (Fiske & Ladd, 2000; McEwan & Carnoy, 2000; Scott, in press).
THE LOCI OF DISCRIMIATION: JUNCTURES, TRANSITIONS, AND EDUCATIONAL CAREER TRAJECTORIES
Beyond obvious acts of overt individual racism and de jure institutional segregation, where does the responsibility of schools for racial discrimination begin? The previous review of social science explanations for the race gap suggests the loci of discrimination are in the junctures and transitions that contribute to students' educational career trajectories and the social dynamics that unfold at these points. How do we make sense of the organizational and institutional dynamics taking place at these loci? Lucas's (2001) model of effectively maintained inequality (EMI) regimes provides a useful guide for viewing the options, actions, structures, sequence, and dynamics where race discrimination occurs in education.
Students' own agency and racial identity are shaped by, and interact with, the culture and structure of the school. At each juncture and transition during K–12 education offers opportunities for students to take paths leading to very different educational trajectories. Students take different paths within each trajectory. The policies and practices by which students are sorted and selected for these paths, and the options students and their families perceive are critical to our understanding of how racial discrimination occurs and how its effects cumulate at each transition in students' educational careers.
Oakes (2002) suggests that we envision educational careers as processes taking place at the intersection of two axes. The x-axis consists of community forces and the school environment. The y-axis consists of the students' educational trajectories with their discrete junctures, at which critical transitions are made between and within schools; these are the conditions in which students are embedded and through which they move as they make critical transitions during their educational careers.
Community forces consist of the wealth and health of the neighborhood and the city, the suburb, or the rural area, its demographic composition, the collective educational history and culture of the community, and the nature of the family-school-community relationships. The school environment refers to the racial composition of the school, its material resources (e.g., funding) and human resources (e.g., access to qualified teachers), the academic press and other aspects of the school climate, and the organization of the school's opportunity structure (how the tracking and grouping practices operate).
Students' educational careers are conditioned by racial discrimination in housing and labor markets, and in public housing policies which result in isolated urban communities with high concentrations of low-income, poorly educated, minority families. Even though a majority of black and Hispanic students do not live in inner city ghettos and barrios, the majority of these students lives in urban areas and is increasingly likely to attend segregated minority public schools. At the same time, the suburbanization of middle class Americans, particularly whites-facilitated by tax and transportation policies-results in more privileged students attending newer, better resourced schools with few peers from low-income, poorly educated families.
Similarly, the decisions and choices students, families, and educators make are conditioned by the quality of the teaching staff, the material resources of the school, the racial composition of students and teachers in schools, staff and student attitudes and beliefs about race and ethnicity, the racial composition of ability groups and tracks, the academic press and the school culture, and whether the home-school relationship is inclusive and welcoming or exclusive and rebuffing.
In many respects, the trajectories of students' educational careers are launched with the critical transition from the home to kindergarten (or to preschool). At each subsequent juncture, discrimination can affect the next stage of the trajectory as racial disadvantage or racial privilege cumulate. Is safe, high quality, affordable childcare in a development preschool available in all communities? As the child moves from the home and preschool to kindergarten, is the public school one of high quality? How ready to learn is the child, and how does the school respond to this level of readiness? Which children are assigned to the certified teachers? Lee and Burkam (2002) and Lee et al. (2001) have shown that poorest, least prepared minority children systematically are assigned to the least prepared instructors in the poorest quality schools; this is one reason why the academic effects of developmental preschool, such as Head Start, do not endure throughout elementary school years (although results of High/Scope's Perry Preschool Project [Barnett, 1996, and Schweinhart, Barnes, & Weikart, 1993] suggest otherwise).
Once in elementary school, is the child identified as eligible for gifted or special education? Identification for either of these categories launches students onto paths that often become master statuses. Once a student enters a gifted track, the additional instruction and enriched curricula-not to mention the self-fulfilling dynamics of labels working through teachers' and parents' expectations-actually create a student who knows more and can do more than comparable students not in the gifted track. The opposite tends to be true for those in special education.
The gifted identification process is often racially discriminatory because of what schools do and do not do. Across the country, whites are more likely than blacks to be identified as gifted, and blacks are referred disproportionately to special education. This is true in part because schools permit savvy, privileged parents to deploy their resources on their child's behalf.10 In addition, schools selectively offer parents critical information about private testing options, or inform them that uncertified students still can take a gifted class. The schools continue to use identification procedures that do not capture the multiple forms of intelligence.
The transition to middle school is another critical junctures in students' educational careers. The academic tracks and courses into which students are placed exert a powerful effect on their present and future opportunities to learn. Middle school tracks determine the parameters of the formal curricula to which students are exposed, the stock of official knowledge they are likely to acquire, the peers with whom they interact, the test scores they will make, and the probable trajectory of their remaining education. Although parents may play an active role, the literature identifies how counselors, teachers, administrators have enormous influence over tracking, and how race and class forces influence the placement process. When we find that placements are correlated with race and social class, we should not be surprised.
Race and class forces also contribute to students' construction of their own social identities, including their identities as learners and as members of a particular race, gender, and class. The structure of opportunity in the school (are top tracks virtually all white while the lower ones are disproportionately black?) interacts with students' complex social identities. Among whites, the racial stratification of school structures signals their privilege; among minorities, it may cue oppositional attitudes or stereotype threat that contribute to racial discrimination in education. What and where students learn in a given year is reflected in part by their test scores and grades. A given year's academic performance then becomes part of the conditions for learning the following year. The effects of this dynamic cumulate during middle school careers and launch students onto disparate academic pathways in high school.
The transition from middle school to high school (or out of school) is another critical juncture. Again, the courses and tracks into which students are placed (a process in which counselors, teachers, and administrators continue to be influential) exert a powerful effect on opportunities to learn in high school. At this point in students' educational careers, stratified opportunities to learn within high school become gateways to the stratified outcomes that follow secondary school.
A CONCLUSION AND REFORMULATION
The core conceptual question guiding this article is "When are racial disparities in education the result of racial discrimination?" I have arrived at three answers. The first is a definition I derived from my review of the social science literature: racial discrimination in education arises from actions of individuals as state actors or institutions, attitudes and ideologies, or processes that systematically treat students from different racial/ethnic groups disparately and/or inequitably.
The second answer is that distinguishing racial discrimination from racial disparities is, in fact, an interesting intellectual, legal, and statistical challenge with possibly important policy implications. But in the everyday lives of black and white students (as well as Latino, Asian, Native American, and biracial youth), this distinction probably is less meaningful than social scientists and policy makers might hope. Distinguishing racial discrimination from racial disparities is akin to cleaning the air on one side of a screen door.11 In the end, the dirty air mingles with the clean, and harms respiration; in the case of education, the cumulative and interacting effects of racial disparities and racial discrimination ultimately harm children.
As the social science literature suggests, the racial gap is the result of complex, dynamic processes that cumulate over time. Most likely, racial disparities at T1 trigger racial discrimination at T2, and these phenomena generate further disparities, and so on. For example, because minority students tend to be concentrated in low performing schools, they are more likely than white students to spend larger portions of classroom time on "kill and drill" to raise their own and their school's standardized test scores. The corollary is true as well: They spend less time on broader and deeper curriculum coverage in social studies, science, and the arts, and in activities that develop higher order thinking skills. This is an unintended consequence of the intersection of high-stakes testing and school racial segregation: Minority students are more likely to receive fewer opportunities to develop higher order thinking skills or to be exposed to richer curricula. Imagine the racial differences in learning after a few years of "kill and drill" in lieu of the richer, deeper, and more engaging curriculum that test preparation displaces. Is it important whether this phenomenon is due to racial disparities or discrimination?
The third answer follows from the first two and is perhaps the most important because it attempts a conceptual leap. Recall, as I stated earlier, the central question I seek to answer is, "According to the social science literature, when are racial disparities in education the result of racial discrimination?" I propose the following reformulation of the question: "When are racial disparities in education not due to discrimination?"
To illustrate the value of this reformulation, let us consider two theories that have been advanced to explain the black-white race gap in educational outcomes: stereotype threat and oppositional cultural framework. Both theories make the agency of black students the core of black underperformance. I argue that racial discrimination in education, in fact, structures and conditions the exercise of that agency by black youths.
Under what historical conditions and
Roslyn Arlin Mickelson
Teachers College Record Volume 105 Number 6, 2003, p. 1052-1086
INDEX OF RESEARCH THAT COUNTS