Findings from educational research can have a major influence on public opinion and even on the outcomes of historic legal decisions. This fact was recently demonstrated by the Supreme Court's decision upholding affirmative action in law school admissions at the University of Michigan. The Court's opinion extensively cited a number of research studies to support their finding that diversity constituted a "compelling interest" for the university law school to use a consideration of race in their admissions policies. (1) Another arena in which social science research can have a significant influence in the public and judicial sphere is in K-12 schooling. Currently, school district leaders, policymakers and the public are still debating whether and in what circumstances race can be used in desegregating or creating racially and ethnically diverse schools. Research on the benefits of racially and ethnically diverse schools could help shape the direction and eventual outcome of that debate. Already in Federal Court cases in Seattle, Washington, and Lynn, Massachusetts, (2) judges have emphasized the importance of obtaining local data to support claims for the benefits that racial/ethnic diversity bring to educational settings. These claims are important because they could preserve the ability of school districts to maintain voluntary school desegregation plans, or to consider race in new student assignment plans.
Over the last half-century, many researchers have studied and written about school desegregation and race in American schools. Most studies on the benefits and costs to school desegregation are primarily from the 1960s and 1970s in response to the changes brought about from Brown, (3) the 1964 Civil Rights Act, and the Green (4) case in 1968--a decision that led to increased enforcement of Brown and authorized busing (Clotfelter, 2004, Hallinan, 1998; Orfield and Eaton, 1996). Moreover, the Swann decision in 1971 particularly influenced the desegregation of school districts in the South because the court ruled that previously segregated districts needed to balance their schools racially, even if that required cross-town busing to do so. (5) The early studies of desegregation concentrated largely on the impact of desegregated schooling on the experiences of African American students and focused on school systems that had been intentionally segregated. However, in the last twenty years there has been a resurgence of scholarly work on the subject, including several important reviews of the literature (e.g., Hallinan, 1998; Schofield, 1995; 2001; Wells & Crain, 1994; Dawkins & Braddock, 1994; Crain & Mahard, 1983) and a newer set of studies by several economists (e.g., Rumberger & Palardy, 2003; Hanushek et al., 2002; Rivkin, 2000; Boozer et al., 1992).
These studies, while much more recent, follow a traditional strand of desegregation research focusing on the impact of desegregated schooling environments on the academic progress of African American students, as measured by standardized test scores. Given the broad mission of public schools and the increasing diversity of today's school age population, it is critical to branch out from the traditional achievement view of benefits to diversity, and incorporate different and equally important outcomes to schooling into the literature. Such outcomes include the impact of diverse schooling environments on civic and democratic engagement, the ability or desire to live and work in diverse settings, and the degree to which schools equally support the academic progress of all students, regardless of race.
In addition, the changing demographics of the country (Reardon & Yun, 2001; Clotfelter, 1999) have led to an increase in the number of Hispanic K-12 students at the same time that Hispanic segregation continues to intensify in certain geographic areas (Frankenberg & Lee, 2003; Laosa, 2001; Valencia, 1991). Although schooling for Hispanic students is becoming more segregated, much of the desegregation literature does not speak to the unique circumstances of Hispanic students in segregated schools, many times not including them in the discussion of important schooling outcomes. Such omissions make it difficult to conceive of what constitutes a desegregated school system when there are three or more races attending schools in substantial numbers. As a result, in addition to expanding the discussion of the impact of desegregation beyond the traditional test based measures, including racial groups that have been largely absent from the discussion is essential. This is particularly important as the courts and school districts are interested in what the impacts of diversity are locally, where the presence of three or four racial/ethnic groups in large numbers is not uncommon.
In order to address these different outcomes and include other racial groups, we present findings from a case study of the Miami-Dade County Public Schools, a district that can legitimately be described as multiracial with three different racial/ethnic groups (black, white and Hispanic (6)) present in large numbers. Miami-Dade schools present an ideal place to explore the relationship between school racial composition and student outcomes for several reasons. First, the multiracial composition of the school district, which features a growing Hispanic population, and a declining white population, is a trend common among many urban school districts across the country. Second, the district has varying degrees of racial segregation present across its schools with some schools racially isolated, some schools with two or more racial groups present in large numbers, and some with at least three groups present in large numbers. Finally, the district, like many other urban school districts, has faced an important set of legal questions, public opinion and policy debates about the role of race in its student assignment plan.
Data and Instrument
The instrument used in this research, the Diversity Assessment Questionnaire (DAQ), is a 70-item student survey. The instrument was developed by researchers at The Civil Rights Project at Harvard University in collaboration with the National School Boards Association's Council of Urban Boards of Education. It was designed to be a classroom-administered questionnaire that asks students about their experiences in their school and classrooms. The survey--targeted for 11th grade students, but suitable for any high school grade--also includes questions about students' future goals, educational aspirations, attitudes and interests. (7) The instrument was piloted in Louisville, KY (Kurlaender & Yun, 2001) and has since been administered in seven school districts around the country.
The data employed for this analysis are based on the survey administered in the fall of 2000 to 10,844 students, attending 33 high schools throughout the Miami-Dade County Public Schools. The survey response rate varies by school, but overall, 49% of all 11th grade students enrolled in a Miami-Dade County public high school completed the DAQ survey. The dataset consists of background information regarding students' racial/ethnic status, gender, whether they were born in the United States, parents' education, and primary language spoken at home. The racial/ethnic classifications used in this analysis are Black, Hispanic, (8) White, and Other. (9)
In addition to individual-level survey data, we obtained school-level data, including school racial composition, percent free and reduced lunch, percent limited English proficient, and percent in special education. Additional descriptive information about the sample is included in the Appendix. Despite the wide range of response rates across schools, and potential selection bias (10) the shear size of the sample and the importance of the district--the fourth largest in the country--provide us with preliminary evidence supporting the impact of school racial composition on these outcomes. (11)
We separate the paper into three sections each focusing on a separate outcome--desire to live and work in multiracial settings, citizenship, and support for educational attainment. We present the results from our descriptive analyses in several ways. First, we describe the outcome and provide a short review of the relevant literature. Next, we provide an exemplar set of questions from the survey that represent each of the respective outcomes. Within each question, we disaggregate the data by the race of the student respondent and the racial/ethnic composition of the school. We adopt three categories of school composition to examine the data--multiracial, Black-Hispanic, and racially isolated schools. We define multiracial schools as those schools that enroll at least 20% of the three major racial groups (black, white and Hispanic). Racially isolated schools are those with at least one race exceeding an 80% share of the school population. (12) Finally, Black-Hispanic schools were those which did not fit into either of the other categories, these schools are more multiracial than racially isolated schools and had at least 30% each of black and Hispanic students and fewer than 80% of either. These schools occupy an interesting middle ground of schools which are not wholly segregated but do not enroll many white students, here minority students interact with each other, possibly garnering some of the benefits that may accrue to more traditionally defined multiracial schools.
To further explore the relationship between school racial composition and student outcomes, we plot the mean values of each outcome by race and by school, sorted by the percentage of white students enrolled. (13) By examining these figures we may be able to discern patterns in the data that relate to the school compositions, and will be able to identify schools that have either very high or low values on that particular composite, or schools that show large...