Twenty-First Century Social Science on School Racial Diversity and Educational Outcomes

  1. The preponderance of social science since the late 1980s indicates a negative relationship between minority concentration and educational outcomes.
  2. Social science research conducted in the 1960s, 1970s and 1980s differs from the studies conducted since 1990 in a number of critical ways. Author finds that the later studies suffer from fewer threats to their internal and external validity compared to the earlier studies. Many of the earlier experimental studies suffer from sample attrition, nonrandom assignment to experimental and control conditions, weak measures of key constructs or incomplete or inappropriate implementation of the desegregation treatment. The earlier studies also tended to have small samples and attempted to evaluate desegregation policies one or two years after implementation.
  3. Better statistical methods coupled with better data quality in the more recent studies enhance the ability of social scientists to isolate the role of school racial composition on outcomes apart from other influential factors like school SES composition, student characteristics and family socioeconomic status.
  4. Authors of amicus briefs cited different social science research. The pro-desegregation respondents’ briefs referred to many more recent research studies than the petitioners’ briefs. The brief of the APA cited fifty-seven studies, the AERA brief cited fifty-four studies, and the 553 social scientists used 213 sources to support their claims that diverse schools were valuable for student outcomes. The overwhelming majority of these studies were disseminated since 1990. Forty-seven percent of the APA’s studies were published since the 2000s with 80 percent appearing since the 1990s. Similarly with the AERA and 553 Social Scientists briefs which cited 79% since 1990 and 77% since 1990 respectively.
  5. In contrast, petitioners briefs by Murphy et. al. cited twenty studies and Amor et. al. cited twenty-nine studies, a majority of which was disseminated prior to 1990. Petitioners’ briefs are dominated by social science conducted in the 1970s and 1980s.
  6. Amicus briefs that rely on incomplete summaries of social science and/or older, weaker social science syntheses cannot adequately inform legal opinions.