School Desegregation and Black Achievement

Meta-analytic techniques were used to explore overall conclusions and variables moderating treatment effects in the research literature on school desegregation and black achievement. Studies were classified on the basis of the threats to their validity as either accepted or rejected for the analysis. For the initial analysis quasi-experimental studies were accepted, yielding an average effect size of .45. The better-designed studies had an average effect size of .34, which was reduced to .16 when adjusted for pretest differences. The National Institute of Education (NIE) convened an expert panel that reviewed and reanalyzed these results. An average pretest-adjusted effect size of .14 was found for the 19 studies selected for analysis by the NIE panel. An average effect size of .20 was found for the better-designed studies that had no selection problems. This is equivalent to two months of educational gain. The largest effects occurred among students moving from highly segregated to predominantly white schools. Reading achievement gains were larger than those for mathematics, but the difference was not statistically significant.