Crain, Robert L., & Mahard, Rita
The Effect of Research Methodology on Desegregation - Achievement Studies: A Meta-Analysis
Johns Hopkins University; Rand Corporation
Determine whether differences in findings of the effects of school desegregation on black achievement can be attributed to differences in methods.
Journal Name or Institutional Affiliation:
The American Journal of Sociology
Vol. 88, No.5, pp.839-854
- Grade of desegregation is significantly related to both magnitude and the sign of the treatment effect.
- Longitudinal randomized experiment always has the largest positive coefficient, and the longitudinal design with justified control group has the second largest.
- A positive treatment effect can be identified.
- Much of the discrepancy among studies of desegregation and achievement can be attributed to methodological differences.
- Desegregation is not a treatment and shouldn't be studied as one.
- Focus should be on the curriculum and community reactions.
Journal Article Empirical Research
Academic Achievement, Communities, Curriculum, Desegregation
Method of Analysis:
Studies on the effects of desegregation on blacks
Unit of Analysis:
- 93 studies conducted in 67 cities on the effects of desegregation on Black students
- Sample limited to studies conducted in desegregated school districts.
- 323 samples of students were identified.
- Discarded studies that were: (a) simple cross-sectional comparisons of test scores of blacks in segregated and desegregated schools, and there was no reason to believe that the two groups were similar (b) studies in which it was not clear that desegregation had occurred by government action.
- Divided all study methodologies into seven types, based on the type of control group comparisons; the designs, ranked from the strongest control group to weakest.
- Divided in 7 types according to their design
- 1) longitudinal, random assignment
- 2) longitudinal, with justified Black control group
- 3) longitudinal, non random controls
- 4) cross-sectional, Black controls
- 5) cohort comparisons
- 6) longitudinal, White control groups
- 7) longitudinal, Test norm sample Comparison
- Coded type of design, direction of effect, dates of desegregation
- DV: Size and direction of effect
- IV: Grade at desegregation, type of design, constant.