Cook, Michael, & Evans, William
Families or Schools? Explaining the Convergence in White and Black Academic Performance
University of Maryland
What has caused the convergence in White and Black academic performance over the past twenty years? Are changes in families or schools responsible?
Journal Name or Institutional Affiliation:
Journal of Labor Economics
Vol. 18, No. 4, pp. 729-754
- Family background and school characteristics explain 25% of the convergence in test scores.
- Most of the convergence 75% is due to narrowing of the gap in test scores between White and Black students who attend the same school and have the same level of parental education.
- Changes in school quality account for little or none of the convergence in test scores.
- Changes in the parental education level account for small part of the convergence and estimates for changes in family income and family structure do not explain because there is no convergence in income or family structure.
- The gap in the quality of schools attended by Blacks and Whites has widened due to the decline in quality of poorer, inner city, and predominantly Black schools.
- By 1988, the gap in White and Black test scores was 6.22 percentage points lower due to changes within schools, and this change accounted for 83% of the overall convergence in test scores.
- In 1970, White reading scores were 3.32 percentage points higher than Black scores due to differences in school quality. By 1988, this difference rose to 4.15, implying that the schools that Blacks attended worsened relative to the schools that whites attended. Thus, changes in school quality explain a negative amount of the convergence in reading test scores.
- Although changes in school quality explain little of the convergence in White/Black test scores, the variation in test scores explained by schools is quite high.
- Three-quarters or more of the convergence in test scores was due to a narrowing of the gap in test scores among white and Black students who attend the same school and with the same level of parental education.
- Although there was a sharp convergence in the levels of parental education, this accounts for only 22%-30% of the overall convergence in test scores.
- Changes in the relative quality of schools can account for little or none of the convergence in test scores. In fact, in the analysis of the NAEP reading tests, we found a divergence in the relative quality of schools that White and Black students attend, while the results for NAEP math tests show that changes in school quality can account for only one-eighth of the convergence in test scores.
Journal Article Empirical Research
Academic Achievement, Achievement Gap, Family, Math, Reading, SES, School Quality
Secondary Survey Data
Method of Analysis:
Unit of Analysis:
- Study analyzes data from the National Assessment of Educational Progress (NAEP), most cited test score data on trends in student performance over time.
- Study uses the trend samples, stratified probability sample representative of all 13 year olds enrolled in public or private schools in a given year.
- Uses primarily data on test scores for 13 year olds in 1970 and 1988.
- Also estimates the decomposition for 17 year olds.
- The data include test scores (reading and math) for Black and White students, and highest level of education for their parents and measures of school quality.
- DV: Test scores
- IV: Demographic characteristics, parental education, race, school fixed effect.
- Multiple observations per school allow for the inclusion of school-fixed effects into the test score equation.