The Coleman Report: Using Inappropriately Designed Achievement Tests

Abstract

The data in the Coleman Report have been interpreted by some as indicating that differences between schools have little impact on achievement. This interpretation is derived from the fact that only about 10% of the variance in the test scores was associated with differences between schools, while about 90% was associated with differences between individuals within schools. If the variance associated with school differences is appropriately compared to the variance associated with attending school one year, then school differences are quite large. It is not appropriate to draw conclusions about achievement when the conclusions are based on results of tests that were designed to maximize individual differences. The Coleman results make a great deal more sense when the test score results are interpreted as reflecting aptitude instead of achievement.