Group Ability Test versus Teachers' Ratings for Predicting Achievement

Abstract

Standardized group tests of ability are sometimes considered biased when group means are unequal. A more accurate measure of bias is fairness in prediction. To assess whether a group ability test predicted achievement fairly, the association of the Verbal Ability score of the Cognitive Abilities Test and the Composite score of the Iowa Tests of Basic Skills was examined for four ethnic groups. The relationship between 25 teachers' ratings of students' ability and the same achievement measure was also examined for bias. Subjects were 403 third-graders (38 Asian-American, 84 African-American, 60 Hispanic, and 221 Caucasian). Multiple regression analyses showed the relationship between scores on the Cognitive Abilities Test and achievement was equivalent for all groups; however, teachers' ratings showed bias against Caucasian and Asian-American students by underpredicting their achievement scores.