A comparison of ethnic, sex and age bias in the predictive validity of English language aptitude tests: some Israeli data

Abstract

The primary aim of this study was to test for ethnic, sex, and age bias in the pre dictive validity of English language aptitude test scores, routinely employed for student selection and placement in the Israeli scene. The analyses were based on the English language aptitude test scores of 824 full time Jewish students studying at a major Northern Israeli university. For data analytic purposes, students were partitioned by ethnicity (Western/Oriental), sex (males/females), and age group membership (18-21, 22-25, 26-29, 30+). English language aptitude test scores, as predictors, evidenced both slope and intercept bias by ethnicity, tending signif icantly to overpredict the first year cumulative grade point average of Oriental students and underpredict that of their Western counterparts. Furthermore, test scores were observed to evidence a small degree of intercept bias by gender, tend ing slightly to overpredict the first year grade point average of males and under predict that of females. Also, test scores are shown to be slightly less valid predic tors of criterion performance for students in the oldest age group (30+ ) when compared to students in the remaining age groups. On the whole, the data are consistent with previous pointing to intercept bias in aptitude test scores by gender and ethnicity. Furthermore, the data are consistent with some previous research attesting to the differential predictive validity of aptitude test scores by age group. The implications of the study for college admission testing policy and bias research are discussed and explicated.