The predictive validity of traditional and nontraditional admissions measures for college performance in students grouped by sex, race, age, and academic risk

Lynn R Cook, Fordham University

Abstract

The predictive validity of standardized aptitude test scores and high school grades for college grades and attrition was compared to that of behaviorally based competency measures for eight student populations. The competency measures were hypothesized to discriminate less against minority students and to measure abilities not reflected in the traditional admissions measures which may contribute to college performance. The competency measures were also hypothesized to have potential in reducing the complexity of current models of college attrition.^ Subjects were 725 students from 16 educational institutions who were interviewed at college entry or early in their first semester. They were administered a questionnaire as part of the Student Potential Program (SPP) sponsored by the Council for Adult and Experiential Learning (CAEL) as well as interviewed. The interview technique, Behavioral Event Interview (BEI), was developed jointly by CAEL and a private firm for the purpose of eliciting recent behavioral information from which specific instances of eleven difference competencies are scored. Cumulative GPA by semester as well as enrollment data were collected over a two-year period.^ Results indicate that the traditional measures not only have predictive validity for college grades for majority and minority students alike but are also generally superior to the presently investigated competencies. Contrary to prediction, the traditional measures do not appear to be less valid in terms of predicting college performance for students differing from the typical white, young, male, low academic risk student. Competencies correlating highest with scholastic performance appear to reflect cognitive ability more validly assessed by the traditional measures. Although some traditional educational measures and competency measures were statistically correlated with attrition, the correlations were weak. Creativity, though not practically useful, appears to be related to college persistence. Failure of older students to score more highly than younger students on the competencies raises questions as to the construct validity of the current competency measures which are thought to be developed experientially. ^

Subject Area

Educational tests & measurements|School counseling

Recommended Citation

Cook, Lynn R, "The predictive validity of traditional and nontraditional admissions measures for college performance in students grouped by sex, race, age, and academic risk" (1990). ETD Collection for Fordham University. AAI9020008.
https://fordham.bepress.com/dissertations/AAI9020008

Share

COinS