Assessing the Validity of the Gre for Students in Psychology: A Validity Generalization Approach

Abstract

Meta-analysis was used to examine the validity of the Graduate Record Examination (GRE) in predicting "success" in graduate school. Results are presented for studies conducted in Psychology/Counseling departments. Validity coefficients were demonstrated to come from a homogeneous population (variance across studies was due mostly to sampling error). Mean validity coefficients ranged from .03 to .37, with an average of .18. Most of the squared population validity coefficients indicate that less than 9% of the variance in the criterion was accounted for. When employing graduate grade point average (GGPA) as the criterion, the GRE did not demonstrate adequate predictive validity. The problem of criteria is discussed, and alternatives to GGPA are presented as potentially superior indicators of graduate school success. Recommendations are suggested for developing a criterion which better reflects what is meant by "success" in graduate school.