Why college grade point average is difficult to predict

Many errors in the selection of college students are inversely related to the validity of the predictors employed. The generally weak validity with which GPA has been predicted gives rise to a substantial number of selection errors, including those that support charges of cultural bias. This situati...

Full description

Saved in:
Bibliographic Details
Published in:Journal of educational psychology 1976-02, Vol.68 (1), p.9-14
Main Authors: Goldman, Roy D, Slaughter, Robert E
Format: Article
Language:English
Subjects:
Academic Achievement
College Students
Cultural Test Bias
Educational Measures
Grading (Educational)
Human
Statistical Validity
Student Admission Criteria
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Many errors in the selection of college students are inversely related to the validity of the predictors employed. The generally weak validity with which GPA has been predicted gives rise to a substantial number of selection errors, including those that support charges of cultural bias. This situation has produced criticisms of standardized tests and a search for alternative models of fair selection. The present investigation with 254 undergraduates suggests that the problem has been misinterpreted. It is a criterion problem rather than a predictor problem. The investigation demonstrates high validity for predicting grades in single classes. Since single-class grades are components in the GPA composite, it seems paradoxical that GPA prediction is inferior to grade prediction in single classes. The explanation for this may reside in different standards of grading in different classes. Implications and remedies are also discussed.
ISSN:0022-0663
1939-2176
DOI:10.1037/0022-0663.68.1.9