The construction, validation, analysis, evaluation and comparison of scales for predicting academic success in college

Date

1970

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

The purpose of this study was (1) to construct scales for predicting academic success during the freshman year at the University of Houston, using a technique which makes it possible to combine the predictive power of several generic variables into a single prediction; (2) to demonstrate, examine and evaluate the application of the method of regression equations to the data utilized in this study; (3) to validate, analyze, evaluate and compare the scales constructed in this study for predicting college academic success. The subjects for the study consisted of a total of 650 randomly selected freshmen from the University of Houston in the Fall of 1966 and Spring of 1967. The total sample group was divided into two groups: the basic or construction group comprised of 550 students (330 males and 220 females) and the validation group (60 males and 40 females). The students in both groups were designated as 'successful' or 'unsuccessful' according to two criteria: 2.0 and 1.6. The predictive variables identified for use in the present study were Total, Verbal, and Math scores, obtained on the Scholastic Aptitude Test and high school grade averages which were computed according to a quality point basis. A total of six 'Predictive Scales' were prepared on the basis of various combinations of these predictor variables and success criteria. The Glueck and Glueck technique of statistical methodology was followed in the construction of these scales which were cross-validated and found to be exceptionally powerful as predictive tools for academic success. For purposes of comparison, the technique of regression equations was applied to the data of this study for the basic sample of 550 students and the the 330 males only, and was found to be, generally speaking. Inferior to the methodology of Glueck and Glueck advocated in the present study. The most salient findings may be summarized as follows: 1. The combination of high school grades and aptitude test performance yield more valid indices of college performance than either method used alone, 2. The overall level of accuracy found In the present study was found to be 97 percent in the cross-validation of the original sample. 3. Prediction can be made most accurately for students with “Total Predictive Scores' falling in the extreme upper end of a distribution.

Description

Keywords

Prediction of scholastic success, Scale analysis (Psychology)

Citation