Predicting success in the College of Engineering, University of Houston



Journal Title

Journal ISSN

Volume Title



The purpose of this study was to determine how accurately academic success of engineering students at the University of Houston can be predicted, from a knowledge of their High School grade point average, and their scores obtained on tests that make up the freshman guidance battery. The battery of standardized tests considered in this study included (1) The American Council on Education Psychological Examination, 1947 College Edition; (2) Cooperative Inter-American Reading Test — part 2 only; (3) Cooperative English Test Mechanics of Expression, forms Z & S; (U) Math. Screening Test, University of Houston; and (5) Kuder Preference Record, Form CM. The criterion used was the grade point average for the first semester's work of the recent high school graduates that entered the University of Houston's Engineering College in the Fall of 1955. The Pearson product moment method was used to find, the intercorrelation between the criterion and each variable and between each variable and all other variables. The Wherry-Doolittle test selection method was used to determine the multiple correlation and the beta coefficients of the various tests. It was found that the maximum predictive value of variables is obtained with the high school grade point averages and the scores provided on two of the tests in the battery. Listed in the order of their relative contribution these are: High school grade point averages, Math. Screening and ACE Psychological Examination score. The beta coefficients, listed in the same order, were found to be .54 .03, and .02. The shrunken coefficient of multiple correlation was found to be .63. It is obvious that no single variable is a very good means of predicting success in the college of engineering. However, when optimum weights are applied to three of the scores, a substantial reduction in errors of prediction is obtained.



Prediction of scholastic success.