Prediction of Undergraduate Achievement in Economics

Abstract

The study employed 18 variables selected from the academic records of students which were available prior to their junior year in college. The sample used in the analysis involved 92 B.A. degree candidates in economics, class of 1970, at the University of Massachusetts. The criterion variable was the grade point average of the advanced courses taken in economics in the junior and senior years of college. In order to discover the relationship between the independent variables and the dependent variable, the computer program B M O 3 R [1] was utilized to calculate the correlation coefficients. The results are presented in Table 1. The freshman and sophomore GPA had the highest correlation of the 18 independent variables. Most variables involving verbal ability had a slightly higher correlation to the criterion than those variables involving quantitative ability. In the second part of the study, a stepwise computer program B M D O 2 R [1] was used to develop a prediction equation with a significant correlation coefficient (R) using the least number of variables. It was concluded that the regression equation with two variables was the most practical and useful for predictive purposes. The prediction equation for the best combination of two variables is given below: