Determinants of student course withdrawals.

Abstract

This study investigates the determinants of a student's decision to withdraw from specific courses in higher education. The study was conducted at Middle Tennessee State University, Murfreesboro, Tennessee, using data from the Fall and Spring semester of the 1990-1991 academic year of four principles courses from four departments.

A simple model of course withdrawal is introduced. It consists of three groups of explanatory variables: student attributes, pre-semester background, and class attribute variables. Interactions among these variables determine the level of satisfaction the student obtains from the course which, in turn, determines the decision to withdraw. The model is tested using the logit technique. For the purpose of comparison, however, ordinary least squares (OLS) and probit results are also presented. The predictive power of the model is measured by count-R{dollar}\sp2{dollar}, and to determine whether the model is replicable, pseudo-R{dollar}\sp2{dollar} is used.

Overall, the models fit well for both the Fall 1990 and the Spring 1991 cohorts. But individually, only several variables show significant relationships with course withdrawal.

The results of the study indicate that the older the student, the less likely he or she is to withdraw from the currently taken course. This finding, however, contradicts to a previous study by Adams and Becker. It is also found that students who attempt less credits are more likely to withdraw from any given course, which is contrary to what is expected.

The number of hours completed prior to the semester tends to be positively correlated with withdrawals. Thus, students with more experience in the sense of having completed more credit hours are more likely to withdraw. In addition, it is also found that students who have tended to persist in the past are less likely to withdraw. Finally, this study reveals that students who like to withdraw from courses tend to have lower grade point averages.