Heads Or Tails (Success Or Failure)? Using Logit Modeling To Predict Student Retention And Progression

Abstract

Using a sample of 2,137 university students and applying the logit model, we find that the probability for students to return in fall 2008 is higher with a higher cumulative GPA, a higher grade for SE 101, and a returning status in the previous semester. Several other explanatory variables are tested and have insignificant coefficients. A few variables such as the Board of Regent’s core requirements (CORE) and high school graduating GPA (HSGPA) have the expected signs and z-statistics closer to one, suggesting that the correlation coefficient may rise if the sample size were larger. The findings suggest that the cumulative GPA is a dominant factor and that the large number of failures in SE 101 may need to be examined in order to fulfill its described purpose: “a course designed to ensure first-year student success.”