Accountancy undergraduate performance: a statistical model

An understanding of the factors that determine a student's final degree classification before the commencement of an honours year is important to many parties. This paper develops a statistical model using logistic regression to predict whether a student will achieve at least an upper second class honours degree in his/her final year. Data were collected from the population of honours graduates of 1998 and 1999 (n = 79) on a thick sandwich Accounting and Finance degree at the Robert Gordon University in Scotland. Information was extracted from student files to form 35 exploratory variables. Statistical analysis was undertaken to identify those variables which appeared to differentiate between students who achieved at least an upper second class degree and those who did not. The statistically significant variables were then entered into a logistic regression model using a forward stepwise approach. The resulting model contained three variables, performance in Financial Accounting in year 2, performance in Managerial Accounting in year 2, and the number of resits in year 1. The diagnostic statistics and classification accuracy suggest a robust model, which suffers little shrinkage when applied to the holdout sample of 43 students who graduated in 2000. The paper concludes with recommendations for counselling students regarding the appropriateness of undertaking an honours degree.