Forecasting Students’ Final Exam: Results Using Multiple Regression Analysis in an Undergraduate Business Statistics Course
Gunawardena Egodawatte *
Department of Global Management Studies, Ted Rogers School of Business Management, Ryerson University, 350, Victoria Street, Toronto, ON, Canada.
*Author to whom correspondence should be addressed.
Abstract
This paper discusses the development of a multiple regression model to predict the final examination marks of students in an undergraduate business statistics course. The marks of a sample of 366 students in the Winter 2017 semester were used to fit the regression model. The final model contained three predictor variables namely two test marks and the homework assignment mark. The marks of another 194 students from Winter 2018 were used to validate the model. The model validation showed that it can be used for future cohorts of students for prediction. The two main objectives of the study were to use the model as a teaching tool in class and to use the model to predict final examination marks of future students.
Keywords: Multiple regression, prediction, multicollinearity, forecasting, performance, statistics education, evaluation in higher education