Abstract
In this article, we study a set of factors underlying student success in a bottleneck business course using statistical and data mining techniques. Factors included learning styles, motivational and other cognitive factors, personality traits, learning analytics, along with background demographic and academic ones. Our analysis yielded interesting insights that show some of these factors play significant roles in predicting both student performance and their propensity to utilize resources that help improve their performance, such as additional support services. The predictive accuracy of both of our models were over 95% (error rate <5%). Moreover, quantile regression models were used to determine factors that specifically affect the performance of low-performing students so that targeted intervention and support services can be developed specifically for them. In conclusion, deeper analytics via statistical models are crucial for forming an in-depth understanding of how to improve student performance in a bottleneck course and this has far-reaching implications for both educators and administrators in higher education.