We investigate business undergraduate mathematics-based courses in a blended environment of
online assignments and exams and offline lectures, and report the impact on academic
performance of factors such as classroom attendance, web-based course supplements, and
homework. We present results from both ordinary least squares and fixed effects, where the latter
method controls for unobserved heterogeneity among students. We discuss biases in estimation
when the ordinary least squares method is used, resulting from the fact that it ignores unobserved
heterogeneity. The fixed effects results suggest that (1) class attendance has a positive impact on
exam score, (2) a student who achieves proficiency in a greater number of Khan Academy skillsets to prepare for an exam takes longer to complete an exam but does not experience a
significant change in exam score, (3) a student who spends more time completing the homework
spends more time completing the exam but does not experience a significant change in exam
score, and (4) students who score relatively higher in homework tend to score relatively higher in
exams and finish in less time than other students.
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education