Students' study activities before and after exam deadlines as predictors of performance in STEM courses: A multi-source data analysis

Abstract

Many college students struggle with regulating the time and effort they invest in classes. We used digital trace data from a learning management system to examine students' behavioral engagement and associations with course performance in four chemistry courses (N = 1596). Results from Study 1a show that behavioral engagement declined across the course, except for high spikes in exam weeks. Students with higher regularity and continued engagement after midterm exams obtained higher course grades, whereas steep increases in study activities shortly before exams did not predict performance. Using a selective subsample of students (n = 51, with 510 observations over time) who identified chemistry as a challenging course, Study 1b explores whether intentions to regulate learning behaviors with goal-directed control strategies lead to changes in behavioral engagement. Intentions to use control strategies lead to short-term changes in behavioral engagement, but students did not implement planned adjustments to their study behaviors in the long run.

Educational relevance statement

This study shows that consistent behavioral engagement in a learning management system over the course of a semester and early increases in learning activities before critical course exams predicted students' academic success in chemistry college courses. Students showed increased behavioral engagement immediately before course exams, but such short-term increases did not lead to better course performance. Instead, regular course engagement, as indicated by click activity in a learning management system, was significantly related to students' end-of-term course performance. Findings from a small and selective subsample of students who perceived the course as particularly challenging (study 1b) further suggest that students' intentions to change their behavioral engagement for the following exam(s) predicted only short-term changes in observed engagement in the learning management system. Thus, these students might benefit from further support to effectively regulate their learning behaviors. Studies 1a and 1b suggest that digital trace data from the course's learning management system can be informative in identifying struggling students, particularly using trace data from weeks around exams.