2018 ASEE Annual Conference & Exposition

This research paper presents findings from a study of the relationship among students’ ongoing academic performance and their co-enrollment in multiple courses during an academic semester. We focus on the potential hazards created by different patterns of concurrent enrollment. Specifically, we model the risk of students’ experiencing academic difficulty and their probability of recovering from academic difficulty given their week to week academic performance in their other coursework.

Our analysis uses weekly academic classifications in an early warning system (EWS) for students in an undergraduate engineering course at a research-intensive university in the Midwest. The EWS gives a weekly categorization of each student’s performance for each course, either “Encourage,” “Explore,” or “Engage,” based on various metrics including: gradebook data, students’ interaction with online course tools and materials, and students’ performances when compared to their peers in the course.

We acquired weekly academic classifications for all students enrolled in one computer engineering course during the Fall 2016 academic semester. Additionally, we collected data from all other courses in which these students were enrolled to examine the impact of experiencing academic difficulty on students’ academic success in their other courses during the semester. We used multilevel event history methods on students’ performance data from the EWS to answer the following research questions:

RQ1) Does experiencing academic difficulty in one course significantly increase students’ odds of experiencing academic difficulty in any of their other courses during the semester?
RQ2) What is the likelihood of students’ recovery from academic difficulty (i.e., moving from an “explore” or “engage” status to an “encourage” status) in one course during the semester? Two courses? Three or more courses?

Initial findings indicate that students who experience academic difficulties in any of their courses are over two times more likely to also experience academic difficulty in at least one of their other enrolled courses. Additionally, the time that it takes to recover from their academic difficulty has a profound impact on the likelihood of recovery. For example, after the fourth week in the “explore” and “engage” models, students’ probability of exiting either of these classifications drops by approximately 50%. For those students who experience academic difficulties in two courses, their likelihood of recovery in both courses is approximately 30%. For those who experience academic difficulties in three or more courses, this percentage drops to less than 10%. Academic difficulty in one course often precedes a snowball effect, whereby students experience increased odds of academic difficulty across their courses and decreased odds of recovery. The full paper will expand upon the results found in this research study, discuss the rest of the analyses from our full statistical model, and suggest how analyses like these can help us understand the conditions for student recovery and provide personalized flags for intervention.