2023 ASEE Annual Conference & Exposition

Considering the increasing demand for engineering graduates, understanding what is limiting students from completing their degrees has been a consistent question posed in the literature. The nontrivial variance in pathways students take in obtaining an engineering degree, especially in cases where students abandon their studies, suggests that longitudinal datasets can hold a wealth of information to uncover factors contributing to attrition. Accordingly, this project uses existing data to explore curricular factors that create barriers for different students by leveraging a new framework for quantifying the impact of such factors, Curricular Analytics. Curricular Analytics uses network analysis to measure sequencing and interconnectedness in a plan of study. There are two intrinsic measurements associated with each course in Curricular Analytics: (1) the blocking factor, which is a count of how many courses are inaccessible to a student upon failing and (2) the delay factor, the longest prerequisite chain through the course. The sum of these factors for all courses in a plan of study, called structural complexity, characterizes a measure of the curriculum’s complexity. This project combines Curricular Analytics with course-taking data from the Multiple-Institution Database for Investigating Engineering Longitudinal Development (MIDFIELD).

Curricular Analytics is a relatively new framework with the potential to address existing research questions and generate new ones. To date, curricular complexity has been used to correlate program quality with a curriculum’s structural complexity and predict four, five, and six-year graduation rates for first-time-in-college and transfer students. This paper/poster reports on the year one activities for this project, which addresses the potential of the framework to bring a renewed perspective to MIDFIELD dataset by creating a new public dataset of curricular information for Civil Engineering, Electrical Engineering, Mechanical Engineering, Chemical Engineering, and Industrial Engineering at 13 MIDFIELD institutions with data current up to 2015. For each institution, we are collecting the previous 10 years of curricular data since the institution’s last record in the appropriate format for network analysis. This results in an upper bound of 650 networks; considering some institutions do not offer all five disciplines of interest, our dataset contains 535 networks. We are exploring these networks longitudinally within and between disciplines and institutions.

The next steps of this project involve creating course-taking trajectories from the course table in MIDFIELD using association analysis. The characteristics for these trajectories, including structural complexity, retaking behavior, and major switching, will be clustered and disaggregated across strata, such as first-time-in-college versus transfer, race, gender, and first-generation status. We plan to correlate structural complexity with ecosystem metrics like discipline stickiness and migration yield. By disseminating these results directly to institutional stakeholders and the broader engineering education community, we anticipate that this project can impact curricular design for all engineering students and help us understand what academic policies are inhibiting degree attainment for diverse groups.