Examining student retention dynamics in 4-year engineering and computer science programs using random forest analysis

Background

Despite the vital function of engineering and computer science (Eng & CS) to innovation and economic development, retention within Eng & CS programs remains a major challenge in the U.S. educational system. Despite extensive research on influential factors of retention, there is a gap in our understanding of how these factors rank in relative importance and their temporal variations.

Purpose

This study aims to identify and compare the factors and demographic variables that predict retention in Eng & CS programs at two critical timepoints: from the first to the third year, and from the third year to completion within 6 years.

Method

The study applied random forest analyses to a longitudinal dataset from the nationally representative BPS 12/17 survey.

Results

High school academic preparation is the most important predictor of both 3-year and 6-year retention. Psychological factors are more crucial to retention in the early stages, while a more diverse, practical, career-oriented set of factors gains importance when it comes to the 6-year retention period.

Conclusions

The study advocates for stage-wise interventions, such as enhanced academic preparation programs and robust support networks in the early stages, and career services and practical support systems as students advance. The findings also highlight the importance of peer support networks and tailored support from institutions.