Investigating Engineering Persistence through Expectancy Value Theory and Machine Learning Techniques

Campbell Bego; Pamela Thomas; Xiaomei Wang; Arinan Dourado

Download Paper | Permalink

Conference: 2022 ASEE Annual Conference & Exposition
Location: Minneapolis, MN
Publication Date: August 23, 2022
Start Date: June 26, 2022
End Date: June 29, 2022
Conference Session: First-Year Programs Division Technical Session 10: Best of First-Year Programs Division
Page Count: 16
DOI: 10.18260/1-2--41559
Permanent URL: https://peer.asee.org/41559
Download Count: 360

Request a correction

Paper Authors

biography

Campbell Bego University of Louisville

visit author page

Dr. Campbell Bego researches engineering learning and persistence as an Assistant Professor in the Department of Engineering Fundamentals at the University of Louisville's Speed School of Engineering. Prior to entering academia, she obtained a BS in Mechanical Engineering from Columbia University, worked in tunnel ventilation (CFD modeling) at Mott MacDonald and AECOM, and received a Professional Engineering license in the State of New York. She draws on these experiences as well as her MS and PhD in Cognitive Science from the University of Louisville to construct meaningful activities in her first-year engineering course. She aims to improve the number of engineering graduates as well as the quality and diversity of the engineering workforce using evidence-based practices and applied theory in the classroom.

visit author page

biography

Pamela Thomas

visit author page

Assistant Professor at the University of Louisville

visit author page

author page

Xiaomei Wang Texas A&M University

author page

Arinan Dourado University of Louisville

Download Paper | Permalink

Abstract

A vast amount of research has revealed that students decide to leave engineering for numerous reasons, including bias and discrimination, low performance in mathematics, belonging uncertainty, financial load, and individual interest. One possible framework to investigate students’ persistence is Situated Expectancy Value Theory (SEVT), which proposes that students’ achievement-related choices and performance are influenced by expectations of success, and subjective task value. In this contribution, we combined SEVT with data science tools to better understand students’ academic decision-making, specifically the decision to persist in engineering. Our goal was to identify underlying patterns that predict persistence as well as illuminate possible interventions. Here, three machine learning tools, namely, clustering, principal component analysis (PCA), and decision trees, were applied to data from two cohorts of engineering students at a large public university. Concerning the SEVT framework, student responses to surveys given at the beginning and end of the first semester, containing established scales for self-efficacy and contingencies of academic competence self-worth (expectancies), and interest in engineering and perceived costs of studying engineering (subjective task values) were used. Demographic data including race, gender, and Pell eligibility, alongside performance data in the form of introductory course grades, GPA, and persistence into Year 2, complete the set of gathered information available to our data analyses. Collectively, we were interested in learning patterns that allowed us to use academic performance and SEVT data to predict engineering retention. Supporting previous findings in the literature, results from clustering analyses and a PCA indicate that performance in the first-semester engineering courses is highly predictive of persistence in engineering. Potential interventions include mid-first-semester feedback and learning interventions, as well as study habits and time management. New findings indicate that the SEVT framework also significantly predicts persistence, especially for mid-level performers. Discussions in the second semester with students are proposed for both research purposes as well as an intervention to improve student persistence.

Citation
Format

Bego, C., & Thomas, P., & Wang, X., & Dourado, A. (2022, August), Investigating Engineering Persistence through Expectancy Value Theory and Machine Learning Techniques Paper presented at 2022 ASEE Annual Conference & Exposition, Minneapolis, MN. 10.18260/1-2--41559

TY  - CPAPER
AB  - A vast amount of research has revealed that students decide to leave engineering for numerous reasons, including bias and discrimination, low performance in mathematics, belonging uncertainty, financial load, and individual interest. One possible framework to investigate students’ persistence is Situated Expectancy Value Theory (SEVT), which proposes that students’ achievement-related choices and performance are influenced by expectations of success, and subjective task value. In this contribution, we combined SEVT with data science tools to better understand students’ academic decision-making, specifically the decision to persist in engineering. Our goal was to identify underlying patterns that predict persistence as well as illuminate possible interventions. Here, three machine learning tools, namely, clustering, principal component analysis (PCA), and decision trees, were applied to data from two cohorts of engineering students at a large public university. Concerning the SEVT framework, student responses to surveys given at the beginning and end of the first semester, containing established scales for self-efficacy and contingencies of academic competence self-worth (expectancies), and interest in engineering and perceived costs of studying engineering (subjective task values) were used. Demographic data including race, gender, and Pell eligibility, alongside performance data in the form of introductory course grades, GPA, and persistence into Year 2, complete the set of gathered information available to our data analyses. Collectively, we were interested in learning patterns that allowed us to use academic performance and SEVT data to predict engineering retention. Supporting previous findings in the literature, results from clustering analyses and a PCA indicate that performance in the first-semester engineering courses is highly predictive of persistence in engineering. Potential interventions include mid-first-semester feedback and learning interventions, as well as study habits and time management. New findings indicate that the SEVT framework also significantly predicts persistence, especially for mid-level performers. Discussions in the second semester with students are proposed for both research purposes as well as an intervention to improve student persistence.
AU  - Campbell Bego
AU  - Pamela Thomas
AU  - Xiaomei Wang
AU  - Arinan Dourado
CY  - Minneapolis, MN
DA  - 2022/08/23
PB  - ASEE Conferences
TI  - Investigating Engineering Persistence through Expectancy Value Theory and Machine Learning Techniques
UR  - https://peer.asee.org/41559
DO  - 10.18260/1-2--41559
ER  -