Quantifying the Impact of Students' Semester Course Load on Their Academic Performance

Shahab Boumi; Adan Ernesto Vela

Download Paper | Permalink

Conference: 2021 ASEE Virtual Annual Conference Content Access
Location: Virtual Conference
Publication Date: July 26, 2021
Start Date: July 26, 2021
End Date: July 19, 2022
Conference Session: Computing and Information Technology Division Poster Session
Tagged Division: Computing and Information Technology
Page Count: 12
DOI: 10.18260/1-2--37630
Permanent URL: https://peer.asee.org/37630
Download Count: 1762

Paper Authors

biography

Shahab Boumi University of Central Florida

visit author page

Shahab Boumi is a Ph.D. student in the department of Industrial Engineering and Management Systems (IEMS) at the University of Central Florida. His main research focus on investigating students behavioral hidden patterns in large scale data sets using supervised/unsupervised machine learning methods, stochastic processes, and optimization tools.

visit author page

author page

Adan Ernesto Vela University of Central Florida

Download Paper | Permalink

Abstract

Students' academic success in science, technology, engineering, and mathematics (STEM) careers is one of the most popular subjects that has gained attention among educational researchers for decades. Many studies have shown students' educational outcomes can be affected by academic factors including high school GPA, SAT score test, student admission type (transfer or first-time-in-college), as well as demographic features such as gender, ethnicity, and family income. Additional studies have investigated the relationship between students' course load and their academic outcomes. In this paper, we define students' course load based on the number of courses they take each semester, which is assumed to have a discrete probability distribution. To assess if students' course load impacts their academic performance, we apply Hidden Markov models, an unsupervised learning method, to classify students into three categories: high-level enrollment, medium-level enrollment, and low-level enrollment. The sequence of the number of courses students enroll each semester during their academic career is fed into our proposed model as input. The output which is a qualitative measure and is not directly observable is the estimated enrollment level for the students. After students' classification, we derive and compare their academic (e.g., cumulative GPA, graduation rate, and DFW rate) and non-academic (e.g., family income level) features for each enrollment level category. Findings show that students who have more engagement with the university have higher academic performance (higher cumulative GPA and graduation, lower DFW rate) than those with lower engagement. Our analysis also demonstrates that students from families with low-income levels are more likely to have lower enrollment levels. Such results indicate that university managers can improve students' educational performance and, subsequently, the university graduation rate by encouraging students to engage more with the university by providing academic and financial support. These results are based on the data collected from the University of XXX from 2008 to 2016 and contain approximately 170,000 students.

Citation
Format

Boumi, S., & Vela, A. E. (2021, July), Quantifying the Impact of Students' Semester Course Load on Their Academic Performance Paper presented at 2021 ASEE Virtual Annual Conference Content Access, Virtual Conference. 10.18260/1-2--37630

TY  - CPAPER
AB  - Students&#39; academic success in science, technology, engineering, and mathematics (STEM) careers is one of the most popular subjects that has gained attention among educational researchers for decades. Many studies have shown students&#39; educational outcomes can be affected by academic factors including high school GPA, SAT score test, student admission type (transfer or first-time-in-college), as well as demographic features such as gender, ethnicity, and family income. Additional studies have investigated the relationship between students&#39; course load and their academic outcomes. In this paper, we define students&#39; course load based on the number of courses they take each semester, which is assumed to have a discrete probability distribution. To assess if students&#39; course load impacts their academic performance, we apply Hidden Markov models, an unsupervised learning method, to classify students into three categories: high-level enrollment, medium-level enrollment, and low-level enrollment. The sequence of the number of courses students enroll each semester during their academic career is fed into our proposed model as input. The output which is a qualitative measure and is not directly observable is the estimated enrollment level for the students. After students&#39; classification, we derive and compare their academic (e.g., cumulative GPA, graduation rate, and DFW rate) and non-academic (e.g., family income level) features for each enrollment level category. Findings show that students who have more engagement with the university have higher academic performance (higher cumulative GPA and graduation, lower DFW rate) than those with lower engagement. Our analysis also demonstrates that students from families with low-income levels are more likely to have lower enrollment levels. Such results indicate that university managers can improve students&#39; educational performance and, subsequently, the university graduation rate by encouraging students to engage more with the university by providing academic and financial support. These results are based on the data collected from the University of XXX from 2008 to 2016 and contain approximately 170,000 students.
AU  - Shahab Boumi
AU  - Adan Ernesto Vela
CY  - Virtual Conference
DA  - 2021/07/26
PB  - ASEE Conferences
TI  - Quantifying the Impact of Students' Semester Course Load on Their Academic Performance
UR  - https://peer.asee.org/37630
DO  - 10.18260/1-2--37630
ER  -