Abstract

Recent research in academic analytics has focused on predicting student performance within, and sometimes across courses for the purpose of informing early interventions. While such an endeavor has obvious merit, modern contructivist learning theory expresses an importance on more individualized support for students. In keeping with this theory, this research describes the development of a model that predicts student performance within a course, relative to their past academic performance. This study is done using the minimum sources of data possible while still developing an accurate model. Useful logistic models using data from the institution’s student information system, learning management system, and grade books some useful findings are developed. While each source of data was able to predict student success independently, the most accurate model contained data from both the grade book and student information system. These models were able to accurately identify students on track to underperform relative to their own cumulative grade point averages within the first seven weeks of a course, aligning with the studied institution’s existing requirements for a manual early intervention system.

Publication Date

8-9-2017

Document Type

Thesis

Student Type

Graduate

Degree Name

Industrial and Systems Engineering (MS)

Department, Program, or Center

Industrial and Systems Engineering (KGCOE)

Advisor

Rachel Silvestrini

Advisor/Committee Member

Scott Grasman

Campus

RIT – Main Campus

Share

COinS