Date of Award

Spring 5-2016

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Educational Studies and Research

Committee Chair

Richard Mohn

Committee Chair Department

Educational Studies and Research

Committee Member 2

Forrest Lane

Committee Member 3

Thomas Lipscomb

Committee Member 3 Department

Educational Studies and Research

Committee Member 4

Kyna Shelley

Committee Member 4 Department

Educational Studies and Research

Abstract

College student retention and graduation are important to students, institutions, and the community. Institutions must commit to understanding why students persist and depart in order to address student success. As a result, institutions and governmental entities have increased the emphasis they place on using data to improve student success and degree completion. An abundance of research suggests that background factors (such as high school GPA and ACT score) combined with environmental factors (such as one’s major and first semester GPA) are predictive of student success. However, the literature has yet to explore the value of ROC curve analysis as a statistical technique to improve decision making. The purpose of this study was to identify the variables that best predicted satisfactory academic progress and degree completion, and model the use of ROC curve analysis.

This study utilized a quantitative approach and secondary data from the institutional research office of a State University System institution in Florida. The findings of the study include the following: (1) logistic regression was successful in identifying factors that were predictive of each outcome, which were further refined using ROC curve analysis; (2) ROC curve analysis successfully discriminated the success/non-success groups based on predicted probability scores and a cumulative risk index; and (3) if a student has more than two risk factors in their profile, they are likely to not achieve any of the outcomes.

ORCID ID

0000-0002-1363-0736

Download

Share

COinS