Higher Education Ph.D. Dissertations

Predicting Student Veteran Persistence

Date of Award

2020

Document Type

Dissertation

Degree Name

Doctor of Philosophy (Ph.D.)

Department

Higher Education Administration

First Advisor

Maureen Wilson (Advisor)

Second Advisor

Andrew Pelletier (Other)

Third Advisor

Kenneth Borland (Committee Member)

Fourth Advisor

Christina Lunceford (Committee Member)

Fifth Advisor

Jessica Turos (Committee Member)

Abstract

The three-fold primary purpose of this study was to: (a) describe student veterans at Bowling Green State University (BGSU) in terms of independent variables, representing students’ input characteristics, environmental factors, and BGSU experiences; (b) identify differences between student veteran persisters and nonpersisters in terms of these variables, and (c) determine how well these variables predicted persistence outcomes.

Astin’s Input-Environment-Outcome (I-E-O) framework (1993) and the work of Bean and Metzner (1985) on nontraditional student attrition were adapted to serve as the organizing framework for this study. The study sample (N = 537) comprised BGSU degree-seeking undergraduates who, based on their military service, received assistance from the U.S. Department of Veterans Affairs (VA), during their first BGSU term, Fall 2009 – Fall 2015.

Descriptive statistical analysis resulted in a detailed picture of the study sample, comparisons of persisters and nonpersisters, and profiles of associate and bachelor’s degree completers. By the end of the study period (August 2017), 174 students (32.4%) had completed a BGSU degree and another 86 had reenrolled for at least one term, Spring 2017 or Summer 2017, constituting 260 persisters, 48.4% of the study sample.

Chi-square tests of independence and independent samples t-tests were used to analyze differences between persisters and nonpersisters. Limited chi-square analyses of a small subset (n = 109) of the study sample failed to find statistically significant differences between persisters and nonpersisters on military experience variables (combat exposure, military rank, and reserve status). Binary logistic regression analysis was conducted to determine which set of variables best predicted persistence status. Significant variables in the best-performing model (overall correct classification, 83.1%; -2LL = 416.633; Nagelkerke R2 = .609) were total transfer credits, VA benefit program, start term, residence hall status, first-term credit hours, first-term GPA, major change, and summer enrollment.

This study, one of the few to employ student-level institutional data to examine persistence of contemporary student veterans, revealed the importance—and the challenges—of using this type of data to gain a more nuanced understanding of this diverse yet somewhat invisible population. Implications for student affairs policy and practice were considered and some future research directions suggested.

Link to Full Text

Share

COinS