Date of Award

5-2018

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Educational Studies

Committee Chair

Yukiko Maeda

Committee Member 1

Anne Traynor

Committee Member 2

Donald Mitchell, Jr.

Committee Member 3

Frank Dooley

Abstract

Despite the overwhelming evidence that higher education data are nested at various levels, single-level techniques such as regression and analysis of variance are commonly used to investigate student outcomes. This is problematic as a mismatch in methodology and research questions can lead to biased parameter estimates. The purpose of this study was to predict cumulative grade point average (GPA) and the likelihood of four-year and six-year graduation while simultaneously accounting for select pre-college characteristics, during-college experiences, and the interrelationship between student-level and major-level predictors. To achieve the desired outcomes, the study applied multilevel modeling techniques to secondary data for new undergraduate students first enrolling at one research institution in the Midwestern United States during Fall 2010 and Fall 2011. Results suggest that approximately 30% of the variation in cumulative GPA, 32% of the variation in four-year graduation, and 48% of the variation in six-year graduation can be attributed to differences in academic majors. Results also indicate that the strength of the student-level predictors of high school GPA, changing one’s major, first-year GPA, and student organization involvement vary across academic majors. Collectively, the study contributes to the application of quantitative research methodology in higher education by demonstrating a more accurate predictive model of academic success for undergraduate students.

Download

Share

COinS