Author ORCID Identifier
https://orcid.org/0009-0006-5790-0343
Date Available
12-13-2024
Year of Publication
2024
Document Type
Doctoral Dissertation
Degree Name
Doctor of Philosophy (PhD)
College
Education
Department/School/Program
Educational Policy Studies and Eval
Advisor
Dr. Kelly Bradley
Abstract
Appropriate methodology choices are of the utmost importance in research, particularly in measurement and education research. When higher education administrators and researchers study topics related to student outcomes, choosing a statistical approach that produces unbiased estimates is of primary concern. Additionally, studying student outcomes in the context of nationally representative samples provides valuable insights into the state of higher education in the U.S. When examining these data, it is critical that the nested structure be taken into consideration prior to making methodological decisions.
The purpose of the following manuscripts is to demonstrate how education researchers can develop, implement and evaluate a framework that accounts for the structure of observations nested within clusters – mirroring common scenarios found in higher education. Through a lens of student persistence within higher education, a framework for student outcomes is presented based on existing research and literature. Methodologies that account for observations nested within clusters are tested across simulated scenarios to demonstrate how bias can be introduced to results. Research questions regarding the effect of online distance learning on student outcomes are presented and tested under this framework and through these analytic approaches.
The specific methodologies of interest for this research include Generalized Linear Mixed-effects Models (GLMM) and Generalized Estimating Equations (GEE). GLMM is a common approach for handling clustered data in education research. GEE is utilized less frequently in education research and is more prevalent in biostatistics. Both approaches can account for clustered data, but yield different, nuanced interpretations. By demonstrating these methods through a commonly researched topic in higher education (student persistence), this research exhibits the value of exploring cross-disciplinary approaches for studies in higher education.
Finally, recommendations for education researchers are provided including details on research question formation, how to select appropriate methodologies that match the research questions, and the importance of multidisciplinary collaboration when investigating complex data scenarios. A review of this research will act as a guide for handling data with observations nested in clusters, as well as how to interpret results from these methods.
Digital Object Identifier (DOI)
https://doi.org/10.13023/etd.2024.452
Recommended Citation
Wilson, Nathaniel P., "AN EVALUATION OF STATISTICAL APPROACHES WITHIN A HIGHER EDUCATION RESEARCH FRAMEWORK: HOW DOES ONLINE DISTANCE LEARNING IMPACT STUDENT OUTCOMES?" (2024). Theses and Dissertations--Educational Policy Studies and Evaluation. 104.
https://uknowledge.uky.edu/epe_etds/104
Supplemental File - Simulation Matrix.csv
R Script Simulation Run 1.txt (7 kB)
R Script Simulation
Included in
Educational Assessment, Evaluation, and Research Commons, Educational Methods Commons, Educational Psychology Commons, Higher Education Commons, Online and Distance Education Commons, Science and Mathematics Education Commons
