Kyle Cox

Kyle Cox

Assistant Professor
Educational Leadership
Cato College of Education, Mebane Hall 266

Kyle Cox is an assistant professor of educational research, measurement, and evaluation at University of North Carolina at Charlotte where he teaches graduate level statistics and research methods courses. His research focuses on improving the feasibility of multilevel studies through design improvements and analytic advancements. This work is applicable across the social sciences as the methods accommodate natural hierarchical structures and complex theories but Kyle is most interested in their application in educational settings. Specifically, Kyle has investigated statistical power in experimental multilevel mediation and moderation studies and is interested in improving the estimation of structural equation models when sample sizes are limited. Prior to joining UNC Charlotte in 2019, Kyle earned his doctorate in quantitative research methodology at the University of Cincinnati after nearly a decade of teaching 6th grade math.


Ph.D.- University of Cincinnati, 2019, Educational Studies: Quantitative Research Methods
M.A.- University of Cincinnati, 2012, Educational Studies: Quantitative Research Methods
B.S.- Miami University, 2005, Education


Structural Equation Modeling
Teaching Advanced Statistics
Research Methods


Research Interests/Areas of Expertise Structural Equation Modeling
Multilevel Experimental Designs
Mediation and Moderation Effects
Mathematics Education


University of Cincinnati Research Council Graduate Student Stipend and Research Cost Award for Faculty-Student Collaboration
2016 CADRE STEM Fellowship
Finalist for Top Proposal to the 2016 American Educational Research Association Annual Meeting: Division D In-Progress Research Gala
2015 Project Excellence Award for Teaching

Selected Publications

Cox, K., & Kelcey, B. (2021). Croon’s bias corrected estimation of latent interactions. Structural Equation Modeling: A Multidisciplinary Journal, 28, 863-874.
Cox, K., & Kelcey, B. (in press). Statistical power for detecting moderation in partially nested designs. American Journal of Evaluation.

Curriculum Vita