A Few Things I’ve Learned About Multilevel Modeling (MLM) Over the Past 20 Years

USC Quantitative Speaker Series (Spring 2025)

Date: April 18, 2025

Speaker: Oi-man Kwok, Ph.D.

Professor
Department of Education and Human Development
Texas A&M University

Video Recording (requires sign in using your USC NetID)

Abstract

Multilevel modeling (MLM)—also known as hierarchical linear modeling (HLM) or mixed-effects modeling—is widely used for analyzing nested data in the social sciences. In this presentation, I will highlight several key concepts in MLM, beginning with variance decomposition and intraclass correlation. I will then focus on predictor centering, emphasizing the importance of group-mean centering for lower-level predictors and its implications for testing contextual effects. The presentation will conclude with practical recommendations for analyzing multilevel data.