Chapter 28: Teaching Multilevel Modeling

Multilevel models (MM) constitute a class of statistical models, which are the preferred analytic tool for analyzing hierarchically nested or clustered data (e.g., when a set of students are randomly selected from randomly selected classrooms then the students are “nested” within classrooms).

Multilevel models are known by several names, including hierarchical linear models (HLM), linear mixed-effects, random coefficients regression models, and covariance components models (Scientific Software International, 2014). When analyzing longitudinal data, multilevel models are often referred to as multilevel growth models. MMs are also known as multilevel regression analyses because the method builds on the regression framework and, conceptually, can be thought of developing regression lines for each higher-level cluster (e.g., schools, neighborhoods, employee teams). MM allows for the study of such hierarchical data in a single analysis while simultaneously accounting for the variability associated with each level of the hierarchy (Raudenbush, & Bryk, 2002).