Applied Multilevel Models for Longitudinal and Clustered Data


  • Ryan Walters, Creighton University

This workshop will be offered in an online video format.

Multilevel models are known by many synonyms (e.g., hierarchical linear models, generalized linear mixed models). The defining feature of these models is their capacity to provide quantification and prediction of random variance due to multiple sampling dimensions (e.g., across occasions, persons, or groups). Multilevel models offer many advantages for analyzing longitudinal data, such as flexible strategies for modeling change and individual differences in change, the examination of time-invariant or time-varying predictor effects, and the use of all available complete observations. Multilevel models are also useful in analyzing clustered data (e.g., persons nested in groups), in which one wishes to examine predictors pertaining to individuals or to groups. This workshop will serve as an applied introduction to multilevel models, beginning with longitudinal data and continuing to clustered data. Multilevel models for (conditionally) normal and non-normal outcomes will be presented. This workshop will focus heavily on appropriate interpretation of all fixed and random effects to ensure attendees appropriately estimate these models and report results. Numerous software examples in Stata, SPSS, and SAS will be provided to allow attendees to begin using multilevel models in their own research. This workshop will include daily opportunities for hands-on practice using instructor-provided in-class activities. Attendees are strongly encouraged to bring their own data as well.

In terms of statistical prerequisites, attendees should be familiar with the general linear model (e.g., ANOVA and linear regression), but no prior experience with multilevel models or knowledge of advanced mathematics (e.g., matrix algebra) will be assumed or required. Knowledge of logistic regression would be beneficial, but not required. The first day will be spent introducing the multilevel model. The second day will be spent fitting random effects in unconditional (aka, no predictor) longitudinal models and discussing the rules of model comparisons. The third day will be spent fitting conditional multilevel models for longitudinal data that include time-invariant predictors (i.e., predictors that are constant within a person) and time-varying predictors (i.e., predictors that are not constant within a person). Heavy focus will be placed on the nuisance of including and interpreting time-varying predictors. The fourth day will be spent fitting conditional models for clustered data. The fifth day will be spent fitting multilevel models for non-normal outcomes.

Registration Fee: Members = $1,500; Non-members = $3,000

Tags: multilevel models, mixed models, hierarchical linear models

Course Sections

Section 1

Location: Online -- Video format,

Date(s): July 13 - July 17

Time: 9:00 AM - 5:00 PM


  • Ryan Walters, Creighton University