In this chapter we introduce developing and interpreting multilevel models. We first define multilevel models and explore how this approach is an improvement on disaggregation and aggregation of data across multiple levels. We then work through four different multilevel models. We provide examples of what kinds of questions can be answered by each model and how to interpret the statistical output. We then explore some additional issues in fitting multilevel models in Stata and consider additional applications of multilevel models.

Nested data arise frequently in clinical research. The nesting might be hierarchical, such as patients nested within clinicians, or it might be longitudinal, such as repeated assessments over time nested within individuals. As articulated in this chapter, whenever and however nesting occurs, it is necessary to account for the statistical dependence of observations within units when analyzing the data. Further, it is important to determine the level(s) of the data at which predictors exert their effects. Multilevel models are a particularly popular and useful approach for addressing these issues. We thus describe these models in detail, illustrating the application of multilevel models in clinical research via two examples. The first example considers nesting of siblings within families and demonstrates the importance of separating within- versus between-family effects. The second example focuses on the application of multilevel models with repeated measures to evaluate within-person change over time. Additionally, we provide a brief survey of other approaches to the analysis of nested data (e.g., cluster-robust standard errors, generalized estimating equations, fixed-effects models).