STAT 7450:Multilevel Statistical Models

3 Class Hours 0 Laboratory Hours 3 Credit Hours
Prerequisite: STAT 7100 and STAT 7210
Data often have a structure that is nested or hierarchical. For example, when investigating student outcomes, we need to consider that students are nested inside classrooms that are in turn nested inside schools. Clustered data violate the assumption of independence of error terms expected of the general linear model family, which includes ANOVA and regression. Multilevel models allow us to analyze data where observations are not independent, correctly modeling correlated errors and avoiding biased standard errors. This course will focus on how to use multilevel (hierarchical) models for dealing with nested data. We will discuss common topics in multilevel modeling, including sample size requirements, centering decisions, assumptions, and slopes and intercepts as fixed and random effects. Our focus will be on two and three level models, as well as growth modeling. The software we will use is HLM and R.