This lecture is a low-level introduction to multilevel models. We begin by looking at examples of hierarchical data structures and these create dependencies in model errors. We look at the distinction between fixed and random model parameters. We then look at an example based an RCT style design of an intervention implimented across 10 clinics and how we would apply a multilevel model with random effects for the intercept and treatment effect. We look at how to specify various models using the nlme package in R, and how to interpret the output.
Learn R alongside these lectures with the discovr package (www.discovr.rocks/discovr/)
Suggested soundtrack:
Wolves in the throne room: Prayer of transformation ( • Prayer of Transformation )