Mixed Models Part 2
In a previous post , we looked at four different linear models for a collection of data where besides x and y values, the data was qualifies by an additional factor called a Group . In those models, we either assumed that the difference in groups had no effect on the model, models 1 and 2, or that the groups were independent, and we could estimate their effects on slope and intercept individually. A compromise between these two approaches is the mixed-effect model , sometimes called a multilevel linear model or a random effects model. The difference is that the groups are assumed to be a random sample from a larger population of groups measuring the same kinds of data. In other words, are the groups we see all the groups we will see or is the data a sample from the world of possible groups? A Linear Model The simplest mixed model is similar to model 2 from the previous post. In this model we expect the Group value to affect the intercept of the linear model. \[\begin{a...