Chapter 3: The basic two-level models

This chapter is mainly intended for reading, with some review questions. We will proceed in three steps. First, we introduce two versions of variance component or random intercept models: the null model without explanatory variables and a model with explanatory variables. In these models, the regression constants (intercepts) can vary among the level 2 units (contexts). Secondly, we add further complexity in the random coefficient model, where the regression coefficients can show variation between contexts. In the final step, we will show models with explanatory variables also at the contextual level. We will conclude the theoretical part with an extended example, before looking at how the data file needs to be organized and estimating the models in SPSS and Stata.

Go to next page >>