One can readily extend the basic two-level univariate regression model to cross-sectional models involving several levels in a data hierarchy, cross-classified data structures (where individuals belong to multiple groups), and multilevel models with categorical outcomes. This chapter provides an overview of three-level, cross-sectional modeling using IBM SPSS Mixed. After building a series of three-level models investigating a random intercept and level-2 slope, it explores the possible presence of cross-level interactions (a macro-level variable explaining variability in a lower-level regression slope) and within-level interactions (i.e., where a third variable may moderate a regression slope at the same level of the data hierarchy). The chapter also introduces strategies for centering predictors in multilevel models to facilitate the interpretation of effects and comparing the fit of successive proposed models to the data. Finally, it also introduces multilevel univariate models with categorical outcomes (i.e., using an extended example of a binary outcome) using IBM SPSS Genlinmixed. This last example serves as an introduction to developing further multilevel models with categorical (e.g., ordinal, count) outcomes.