Another variation on the basic three-level modeling framework is a multivariate multilevel model—a model with more than one dependent variable (defined at level 1), with individuals at level 2 and groups at level 3. The multivariate multilevel model follows directly from the traditional single-level multivariate analysis of variance (MANOVA) model. This chapter provides three extended examples of two- and three-level models with multivariate outcomes using IBM SPSS Mixed. The first example is one where the goal is to incorporate several survey items to define one or more latent constructs. This type of latent variable formulation allows the incorporation of measurement error and possible missing data on items in the analysis. The second example occurs when there are two or more observed outcomes, for example, where the researcher might examine students’ reading, math, and language test scores simultaneously, rather than modeling each outcome separately. This type of multivariate formulation is helpful in adjusting parameters for the expected correlation between students’ performances on each test. The third example is one where researchers are interested in examining individual development in two or more domains simultaneously by defining a parallel growth model rather than examining each developmental trajectory separately.