Commonly used single-level analyses may not be appropriate for studying change in schools since curriculum evaluation data are often nested within a hierarchical or multilevel structure (e.g., students are nested in classes and/or schools). Fortunately, the hierarchical linear model (HLM) provides a better alternative for analyzing this type of nested data. The purpose of this presentation is to discuss the process of using hierarchical linear modeling to evaluate curricula and children’s performances through the illustration of a study that evaluated a statewide curriculum and student learning in physical education in the areas of motor skills, personal-social development and fitness performance. Nineteen teachers using either the statewide curriculum at high implementation levels or not using the curriculum participated in this study along with their 1,209 Kindergarten through second-grade students. Student data were collected on four outcome variables (i.e., 600-yard run/walk, hop, overhand throw, and self-reported personal/social behavior). Data were analyzed using the HLM by level and role including student outcomes, student predictors and school predictors. Two key factors contributed to the observed student differences, physical education (PE) specialists and use of the statewide curriculum lessons. Students taught with the statewide curriculum had significantly faster 600-yard run/walk scores and higher self-reported personal/social behavior scores. This study illustrates how the HLM can be used to study curricular differences in PE programs, therefore, eliminating problems associated with single-level analyses with multilevel data.