ABSTRACT
Many examples of educational policy research involve the analysis of hierarchical, or nested, data structures. These data structures are prevalent in research on educational organizations (i.e., where students are nested in classrooms, schools, districts, and states) and policy systems (i.e., where individuals are nested in various policymaking groups). Because of the limits of statistical model techniques, until recently, researchers had to make choices about whether to ignore the groupings and analyze the data on individuals, or whether to ignore the individuals and analyze the data for groups only. Neither solution took full advantage of the complexity of the data structure. Concerns in various fields with how to analyze hierarchical data properly led to the development of multilevel models as extensions of linear models over the past few decades.
