ABSTRACT

The technique of multilevel modeling is now well established and has been incorporated

into most standard statistical packages. It is convenient for present purposes to consider

educational data that exhibit a hierarchical structure of students nested within classrooms

nested within schools-a three-level structure. Other examples are repeated-measures

data with occasions nested within subjects, and surveys with people nested within

households, both of these being two-level data structures. More-complex data structures

such as cross-classifications and multiple-membership structures are extensions for which

models have been developed. Goldstein (2003) provided a detailed exposition with

references to further application areas.