ABSTRACT
This chapter presents situations where the nesting of subjects within groups is more complicated than in typical hierarchical data sets. In cross-classified data structures, lower-level units (e.g., individuals) are simultaneously nested within separate higher-level units (e.g., student teacher training institutions and initial districts who hire them). Researchers utilizing such data may wish to examine how these different combinations of teacher preparation institutions contribute to student teachers’ transition to initial school districts and their likelihood of obtaining tenure. Fortunately, examining cross-classified data structures does not require that factors associated with random effects be hierarchical factors in the design. The chapter also introduces multiple membership models using IBM SPSS Mixed, extending the cross-classified model to situations where individuals might be members of more than one unit at the same classification level. For example, suppose the outcome for each individual is obtaining an undergraduate degree. In that case, some students in the data set might attend more than one postsecondary institution, depending on whether they entered a two-year institution first and then transferred to a four-year institution to finish their undergraduate degree. Additionally, some students may have entered one four-year institution and then transferred to one or more other institutions.
