Search
Program Calendar
Browse By Day
Browse By Time
Browse By Person
Browse By Room
Browse By Unit
Browse By Session Type
Search Tips
Annual Meeting Registraion, Housing and Travel
Personal Schedule
Sign In
When data are nested, traditional multilevel mediation models can produce biased estimates of indirect effects when the within- and between-level effects are conflated. They are also limited by their inability to accommodate upper-level dependent variables. We describe a multilevel structural equation model that can be used to decompose observed L1 variables into within- and between-level latent components that control for mesurement and sampling error, and remove bias in estimates of indirect effects. We also illustrate this approach through investigation of the mediating role of student aggressive attitudes in associations between teacher reports of school climate and school suspension rates. Results are based on a statewide sample of 59,483 students and 11,328 teachers from 295 schools.
Yuane Jia, Rutgers University
Timothy R. Konold, University of Virginia
Kelly D Edwards, University of Virginia
Dewey G. Cornell, University of Virginia