Analysis of Intra-classroom Correlation: Implications for Random Assignment

This project is intended to provide empirical information needed to design adequately powered studies.

The project will estimate school-level, classroom-level and student-level variance components for measures of student achievement—with and without covariates (especially pretests)—from data for a set of school districts in North Carolina and Florida. The basic research questions to be addressed are: (1) How do values of the parameters vary by grade level and subject? (2) How do the values of these parameters vary by type of school district and type of school? (3) How does the predictive power of different types of student-level, classroom-level and student-level covariates vary?