Design Parameters for Planning Longitudinal Experiments in Education

• In Event: AERA Poster Session 4
  In Poster Session: Poster Session (San Diego)

Abstract

The design of longitudinal experiments in education requires information about reliability coefficients and intraclass correlations to conduct statistical power analysis and determine the sample sizes necessary for adequately powered studies. In addition, covariate effects (i.e., the proportion of variance explained by the covariates) is also an important element of conducting power analysis in longitudinal experiments. This study presents a compendium of reliability coefficients, intraclass correlations and covariate effects for planning two- and three-level longitudinal experiments in education. Our models involved reading and mathematics scores as outcomes, and student demographics (e.g., gender, race, SES) and school characteristics (e.g., school composition, size, location) as covariates.

Authors

• Wei Li, University of Florida

• Spyros Konstantopoulos, Michigan State University