Power Analysis for Interaction Effects in Longitudinal Multisite Randomized Designs

• In Event: Experiment Designs of Educational Research Studies

Abstract

Randomized control trials often incorporate a longitudinal component where for example students are followed over time, and student outcomes are measured repeatedly. Besides examining how the main effects of the intervention change over time, educational researchers are also interested in whether the effects of an intervention differ conditional on individual or cluster characteristics (i.e., moderators) such as gender, ethnicity, or school urbanity. This study provides methods of power analysis to detect interaction effects in longitudinal multisite randomized designs. Power computations take into account effect size, clustering effects, the number of measurement occasions, sample sizes at different levels, and the variance of the moderator. An illustrative example shows how power is computed.

Authors

• Wei Li, University of Florida

• Spyros Konstantopoulos, Michigan State University