Comparing Regression Discontinuity and Multivariate Analyses of Variance

Abstract

The purpose of this study was to compare the results of a regression discontinuity design (RDD) with those of an experimental design of a tiered vocabulary intervention for children at risk for reading disability to determine RDD’s feasibility as a research methodology for this type of study. Researchers reanalyzed an archival dataset of a previous study that determined the effectiveness of the intervention through a randomized experiment. Results were comparable for the receptive outcome scores in direction, magnitude, and significance, although the magnitude of the treatment effect for the RDD was slightly higher. Results were also comparable for contextual outcome scores with respect to direction and magnitude. We present possible reasons for the different results between studies, limitations, and future research.