Experimental Impact Under Sufficient Implementation: A Moderator Analysis Approach

• In Event: Roundtable Session 97
  In Roundtable Session: Accommodating Data From Quasi-Experimental Designs

Abstract

This paper develops a new approach to estimating the impact of a program under the condition where the subjects receive adequate exposure to the program. The method uses pre-randomization characteristics of subjects to predict the probability of receiving the program with adequate exposure in both conditions. The moderating effect of the predicted probability on the impact of the program provides the estimate of impact under adequate exposure. The method build on prior works but is novel in that it uses a dichotomous measure of exposure status. Three variants of the method are applied to results from a randomized trial. Results are discussed and limitations are enumerated.

Authors

• Boya Ma, Empirical Education Inc.

• Ping Yi, Empirical Education Inc.

• Andrew P. Jaciw, Empirical Education Inc.