Difference-in-Differences Regression Modeling

• In Event: Linear, Nonlinear, Latent, and Machine Learning Modeling for Educational Research

Abstract

An interesting multiple regression application is the difference-in-difference regression model (DID). The DID model estimates the differential impact of a treatment on subjects in an experimental group versus a control group over time. The DiD regression model includes a key variable, time, in the analysis. The passage of time will change the value of the dependent variable from pre-treatment to post-treatment in a research study. The DiD regression model accounts for the effect of time, thus allowing for interpretation of the treatment effect on the dependent variable over and above the passage of a time effect in the study. A DiD regression model will be demonstrated to illustrate how the equation is written, data analyzed, and interpretations presented.

Authors

• Randall E. Schumacker, University of Alabama

• Todd Sherron, Texas State University