Difference-in-Differences Regression Model and Comparable Methods

Abstract

Multiple regression model applications have expanded since the 1960s, especially with the development of computers and computer software. The Difference-in-Differences regression model has received much attention in determining treatment effects between experimental and control groups. There are two other regression applications that are comparable models in analyzing treatment effects. The other two similar applications are regression-discontinuity and test of slope differences; sometimes referred to as a test of parallelism. This article presents the similarity between these regression applications, which is a test of interaction between the group variable and a continuous variable (time; scores) that usually involves a cut-off point before and after a treatment condition. The treatment effect is shown to be the same in all three methods.