Application of Difference-in-Difference Techniques To the Evaluation of Drought-Tainted Water Conservation Programs

A nonexperimental evaluation's validity can be threatened by a number of factors that are well known in the program evaluation literature. This article illustrates some of the internal validity threats that arise when evaluating the impact of water conservation programs implemented during a drought. These include differential response to the drought because of imbalance in the control and participantgroup characteristics, self-selection bias, and the impact of measurement error. This article shows how to deal with these problems in the context of water conservation programs when high-quality disaggregate data are available.