Nonexperimental program evaluations are performed under numerous constraints that require a judicious blend of theory, attention to data quality, and judgment. This paper describes the conceptual and measurement issues that arise while evaluating nonresidential water conservation programs implemented during a drought in the absence of a true control group. Sensitivity analysis techniques are presented to assess the impact of the drought on comparisons of preprogram and postprogram water consumption. ¿ American Geophysical Union 1996