The importance of using natural experiments and experimental data in economic research has long been recognized. Yet, it is only in recent years that these approaches have become an integral part of the economist's analytical toolbox, thanks to the efforts of Meyer, Card, Peters, Krueger, Gruber, and others. This use has shed new light on a variety of public policy issues and has already caused a major challenge to some tightly held beliefs in economics, most vividly illustrated by the finding of a positive effect of a minimum wage increase on the employment of low-wage workers. Although currently in vogue in economic research, the analysis of experimental data and natural experiments could be substantially strengthened. This paper discusses how analysts could increase the precision with which they measure treatment effects. An underlying theme is how best to measure the effect of a treatment on a variable, as opposed to explaining a level or change in a variable.