Managers of government technology programs are under increasing pressure to demonstrate the effectiveness of their programs. In this paper we examine the issues involved in credibly evaluating such programs in the context of recent efforts to evaluate manufacturing extension programs in the U.S. We provide a stylized model of the dynamic competitive environment in which the plants and firms targeted by these programs operate and discuss its implications for evaluation. We compare and contrast the various methodologies and data sets used to evaluate manufacturing extension. We conclude that the best currently available method for measuring the overall effectiveness of programs such as manufacturing extension is to combine program administrative data with existing panel data sets.