In an era when it has become fashionable to advocate that social policy be guided by evidence, social scientists should be riding high. But there is a problem: The available evidence on what works in the social-policy arena is typically something less than definitive. Experts disagree. Sometimes the only consensus that can be mustered among researchers is that" more research is required," often a dubious assertion when there has been voluminous research already. Meanwhile, policy-makers are left free to either ignore the research evidence, or to search out an expert who supports their position.

A case in point is the recent report on gun violence of an expert panel of the National Research Council. In one topic after another, the NRC's blue-ribbon panel concludes that the existing evidence is inconclusive (Wellford, Pepper, & Petrie, 2005). 1 It calls for an investment in better data, the invention of better