In this article, ten years of police, crime and other data from London was used to investigate the potential effect of stop and search on crime. Using lagged regression models and a natural experiment, the authors show that the effect of stop and search on crime is likely to be marginal, at best. The claims that stop and search is an effective way to control and deter offending are concluded to be misplaced. The article suggests that, first, in a legal sense each and every stop should be justified in itself, not in that it has some putative wider effect on crime, and, second, in a sociological sense, that stop and search is a tool of social control widely defined, not crime-fighting, narrowly defined.