The evidence that changes in policing policy made "all the difference" is largely circumstantial. Homicide rates were at an all-time high in 1990-91 and had begun to decline before any radical changes in policing policy were instituted. The 1985-91 "murder spike" had been attributed largely to the simultaneous expanding crack cocaine "epidemic" so the subsequent reduction in murder is related logically to the contraction of crack cocaine markets in the 1990's. There is some tentative support for the impact of policing on an already falling crime rate, but the changes in policing between 1991 and 1997 cannot adequately be described as "zero tolerance." The changing social context (decline in the use and dealing of crack cocaine), together with the combined effect of preventive work among local communities and a rejection of crack cocaine and guns by a new generation of youth provides a credible alternative account to that generated by the media-smart New York City police. The author advises that when violent crime declines, there must be rigorously collected evidence to explain what is happening. Without such evidence, explanations will be offered by powerful people with a political interest in claiming that they are responsible for improving the "quality" of city life. 42 footnotes and 72 references