Story Name: Police Manpower and Crime

Datafile Name: New York City Crime

Abstract: The datafile contains percent changes in manpower and seasonally adjusted changes in weekly auto thefts and larcenies for the 25 precincts in New York City from a base period of 27 weeks in 1966 to an experimental period of 58 weeks in late 1966 and 1967. During the experimental period police manpower assigned to Precint #20 was increased by about 40 percent. Precinct #20 lies south and west of Central Park, extending to the Hudson River. If a simple regression of change in crime rate on change in police manpower assigned were conducted, Precinct 20 would exert undue influence on the regression. On the other hand, indexing Precinct #20 with a dummy variable and running the a two-predictor regression of crime rate on manpower change and the Precinct #20 dummy variable provides a way of estimating the effects of the manpower experiment under certain important assumptions.

Image: Scatterplot of change in thefts by change in manpower