<P> The design took three different police beats in Kansas City, and varied patrol routine in them . The first group received no routine patrols, instead the police responded only to calls from residents . The second group had the normal level of patrols, while the third had two to three times as many patrols . </P> <P> The experiment had to be stopped and restarted three times because some patrol officers believed the absence of patrols would endanger citizens . (Kelling et al. 1974). </P> <P> This was upheld for 12 months, from 1 October 1972 to 30 September 1973 . </P> <P> Victim surveys, reported crime rates, arrest data, a survey of local businesses, attitudinal surveys, and trained observers who monitored police - citizen interaction were used to gather data . These were taken before the start of the experiment (September 1972), and after (October 1973), giving' before' and' after' conditions for comparison . </P>

The kansas city preventive patrol experiment was conducted over a period of