Do visits or time spent in hot spots matter most? A randomised control trial in the West Midlands Police

Randomised Control Trial Status


Trial Hypothesis

​The hypothesis for this small scale randomised control trial is that shorter more frequent patrols of hot spots will result in greater deterrence of street crime and ASB calls for service.

​Shorter more frequent patrols being 5 minutes in length conducted 9 times during a late shift compared with longer less frequent patrols of 15 minutes conducted 3 times during a late shift. Will creating more frequent initial deterrence and leaving less time for deterrence decay to kick in cause there to be less crime? Research to date suggests that the threshold of patrol time required to create a lasting deterrent effect is around 10 minutes (Koper, 1995). In contrast, however, there is now emerging evidence that suggests the frequency of visits may have more of an influential (Ariel, 2015) role.

Geographical area

​Birmingham West and Central Local Policing Unit, Birmingham, West Midlands.

Research Institution / Organisation

University of Cambridge

In Collaboration With

West Midlands Police

Contact Name

Simon Williams

Project Start Date

June 2015

Participants - inclusion criteria

  • Random assignment of days to 3 x 15 minute patrols or 9 x 5 minute patrols
  • Geo-fenced patrols
  • Patrol Data fed back to patrol officers
  • Street Crime
  • ASB tracked 

Target Sample Size

  • ​7 Hot spots
  • 100 days
  • Activity Analysis of over 2000 individual patrols
  • Sample Treated as population data, effect size measured using Cohen's d and descriptive statistics

Study Design

An experiment was designed in which 7 hot spots were randomly allocated to one or other patrol mode for a period of 150 days between June and November 2015. Patrol visits were tracked using patrolling officers’ personal issue G.P.S (global positioning system) ‘Airwave’ radios where patrol information was fed back and officers help to account for the number of patrols conducted. Although this research took place over 150 days the results presented are based on 100 days of patrol as a result of a breakdown in ‘geo-fencing’ software during the last 50 days.

Summary of Findings

​Fewer units of longer duration are associated with greater crime falls, indicating that they are more effective than more frequent shorter patrols. The findings from this experiment confirm Koper’s (1995) finding that longer units of 10-15 minutes duration are more effective.
Additionally activity analysis of police constable and police community support officers overwhelmingly indicates that the highest frequency outputs, accounting for nearly 90% of all activity during 15 minute patrol days, do not require police powers (i.e. Community engagement and visits to high demand crime and ASB micro-locations within hot spots).

Do visits or time spent in hot spots matter most? A randomised control trial in the West Midlands Police - Final report

Date last updated

Monday 16 July 2018
Return to Research Map