Share
This summary is provided by Michelle Rigozzi.
The Rise in Evidence-Based Policing: Targeting, Testing, Tracking
Prof Lawrence Sherman, Wolfson Professor of Criminology, University of Cambridge. 31 October 2013.
The ‘Triple-T’ strategy is a new paradigm showing great promise within police force decision making. Improvements in data collection technology and analysis have assisted the development of this strategy. Its three principles are:
1) Targeting scarce resources by focussing on predictable concentrations of crime and disorder;
2) Testing police practices to help choose those that work best to reduce harm, and;
3) Tracking the delivery and effects of police practices.
These tasks go beyond randomised controlled trials. They require the police to continuously focus resources, test hypotheses and track what they are accomplishing by means of generating and using empirical evidence.
Use of evidence requires thinking in a certain way. In the language of Kahneman and Tversky, thinking can be separated into fast, unconscious thinking (what they call System I) and slow, conscious, reasoned thinking (System II). Evidence-based policing needs the latter: slow System II thinking, rather than fast System I thinking.
This ‘slow thinking’ can have dramatic effects on cost-effectiveness and the reduction of harm, as shown by the following controlled experiment:
Police in Trinidad and Tobago used a colonial policing pattern whereby police stayed in the police station without going on patrol. Crime there is almost 40 times higher than in England and Wales, with gun and gang crime outweighing other types of crime. The Cambridge Institute of Criminology conducted an experiment in Trinidad and Tobago in which the Triple-T strategy was implemented in half of the police districts. The other half (the control group) had matching levels of serious crime.
In September and October this year, six months into the experiment, the rate of homicide in the Triple-T districts was half of that in the control districts (17 compared to 34 instances). The rate of near fatal wounding and shooting was also almost halved (17 compared to 30 instances).
What was done differently in the Triple-T districts?
- Targeting: The focus was on homicide and gun violence in hot spot locations, where homicide happens repeatedly. This is hyper-local policing, which relies on police being visible in the hot spot location. Hot spots were also targeted at ‘hot times’, i.e. when homicides per hour were found to be highest (between 6pm and 2am in this case);
- Testing: Tested practices were used. These were (a) to stop and frisk people for handguns, (b) to greatly increase patrol time and visibility in hot spots, and (c) to convey a message of respect, even if gun carrying is suspected.
- Tracking: Both crime and policing were tracked. In particular, police patrols were monitored using GPS and fed back to police in a report every two weeks. Team leaders were held accountable for their teams and patrol results were compared between teams.
This use of Triple-T appears to have reduced serious crime in Trinidad and Tobago.
There is evidence that use of Triple-T is increasing in cities such as London and New York, where police have also been focusing on hot spots and hot times. Testing suggests that the optimal duration of a patrol in a hot spot is 15 minutes (the longer the police presence, up to 15 minutes, the longer the period of crime-free time in that spot after police have left); this practice has had uptake in a few cities in the UK and US, but not yet in most cities in any country. Opportunities for tracking and managing patrol time using GPS are also used in experiments, but have not yet been utilised in general police management.
To improve the quality of evidence-based practice, two suggestions were made. The first is that targeting requires systematic ranking and comparison of levels of harm. Professor Sherman proposes a Crime Harm Index that weights each crime counted by the number of days in prison. Secondly, testing can also be improved through making police aware of such features of the data as regression to the mean and how they cannot be used to infer the success of a policy without more information.
Finally, two challenges for the use of Triple-T were raised by Professor Sherman. Firstly, should police resources be distributed equally rather than targeted on concentrations of crime? Professor Sherman argues that this is like “giving aspirin to someone who isn’t sick - the aim should be for equality of results rather than equality of resources”.
The second challenge is policy decisions being “doomed to success”, as put by the London Metropolitan Police. There are pressures to make every policy decision sound like it was a success. This limits feedback on what works. The argument presented in this presentation is that if more research is carried out by the police and elsewhere, this will provide data to help decide what works best, in accordance with the Triple-T strategy.
Banner image from EUPOL Afghanistan via CC4.0