Evidence-Based Policing

Evidence-Based Policing

View Segments Segment :

  • Citations
  • Add to My List
  • Embed
  • Help
  • Citations
  • Add to My List
  • Embed
  • Help
Successfully saved clip
Find all your clips in My Lists
Failed to save clip
  • Transcript
  • Transcript

    Auto-Scroll: ONOFF 
    • 00:01

      [MUSIC PLAYING][SAGE video tutorials][Evidence-Based Policing]

    • 00:11

      LAWRENCE SHERMAN: Hello, my name is Lawrence Sherman.[Professor Larry Sherman, Director of Criminology,University of Cambridge]I'm the director of the Institute of Criminologyat the University of Cambridge in England,and I'm a professor of criminologyat the University of Maryland in the United States.I run a program for police executiveswho come from all over the world to study at Cambridgeand then to take out to their own police agencies

    • 00:32

      LAWRENCE SHERMAN [continued]: what they've been studying, to apply it to police work.And it's based on the whole frameworkof evidence-based policing, whichhas been developed, really, over a 40-year periodin collaborations between universities and policeagencies, first in the United Statesand then in many other countries.

    • 00:54

      LAWRENCE SHERMAN [continued]: In this tutorial, we will provide an introductionto evidence-based policing, whichhas three key ideas-- targeting, testing, and tracking.Those three ideas come together in a Triple-T Against Crime,in which police are using big data in ways never usedbefore in policing to simultaneously pick

    • 01:17

      LAWRENCE SHERMAN [continued]: the most important parts for police efforts.They test the most effective methodsto deal with those targets.And they track whether those methodsare being used and delivered across big organizationscovering many square miles of space.And put it all together in what is called evidence-basedpolicing-- the evidence from the data about all

    • 01:40

      LAWRENCE SHERMAN [continued]: of these things, which is very different from kindof "good enough" or opinion-based policing,which doesn't have the same precision asevidence-based policing to create exactlyright decisions of the biggest impact for targets,deciding what works through testing,and making sure it happens through tracking.

    • 02:00

      LAWRENCE SHERMAN [continued]: [Targeting]Now, when we start with targeting,we are using what's called a powercurve, in which a small fraction of any populationis producing most of the harm or benefit that comes outof that population.17% of the offenders yield 50% of the arrests.

    • 02:21

      LAWRENCE SHERMAN [continued]: 3% of the addresses in the city may produce 50% of the callsto the police.1% of the officers yield the most complaints.And if crime's going to go down, as it did in Seattle over 15years, almost all of it happened at a tiny portion of the streetsegments in that city.So that portion of any distribution,

    • 02:43

      LAWRENCE SHERMAN [continued]: which we call the power few, is somethingthat Tokyo identified in looking at the concentrationsof violent crimes at areas mostly around railroad stationsand mass transit stations, with clear evidence that if you wereto try to police the whole city of Tokyo equally,you would be wasting your time in places that

    • 03:05

      LAWRENCE SHERMAN [continued]: never have any violent crime.But you could be preventing violent crimeby concentrating police more where the violent crimeitself is concentrated.So that kind of targeting of places, offenders, victims,or gangs can allow us to identify concentrations--and possibly even the causes of those concentrations--

    • 03:26

      LAWRENCE SHERMAN [continued]: so that we can set a high priority,pay less attention to the low-priority targets,and get the overall crime rate downwith this targeting process of evidence-based decision-making.Which we can then use for a separate kindof decision-making, from testing--

    • 03:47

      LAWRENCE SHERMAN [continued]: from comparing two kinds of methodsand then asking which method works better.Which one costs less?Which one gets the best result for the same cost?Once we know the answer to that question and we say,let's do this one, then we can track whether it actuallyhappens-- and whether it helps us to get crime down where

    • 04:07

      LAWRENCE SHERMAN [continued]: and when we want it to go down, whether we are providingthe policing or other security to the places and timesthat we want to provide it.And what's this match between the risk of crimeand the prevention efforts of crime?That can all come through tracking,which creates a continuous process of looking at what

    • 04:31

      LAWRENCE SHERMAN [continued]: we know from published scientific evidence in journalsthat are read all over the world,in looking at the evidence from the in-house data in any policedepartment, using that to produce guidelinesand then to measure police outputs,the activities like arresting or patrollingthat they undertake to reduce crime.

    • 04:53

      LAWRENCE SHERMAN [continued]: And then to see whether the outcomes-- crimeor repeat offending or even auto accidentdeaths-- whether those outcomes get better.This is something that's been used in many fields,including elections.And it is something that we can say that, in a way,police have actually been leading on.

    • 05:13

      LAWRENCE SHERMAN [continued]: And now even business is talking about howthe Triple-T in policing could help to improvethe performance of businesses.So let's just go back slowly through the conceptof targeting.If we consider it's not just the places whichcan be called hot spots, the power few places with the most

    • 05:34

      LAWRENCE SHERMAN [continued]: crime, it's also when the crimes occur.And even hot spots are not very hot most of the time.But it's predictable, such as Friday or Saturday nightbetween 10:00 PM and 2:00 AM, that youmight have most of the violence in the whole week concentratedat those hot spots at that time.

    • 05:55

      LAWRENCE SHERMAN [continued]: Similarly, repeat offenders commit most of the crimes.But the most dangerous offenders maycause most of the harm-- most of the murdersor rapes or other serious crimes--which might make them higher priority than repeat offenders,who commit relatively minor crimes with high frequency.Repeat victims are a potential target.

    • 06:16

      LAWRENCE SHERMAN [continued]: And in fact, many victims go on to become offenders.And if we think about the way police deal with victims,not just as an obligation but alsoas a form of preventing those victims from becomingcriminals, we could do our targeting analysis and resourceallocation with a wider understanding of the benefits

    • 06:37

      LAWRENCE SHERMAN [continued]: that we'd receive.There are other kinds of targets here.Only 2% of the offenders under probation supervisionin Philadelphia produced 75 times more charges for homicidethan the lowest-risk 60% of the offenders in Philadelphia.That shows why it's useful to concentrate on the 2%

    • 06:60

      LAWRENCE SHERMAN [continued]: and to spend a lot less time on that 60%.So what do we do with that 2%, or 5%?Well, we need to test what works best.We need to have a defined comparison between twodifferent methods, with one of them perhapsbeing the current method and the other onebeing a new method where you try to hold

    • 07:21

      LAWRENCE SHERMAN [continued]: all other conditions constant.[Testing defined-A Fair Comparison-Between two different methodsE = Experimental Method (new)C = Control Method (current)- All else equal- Which one is better?- By what criteria?And that's the kind of scientific evidencethat we see.Not Sherlock Holmes with a magnifying glass,looking at fingerprints, but statisticianslooking at statistical distributionsto see if policing this problem that wayworks better than policing it this way.

    • 07:42

      LAWRENCE SHERMAN [continued]: Good evidence creates not just a trend in what'shappening with the new program, but it compares itto what would have happened without the program.[Randomized Controlled Trial RCT:COMPARISON or NET difference]And that dotted line is what we see as whatthe control group tells us.And in this case, it showed that whenpolice used a restorative justiceconference with violent offenders,

    • 08:04

      LAWRENCE SHERMAN [continued]: their crime went down.But similar offenders who were prosecuted instead,by a lottery choice, they had an increase in crime.So the conclusion is that the new program worked betterthan the current practice.And that's what we can then use to trackwhat the police are doing.

    • 08:25

      LAWRENCE SHERMAN [continued]: Are they delivering these restorative justiceconferences, rather than just prosecuting offenders?Are they patrolling in the high-crime hot spots,which we can measure by global positioning satellite dataand feed back to the officers?And if we look at this graph from Trinidad and Tobago--[National Hot Spots Average & Percent Cumulative MurderCharge (2013 15)In which an increase in the murder

    • 08:47

      LAWRENCE SHERMAN [continued]: was followed by an increase in police patroltime in the homicide hot spots of the island of Trinidad,we see that as the yellow line, the activity of policein those hot spots, increased, the homicide rate went down--and in fact, dropped, by early 2015, by 45%

    • 09:10

      LAWRENCE SHERMAN [continued]: below the year before.And then as often happens, it came back up a little bitto create more of the kind of fluctuationthat you see in the stock market or anything else--but with a long-term shift to a lower homiciderate with well-tracked data about what the police are

    • 09:31

      LAWRENCE SHERMAN [continued]: doing, where and when it's needed most to be done.This is fed back, as you see in this picture here of a CompStatmeeting in a police station in Trinidad--not in a police headquarters but with the officers whoare actually doing the patrol, so theycan see whether they're doing what is most effective.

    • 09:51

      LAWRENCE SHERMAN [continued]: They can also see that if they stop doing it,the crime may rise, and they'd betterget back to doing the 120 minutes per day of patrolin homicide hot spots, rather than letting it slip downto 50 minutes per day or lower levels thatwere less effective in preventing homicides.

    • 10:12

      LAWRENCE SHERMAN [continued]: So there it is, the three Ts-- targeting, testing,and tracking.[Key points- Targeting- Testing- Tracking]In this tutorial, you should havelearned how those three Ts come together to empower the policeto use the evidence that is sitting in so many computersand has never been put together before in this way.And that's how we see the 21st century

    • 10:35

      LAWRENCE SHERMAN [continued]: as the knowledge-driven century for making a safer societyand supporting democracy in ways that were unimaginablea generation ago.[SAGE video]

Evidence-Based Policing

View Segments Segment :


Professor Lawrence Sherman discusses the Triple-T Against Crime: targeting, testing, and tracking. The three T's empower police to use evidence that they already have and look at it in a new way. Between the three, they find hot spots of crime, discover the best ways to deal with crime areas, then monitor how their solutions work.

SAGE Video Tutorials
Evidence-Based Policing

Professor Lawrence Sherman discusses the Triple-T Against Crime: targeting, testing, and tracking. The three T's empower police to use evidence that they already have and look at it in a new way. Between the three, they find hot spots of crime, discover the best ways to deal with crime areas, then monitor how their solutions work.

Back to Top