Crime Harm and Problem Oriented Policing Dr. Peter Neyroud Institute of Criminology
A Pracademic career Police Chief (Thames Valley and National Policing Improvement Agency) Academic Researcher, author and teacher at Cambridge 2
The Havant Policing Scheme 1980 as a bright young constable (Weatheritt, 1986)
The Havant Policing Scheme 1981 my design for local policing A conscious shift from the 3 R s to: Problem-solving teams Intelligence-led focus on crime and incident concentrations The use of technology to support targeting Enhancing the supervision to ensure tracking of activity
Incident Pattern Analysis 1984 the maps can show specific institutions or locations requiring special attention
The Goldstein s Point is POP: I read the book and tried to do it At the heart of this more honest approach to policing is the realization that the objective in attempting to bring about change is not simply to improve the police but rather to solve community problems p. 179) Measuring something meaningful with whatever data you have Using it to allocate police resources Measure success in reducing harm
Why did those initiatives fail? Targeting Havant Policing Scheme targeted areas/beats rather than microplaces We never measured the impact on fear of crime and legitimacy The data was neither big enough nor the technology powerful enough to analyze Testing None of them were tested experimentally or replicated Tracking: We did not rigorously track and manage Implementation..low dosage and low fidelity
Not all problems or crimes are created equal
13th November 2007 Targets let dangerous criminals escape net Police are neglecting to tackle serious, violent crimes and focusing instead on more minor offences as they strive to meet government targets, the man charged with shaping the future of policing in England and Wales has admitted. Peter Neyroud, Chief Executive of the National Policing Improvement Agency, said that over the past five years police had focused on increasing the number of offences brought to justice. But the former chief constable admitted that this meant that catching a murderer carried no more importance than apprehending someone who had stolen a bottle of milk.
Goldstein on Problem Identification Behavior Territory Persons Time What should get the highest priority? (p. 77) Impact of the problem Presence of Lifethreatening conditions
Goldstein s search for alternative strategies Concentrating attention on those Individuals who account for a disproportionate share of the problem (p.104) Places which account for a disproportionate share of the problem Victims who experience a disproportionate share of the problem
Developing the ideas: the power few and harm reduction
Danish Crime is Down
What About Harm?
The Cambridge Crime Harm Index
Producing Sentencing Guidelines
Calculating CCHI Values CCHI score = 365 X 5 = 1825 CCHI Screenshot from: http://www.sentencingcouncil.org.uk/wp-content/uploads/final_sexual_offences_definitive_guideline_content_web1.pdf
Screenshot from: http://www.sentencingcouncil.org.uk/wp-content/uploads/final_sexual_offences_definitive_guideline_content_web1.pdf
Screenshot from: http://www.sentencingcouncil.org.uk/wp-content/uploads/final_sexual_offences_definitive_guideline_content_web1.pdf
Calculating CCHI Values CCHI score = 365 X 5 = 1825 CCHI Screenshot from: http://www.sentencingcouncil.org.uk/wp-content/uploads/final_sexual_offences_definitive_guideline_content_web1.pdf
Calculating CCHI Values for crimes where the starting point is a community order 80 hours unpaid work 8 hours per day Gives a CCHI value of 10 for ABH Screenshot from: http://www.sentencingcouncil.org.uk/wp-content/uploads/mcsg-april-2017-final-2.pdf
From Sentencing Guidelines to CCHI No of days in prison
Findings: Sentencing analysis 40,100 880,000 Court sentences Western Australia 2010-2017 Sentence outcomes: 46,100 First time offenders / offences 29,700 Imprisonment Detention Community sentence Monetary fine 40,100 Offences within 103 sampled categories Converted to CHI days Tallied per offence category
Findings: Validity CHI comparison (House, 2017) Offence WA Max WA CHI UK CCHI UK CSS NZ CHI NZ JSSS Murder 9125 5713 5475 7979 1629 12045 Manslaughter 7300 2203 3825 7979 1687 1983 Attempted murder 2555 2737 1460 4663 2187 2849 Grievous bodily harm 3650 1419 1460 1965 425 892 Assault occasioning bodily harm 3650 38 20 184 108 192 Common assault 745 19 1 16 13 12 Sexual assault (Rape) 7300 1293 1825 2895 1172 3627 Deprivation of liberty 3650 235 548 123 330 1107 Armed robbery 5110 830 365 746 742 1738 Robbery 3650 572 365 746 155 475 Dwelling burglary 7300 338 20 438 63 171 Motor vehicle theft 2555 35 5 124 12 177 Theft 5110 5 2 86 10 25 Fraud 2555 68 10 200 48 170 Damage (Arson) 9125 526 33 185 110 474 Graffiti 993 6 2 7 5 1 Damage (Property) 3650 7 2 7 45 58 Disorderly conduct (public fear) 59 5 5 10 2 2 Breach of restraining order 365 5 6 54 9 73 Threatening behaviour 2555 10 10 280 4 14 Correlation WA CHI and UK CCHI = 0.93 WA CHI and UK CSS = 0.90 WA CHI and NZ CHI = 0.82 WA CHI and NZ JSSS = 0.93 WA Max and all others =0.41 to 0.56
What does this mean in practice?
Triple-T Against Harm Testing Targeting Tracking Evidence- Based Policing
Triple-T Against Harm Testing Targeting Tracking Evidence- Based Policing
Campbell Collaboration Systematic Review http://www.campbellcollaboration.org/lib/pr oject/24/
Hotspots policing has a small but significant effect on crime counts but what about crime harm?
Harm-spots v Hotspots
The Law of Crime Concentration: Birmingham: 5% of Places = 50% of Crimes
CHI Harm Spots same scale
Hot vs. Harm Spots
What s The Difference? Hot Spots are less concentrated than harm spots they cover more area of land Harm spots are taller than hot spots More of the harm in Birmingham is concentrated into less land than the same percentage of crime counts Police can target fewer places, less space, with for more harm reduction using harm spots than by using hot spots.
Findings: CHI Terrain Mapping
And when Hotspots become Harm-spots?
Context matters? V
Resource Allocation: Automatic Number Plate Recognition Hits
High and Low Crime Harm Offences/ Alerts High Crime Harm Offences/ Alerts Low Crime Harm Offences/ Alerts
Mean Crime Harm Per Alert Mean Crime Harm Per Alert Mean Crime Harm Per Alert Difference in Crime Harm by Outcome 140 100 106.3 120 100 80 60 40 20 0 112.3 No Response 80 60 66.9 40 59.4 20 p = 0.055* p = 0.245 0 Response Not Intercepted Intercepted 100 105.8 80 60 40 45.6 20 p = 0.069* 0 No Action Criminal Justice Action
Targeting Most Dangerous People: Philadelphia, UK High Risk (2%) Neither High nor Low Risk (38%) Low Risk (60%)
Victims:
Harm, Problem Oriented Policing and measuring success measuring effectiveness entails far more than counting familiar variables that lend themselves to counting, such as numbers of arrests, numbers of reported crimes and numbers of calls for assistance new measurements of effectiveness can be developed. (p.147)