Become a ProblemSolving Crime Analyst

In 55 small steps

1

Ronald V. Clarke

John Eck

Jill Dando Institute of Crime ScienceUniversity College London

Become a Problem Solving Crime Analyst

In 55 small steps

Target/victim

Guardian

Place

Offe

nder

ManagerHa

ndle

r

CRIME

Published by

Jill Dando Institute of Crime ScienceUniversity College London29/30 Tavistock SquareLondon WC1H 9QU

© Ronald Clarke and John Eck 2003

All rights reserved: no part of this publication may be reproduced, stored in a retrievalsystem, or transmitted in any form or by any means, electronic, mechanical, photocopying,recording or otherwise without the prior written permission of the Publishers or a licencepermitting copying in the UK issued by the Copyright Licensing Agency Ltd, 90 TottenhamCourt Road, London, W1T 4LP

First published 2003

ISBN 0-9545607-0-1

British Library Cataloguing-in-Publication DataA catalogue record for this book is available from the British Library

Project management by Deer Park ProductionsDesign and typeset by Pantek Arts LtdPrinted and bound by Ashford Colour Press

Ronald Clarke

Ronald Clarke is Professor of Criminal Justice at Rutgers, TheState University of New Jersey, and Visiting Professor at the JillDando Institute of Crime Science. He worked for many yearsin the Home Office Research and Planning Unit where he con-tributed to the development of situational crime preventionand the British Crime Survey. He has recently undertaken casestudies of problem-oriented policing together with HermanGoldstein. He helped to develop the Center for Problem-Oriented Policing and the Problem-Oriented Guide Series forthe US Office of Community Oriented Policing Services. DrClarke is chair of the judges for the annual Herman GoldsteinAward for Excellence in Problem-Oriented Policing. He can bereached at: rclarke@andromeda.rutgers.edu

John Eck

John Eck is Professor of Criminal Justice at the University ofCincinnati. He has contributed to the development of problem-oriented policing since 1984 when he studied the first full scaleattempt to implement the concept in the United States atNewport News, Virginia. He helped to develop a number of nowstandard techniques in problem solving, including the SARAmodel and the crime analysis triangle. He was a cofounder ofthe International Problem-Oriented Policing Conference heldannually in San Diego. Dr Eck has served as consultant to theLondon Metropolitan Police and since 1999 has been a judge forthe Tilley Award for Excellence in Problem-Oriented Policing. Hecan be reached at: john.eck@uc.edu

In assembling this manual we have drawn freely on the work of many colleagues and friends towhom we owe a large debt of gratitude. Those who have commented on drafts of or suppliedmaterial for the manual include: Pat and Paul Brantingham, Andy Brumwell, Rachel Boba, KateBowers, Barbie Brookover, Spencer Chainey, Paul Ekblom, Graham Farrell, Bonnie Fisher,Herman Goldstein, Alex Hirschfield, Shane Johnson, Scott Keay, Johannes Knutsson, GloriaLaycock, Mangai Natarajan, Michael Maxfield, Ken Pease, Nanci Plouffe, Barry Poyner, JerryRatcliffe, George Rengert, Nick Ross, Kim Rossmo, Karin Schmerler, Nick Tilley, Julie Wartell, BarryWebb and John Wooldredge. We owe special thanks to Marcus Felson who helped design themanual and who supplied material for some of the sections. Finally, we thank the Home Office forsponsoring our work.

Acknowledgements

Prepare yourself1. Read this first

2. Rethink your job

3. Be the local crime expert

4. Know the limits of conventional policing

Learn about problem-oriented policing5. Become a POP expert

6. Be true to POP

7. Be very crime specific

8. Be guided by SARA – but not led astray!

Study environmental criminology9. Use the crime triangle

10. Never forget opportunity makes the thief

11. Always ‘think thief ’

12. Expect offenders to react negatively

13. Don’t be ground down by the displacement pessimists

14. Expect diffusion of benefits

Scan for crime problems15. Say Cheers! when defining a problem

16. Know what kind of problem you have

17. Study the journey to crime

18. Know how hot spots develop

19. Learn if the 80–20 rule applies

Analyse in depth20. Formulate hypotheses

21. Diagnose your hot spot

22. Know when to use high-definition maps

23. Pay attention to daily and weekly rhythms

24. Take account of long-term change

25. Know how to use rates and denominators

26. Identify risky facilities

The 55 Steps

27. Be ready for repeat victimisation

28. Consider repeat offending

29. Know which products are CRAVED by thieves

30. Look for crime facilitators

31. Check you have answered the five ‘W’ (and one ‘H’) questions

Find a practical response32. Accept your key role at response

33. Increase the effort of crime

34. Increase the risks of crime

35. Reduce the rewards of crime

36. Reduce provocations

37. Remove excuses for crime

38. Find the owner of the problem

39. Choose responses likely to be implemented

Assess the impact40. Conduct a process evaluation

41. Know how to use controls

42. Consider geographical and temporal displacement

43. Examine displacement to other targets, tactics and crime types

44. Watch for other offenders moving in

45. Be alert to unexpected benefits

46. Expect premature falls in crime

47. Test for significance

48. Calculate costs

Communicate effectively49. Tell a clear story

50. Make clear maps

51. Use simple tables

52. Use simple figures

53. Design powerful presentations

54. Become a good presenter

55. Contribute to the store of knowledge

You who read this manual are more important than perhapsyou think. Crime analysts are not well-known to the generalpublic. You don’t star in peak-time TV series or big-screenmovies as do behavioural profilers or forensic scientists. Evensome of your colleagues in the police aren’t sure what you’reabout. But you are the new face of policing.

For years the police have contented themselves with chasingindividual crimes after they have taken place. Crimes havebeen regarded as episodes to be detected, and if they result ina conviction the case is thought to be ‘solved’.

This is manifestly mad. So mad, in fact, it is astonishing that soci-ety hasn’t rumbled it, complained very loudly and demanded asmarter approach. Running after crooks relentlessly is too late,like catching the horse (if you’re lucky) after it has bolted for thehundredth time rather than learning how to lock the stable door. It is as though when aircraftcrashed we contented ourselves with finding someone to blame rather than changing proceduresor amending designs.

When people do consider causes of crime they tend to talk of distant issues that cannot bechanged quickly (like parenting or poverty); they neglect the more immediate causes – thingsthat it is often quite easy to influence. Indeed, it is no exaggeration to say that answers to crimeare lying all around us waiting to be picked up. That is what this manual is all about.

But who will champion this new approach when almost all those in the crime industry have avested interest in the status quo? The media prefer to see crime as a series of individual humandramas which every now and then reward them with juicy headlines. Lawyers are steeped in tradi-tional ways of doing things (indeed they are taught that precedence is a virtue) and are broadlycontent with a system which puts them at the centre and feathers their nests. Most politicians,shuffling from one policy portfolio to another, reckon crime can be tackled intuitively which, forthem, means being tougher if they’re of conservative inclination, and softer if they’re liberal. Andmany criminologists have been too interested in theorising to be of any practical value to anyonebut themselves.

It has been left to a new breed of thoughtful police officers, plus a few diligent and unsung civilservants, and one or two enlightened politicians in high places, to recognise that a new approachis needed. This smarter way is based on the work of a precious minority of academics, many ofwhose names you will come across in these pages. They are mostly criminologists, but their brandof criminology is distinctive. For one thing it is intensely practical. It is concerned with outcomes.For another, it is much more truly scientific and evidence-driven than the impenetrable analysisthat sometimes passes for good work in social science essays.

In short, they are consultants in crime reduction. Yet there has been nothing to distinguish theseintensely practical people from the great majority of sociological theorists that populate schoolsof criminology and criminal justice. This led me to coin a term for them and for the new approachthey champion: crime science.

Crime science has three features. Its sole purpose is to reduce crime and so reduce victimisation.It is scientific in its methodology, by which I mean it aims for the same high standards of evidencethat would be accepted by physicists or aeronautical engineers. And thirdly it is multidisciplinary:

Foreword

it recruits every possible skill towards its cause. That is why in the following pages you will seeideas that come from geography, psychology, mathematics, epidemiology (the study of how dis-ease spreads), economics and many other schools of knowledge.

Not all the academics I admire call themselves crime scientists, though I hope one day they will,and as in any discipline there are healthy disagreements, not least about scientific methodologies.But the concept of crime science groups together those who care about hard evidence, who areconcerned with how to change things, not merely theorise. Crime scientists look for patterns incrime so that they can find ways to disrupt it.

One of the most important breakthroughs has been in rethinking the role of the police them-selves. It involves a change in emphasis from chasing criminals to outwitting them. The idea,developed by Herman Goldstein, is wonderfully simple: instead of being reactive to crime, con-centrate on your biggest problems and seize the initiative. The name given to the idea is equallystraightforward: problem oriented policing, or simpler still, POP.

In reality POP involves some pretty sophisticated stuff. It requires smart thinking. Crime sciencecan create new tools to make POP possible, but who are the professionals who will do the analysisof data and identify patterns at the local level? Who will construct hypotheses on how to interveneand put those ideas to the test? Who will create strategies that the police and other crime reduc-tion partners can then act on? Who can gently but firmly lead society away from being so reactiveto the new smart way of detecting baddies quickly and heading off trouble before it starts? Need Isay it? You.

Police analysts will become more important, indeed increasingly will be seen as crucial, if we areto tackle crime more intelligently. You are the brains, the expert, the specialist, the boffin.

Your role requires three qualities that do not always come naturally: application, scepticism, andpersistence. Application because much of what follows on these pages has to be learned andunderstood; it is counterintuitive, or at any rate is only self-evident once you have worked yourway through it. And application is essential in your day-to-day work because good science needsgood evidence and good evidence usually means hard work in finding, understanding and pro-cessing the data. Cutting corners is almost certain to turn you from a crime scientist into a conartist. You will need scepticism because that is the very foundation of good science. Never takeanything for granted. Beliefs come free, but evidence is costly, sometimes challenging our ownprejudices. Be sceptical reading this manual, be sceptical about what colleagues tell you (how dothey know, how reliable is their evidence, how much do they really understand the issues?), besceptical about data (do they really tell you what they purport to tell you?) and remember that sci-ence is not a set of subjects like physics or biology, but a rigorous methodology for testing andre-testing things we believe to be true. And you must be persistent because these are early days incrime science; some people will disregard you, others will think you are meddling in things thatdon’t concern you, most will need persuading.

Forgive us if you know much of this already. To those who are already well-versed in crime sciencethis manual may seem simplistic. But the truth is very few people yet understand these concepts,and to most readers nearly all of what follows will be new. I hope it will be not only extremelyuseful to you as a handbook, but exhilarating too.

Nick RossBBC Crimewatch UKChairman, the Jill Dando Institute of Crime Science, UCL

Foreword

This manual assumes that you are alreadyworking as an analyst and that you are accus-tomed to providing the kind of informationneeded to support police operations. Thismeans that:

1. you know how to use modern computingfacilities and how to access and manipulatecomprehensive databases;

2. you know how to use software to mapcrime, to identify hot spots and to relatethese to demographic and other data;

3. you routinely produce charts showingweekly or monthly changes in crime atforce and beat level, perhaps to supportCompstat-style operations;

4. you are accustomed to carrying out smallinvestigations into such topics as the rela-tionship between the addresses of knownoffenders and local outbreaks of car theftand burglary;

5. you have probably carried out some before-and-after evaluations of crackdowns, say onresidential burglaries or car thefts; and

6. you have some basic knowledge of statisticsand research methodology such as pro-vided by an undergraduate social sciencedegree.

This manual builds on this experience to pre-pare you for a different analytic role as a keymember of a problem-solving team. Indeed, thelatest writings on problem-oriented policingsee crime analysts as central to this new way ofpolicing communities. They argue that many ofthe weaknesses of current practice in problem-oriented policing result from the insufficientinvolvement of well-trained crime analysts ateach stage of the problem-solving process.

The manual prepares you for this new role byproviding you with a basic knowledge of prob-lem-oriented policing and the related fields of environmental criminology and situationalcrime prevention. These fields are encom-passed by the new discipline of crime scienceand you cannot adequately function as a problem-solving crime analyst without beingconversant with them. Nor can you adequatelyfunction in this role unless you rethink your

job, and the early sections of the manualexplain how you must take a more proactiveapproach. You cannot simply wait for yourpolice colleagues to come to you withrequests for information. Instead you musttake the initiative at every stage of the projectin defining the scope of the problem-solvingeffort, in trying to analyse the causes of theproblem, in helping to find an effectiveresponse and in setting up the project so thatit can be evaluated and the police can learnfrom the results.

The manual also assumes that analysts whotake on this new role are interested in con-tributing to the development of theirprofession. Assisted by vastly improved data-bases and powerful computing hardware andsoftware, crime analysis is on the verge ofbecoming an exciting new specialty. Indeed, ithas already begun to attract a cadre of well-trained and highly motivated professionalswho are vital to the development of policing inthe twenty-first century. You can make yourcontribution by communicating the results ofyour work in professional meetings and in thejournals. By doing so, you will not only helpyour profession and policing in general, butyou will become a more informed and valu-able resource to your own force.

The manual is short enough to get through in aweekend. It would be hard work and probablyworth doing, but it was not designed to be readand then shelved. Instead, we hope that you willfind it to be an indispensable reference sourcethat you will keep near your desk, consulting itwhenever you need in the course of a problem-solving project. This is why it is designed to berobust, allowing for continuous use. When openat a particular step it is designed to lie flat onyour desk so that you can consult it easily whenworking at your computer.

We have arranged the steps to follow logicallyone from another, in line with the SARA model(Scanning, Analysis, Response and Assessment),though each is self-contained and deals with aspecific topic. This should avoid your having toleaf through the manual, jumping from place to

1 Read this first

place, when dealing with a particular topic. Toget the best out of the manual you should bethoroughly familiar with the list of contents andyou should have browsed through sections thatinterest you to get an idea of the coverage. Butyou need only study a particular step when youhave an immediate need for the information itcontains. In any case, this is the best way tolearn: to seek and apply information when youhave a practical need for it.

In some cases, we do deal with a particulartopic in more than one place. For exampleStep 14 provides a general introduction to theconcept of displacement, while Steps 42 and 43explain how to check for various forms of displacement at the evaluation stages. Thecombined glossary and index at the end of themanual should help you find where a topic ismentioned in more than one place.

We have not referenced the manual as fully aswould be needed for an academic publication.There are several reasons for this. We havealready tried to distil the essentials of the liter-ature at each step. We also doubt that busycrime analysts will have much time for aca-demic reading. Lastly, few of you will haveready access to the specialised libraries thathold this material. But occasionally you willneed to know more about a topic, and at eachstep we identify key articles or books that youshould be able to obtain more easily. If youneed help with references, feel free to emailone of us at the addresses given earlier. We

would also be glad to receive any commentson the manual, especially suggestions forimprovement, which could be useful if we pre-pare later editions. Don’t be shy aboutsuggesting your own analyses for inclusion!

Embrace both SARA and NIM

Policing is beset by new fads that follow hot uponone another and almost as quickly disappear whensomething new arrives. Many seasoned officersplay along for a while, waiting for management tolose interest when they can get back to businessas usual. Neither problem-oriented policing nor theNational Intelligence Model should suffer this fate.The one provides a standard methodology fortackling specific recurring crime and disorderproblems harming a community. The other is astandard approach to the collection, analysis anddissemination of intelligence that will ensureuniform practice across the country. Both modelsput the crime analyst at centre stage because theytake it as given that policing must be evidence-led.The diagram below prepared by the NationalCriminal Intelligence Service shows how POPcomplements the National Intelligence Model. TheScanning, Analysis, Response and Assessmentstages (SARA) of a POP project sit easily alongsidethe intelligence cycle promoted by NIM: Collection,Evaluation, Collation, Analysis, Recommendationsand Review. Moreover, one of the four intelligenceproducts that NIM identifies are ‘problem profiles’,which are analyses of specific recurring problemswith recommendations for solutions. Soundsfamiliar, doesn’t it?

PREPARE YOURSELF

PR

EP

Business

Manage Crime

Manage Criminals

Manage LocalisedDisorder

Reduce OpportunitiesFor Crime

Manage Enforcementand Community Issues

Outcomes

Community Safety

Reduced Crime

Arrested/DisruptiveCriminals

Managed Hot-Spots

Potentially DangerousOffenders Controlled

KEY INTELLIGENCE PRODUCTS

STRATEGICASSESSMENTS

TACTICALASSESSMENTS

TARGETPROFILES

PROBLEMPROFILES

Tasking &Coordination

Group

PrioritiseIntelligence

Work

Level 1 - Local Issues

ANALYTICALPRODUCTS

SYSTEMPRODUCTS

KNOWLEDGEPRODUCTS

RESPONSESCANNING ANALYSIS ASSESSMENT

TASKING AND COORDINATING PROCESS

Like most crime analysts, you probably thinkof your job in rather modest terms. You do notsolve crimes single-handed. Nor do you takethe lead in finding new crime patterns andpersuading others to deal with them. Instead,you crunch data for those who do the ‘real’work of finding better ways to arrest criminals.You respond to requests for the latest statisticson burglary or car theft from beat officers andsergeants. You map crime for weekly meetingsso that the superintendent knows where todemand more effort. And you compilemonthly statistics that others need for theirreports. In other words, you sit in the backseat while others do the driving, asking foryour help only when they need it.

This manual will help you rethink your role.Even someone sitting in the back seat can helpthe lost driver find direction. Control overinformation is crucial, and the ability to analyseit is all-important. The person who learns howto do so becomes an essential member of theteam. But we are not talking here about poweror status. We are referring instead to a chal-lenge facing all police forces: how to solveenduring and repetitive crime problems. Thinkof yourself as a member of a team helping tosolve these problems, with a particular role inthat team. As you use this manual you willbegin to see how to perform that role and youwill also see how essential it is.

To play that essential role, you need to knowmore. Surprisingly, the most important know-ledge you may need to add is not computerskills or mapping ability, important thoughthese are. You need to learn more about crimeitself, to become a resource to your depart-ment as an expert on crime in your local area.If there is a new burglary wave, you should bethe first to know and the first to tell. Run thestatistics, map them and get the basic factsyourself. If you wait, others will say what ishappening without any factual basis. Oncemore you will be relegated to the back seat.You are the facts person and you must findthings out as soon as possible, using the bestmeans possible. This will often mean going

beyond police data, and this manual will tellyou how to find and use other data sources,including interviews with victims and offend-ers and records of crime kept by businesses.Becoming a source of information is a firststep. The ideal is to also be a source of advice.Whether you can do this depends on yoursupervisor’s openness, but at least you canprovide options or use the suggestions ofothers to inform decisions.

As a crime expert you should also know howbest to control it. In particular you shouldknow what works in policing and what doesnot. How effective is random patrol? Howoften do police come upon a crime inprogress? How often are crimes solved laterthrough patient detective work or forensic evi-dence? How productive are stakeouts andsurveillance in terms of arrests? How much docrackdowns cost in officers’ time? What are thearrest rates for different kinds of crimes? Howmany crimes of different kinds are evenreported to the police? Knowing answers tothese questions will tell you why even themost hard-working officers are relatively in-effective in preventing crime and why anincreasing number of police forces are nowturning to problem-oriented policing.

The main purpose of this manual is to tell youabout problem-oriented policing and about thevital part you can play in its implementation. Ithelps you distinguish problem-oriented polic-ing from other forms of community policing. Itshows you how problem-oriented policing canbecome more effective by using environmentalcriminology and situational crime prevention(or more broadly crime science). It describeseach of the four stages of a problem-orientedproject – scanning for crime problems,analysing a specific problem in depth, respond-ing to the problem by implementing solutionsand assessing the results of the project – andgives examples of the data and information thatyou could provide at each stage.

These stages of a problem-oriented projectwill require you to remain working on a single

2 Rethink your job

project much longer than in your traditionalanalytic role. You can expect to stay with aproblem-solving project for weeks or months,rather than just the few hours needed to plot aburglary hot spot or provide a monthly report.Where a detailed assessment of results isneeded, your involvement might even stretchover more than a year. You may have toexplain this to officers who come to you forhelp. At first they may be surprised that youexpect to stick so long with a project, but soonthey will appreciate your commitment tomaking the effort worthwhile.

Your time has been wasted if you cannot com-municate the results of your work. Latersections of the manual give some suggestionsfor communicating more effectively by tellinga story using simple maps and tables. Your pre-sentations should try to lead to a course ofaction, but you must always explain the limitsof your data and tell officers where your rec-ommendations are based on best guessesrather than facts.

This manual cannot tell you everything youmust know to be a problem-solving analyst.You must seek constantly to enhance yourprofessional skills and learn about the latestdevelopments in relevant fields. You must readmore widely and explore other sources ofinformation. Additional readings are recom-mended throughout this manual, but you willalso have to find material for yourself. A goodway to do this is through networking with ana-lysts in other departments and by attendingprofessional meetings of analysts, police and

criminologists. And try to pass on lessons youhave learned by making presentations at thesemeetings of valuable or novel analyses youhave undertaken. In short, you should beginto see yourself as more than just a technician,skilled in manipulating and presenting data.You should become more like a researcher –albeit with a highly practical focus – one whois bringing the very best that science can offerto make policing more effective. By the sametoken, also recognise that you are part of an emerging profession, which you can help to develop.

Rethink your job

Become a crime expert

Know what works in policing

Promote problem solving

Seek a place on the project team

Learn about environmental criminology

Hone your research skills

Communicate effectively

Enhance your profession

Become a crime scientist!

PREPARE YOURSELF

PR

EP

Anthony Braga (2002). Problem-Oriented Policing andCrime Prevention. Monsey, New York: Criminal JusticePress.

Read more

How often have you been asked the followingkinds of questions in your work?

What are burglars taking from shops?Where do thieves fence their goods? Which are the most troublesome pubs?Which cars do joyriders favour?Do better-lit streets have less crime? What is the street price of heroin?Where do men find prostitutes?Who is victimising Asian shopkeepers? Is crime declining on any council estate?

Some you may not have been able to answer,others only after a special analysis. But sup-pose you had the answers to them and manymore at your fingertips? Suppose you were theexpert on crime in your force area? In fact,nobody else can fill that role:

l The individual officer on the beat is toobusy with local crime.

l The detective has specific cases to solve.l The sergeants and inspectors are super-

vising their own officers.l The superintendents are providing leader-

ship for their units.l The chief and deputies are busy with the

bigger picture.

In short, nobody can see the whole picture.But if you could become the expert on localcrime that would give you an importantchance to make your department moreinformed, efficient, and capable of using itsresources to reduce crime. It would providemore chance to warn citizens, to detectoffenders, and to institute prevention efforts.In short, you could help a lot of people bygathering the right information.

But to become the local crime expert youmust try hard to become more informed thanothers. You must talk to officers about whatthey are seeing. Remember the late shift mightnot see officers on the early shift, and those onone side of town might not see officers on theother. They often talk about exceptions, notthe rules, about what made them angry, notabout the routine. Yet the routine is the breadand butter of crime analysis.

Go out on patrol with officers as often as timepermits. Not only will you get to know more ofthe officers in your force, but you will also get amuch better ‘feel’ for their work and the prob-lems they face on the street. You will also getearly warning about any new problems they areencountering. Sitting in regularly with dis-patchers provides the same sort of benefits andwill give you fresh perspective on your work.

Crime scenes receive a good deal of attentionin serious crimes, but ordinary crime scenes aretoo often neglected. You may not have the timeto visit these often, but you can look closely atcrime reports. Make a practice of pulling abatch each week and going through them tosee if there is anything new. Reviewing crimedata in automated systems, whether calls forservice data or crime reports, can also be veryproductive. Pay particular attention to failedcrime attempts. Some offenders have a trial anderror process, as they try to find new ways toget something for nothing. Those trying tocheat ticket machines or ATMs do not alwaysfind a method that works. But when they findit, the word will spread. You should provide anearly warning system to the police and otherlocal officials, as well as the main network.

Very often a local crime problem is also foundelsewhere. Your force may begin to experiencea rash of thefts from building sites when thishas never been a problem before. But you canbe sure that somewhere else has suffered thisproblem. That’s why it is important to be alertto changes in crime targets and modusoperandi. The Internet is a good source ofinformation about what crime others areseeing. You should also ask your colleagues inother forces, especially nearby forces. Theymay be experiencing exactly the same prob-lem, with perhaps the same group ofoffenders involved.

Do not limit yourself to police, for many otherpeople know a lot about particular crime prob-lems. Housing officers often know what drugproblems are growing or fading. City engineerscan see blight developing and notice indica-tions of crime before they are apparent to others. Publicans know about underage

3 Be the local crime expert

drinking, poor serving practices and sloppymanagement (in other pubs, of course!). Buscompanies keep records of assaults, vandalismand other crime problems that plague them.Head teachers know all too well about bullyingand vandalism on school premises. Small busi-ness owners are especially likely to noteproblems involving their premises. For exam-ple, a chemist knows what is being stolen fromhis shop or whether intoxicated people arehanging out nearby. Private security guards areoften the first to know about a crime.Unfortunately, most of their contacts withpolice are about single incidents. But they oftenhave information that can contribute to yourunderstanding about general patterns of localcrime, including changes in those patterns.

Offenders themselves are surprising sources ofinformation. Although they might not want toadmit doing anything themselves, they usuallyare willing to talk about ‘how it is generallydone’. Many offenders are actually quite talka-tive about the craft of offending, and will tellyou exactly how they pick targets, fence valu-ables, what offenders are looking for thesedays, and the like. Asking your police col-leagues to obtain this information fromoffenders can sometimes be very useful.

Lastly, victims often know a good deal aboutcrime circumstances. For non-contactoffences, they cannot usually give a precisetime of offence. But you can still learn from aburglary victim where an offender broke in,what they know to be missing, what room orfloor was left alone, etc. And talking to othervictims can be equally revealing.

How to become expert on crime inyour area

Talk to officers about what they are seeing.

Look closely at crime reports and visit crime scenes.

Go regularly on patrol with officers and sit withdispatchers.

Check failed attempts at crime to learn exactly whathappened.

Talk to other local officials about specific crime problems.

Exchange information with business and private security.

Explore sources outside your area for changes in crimetargets and methods.

Ask officers to question offenders about their methods.

Ask officers to talk to victims about exactly when, where,and how.

Get away from your computer!

Thinking ahead and learning fromunsuccessful attempts

Crime analysts in Chula Vista, California, knew that thebuilding boom in their city could worsen the residentialburglary problem. The houses being built were intendedfor affluent couples that would be out for most of theday. Daytime burglaries were already the predominanttype of residential burglary for the city. The analystsdecided to examine the effectiveness of existing securityprecautions to see if any of them could be built-in tonew homes or suggested to homeowners. First, theycompared completed burglaries with unsuccessfulattempts for an 18-month sample of 569 homesthroughout the city. This indicated that deadbolts shouldbe installed on the side doors of new houses as well ason the front doors. Second, they interviewed 250victims and 50 burglars and discovered that not oneburglar attempted to enter a house by breaking adouble-glazed window. This led to the recommendationthat all windows in new housing be double-glazed andmeet strict forced-entry standards.

PREPARE YOURSELF

PR

EP

Completed Unsuccessful Effective?*Burglaries Attempts

Dusk-to-dawn light 28% 29% NO

Indoor light on 26% 29% NO

Indoor timer light 9% 11% NO

Deadbolt only on front door 28% 25% NO

Deadbolt on front and side doors 15% 29% YES

Outdoor motion detector 23% 36% YES

Radio/TV left on 9% 18% YES

Alarm company sign 19% 36% YES

* ‘Yes’ means present in a larger proportion of unsuccessful attemptsthan completed burglaries

One of the most important functions of polic-ing is to prevent crime and you should knowthe research on how effectively they do this.

Conventional policing uses three strategies toreduce crime:

1. General deterrence endeavours to createthe public perception that the risks andpenalties of offending are high, so anyoneconsidering such behaviour will refrain.

2. Specific deterrence attempts to commu-nicate a perception of high risk and penaltyto specific individuals, so they will refrainfrom committing crimes.

3. Incapacitation tries to remove activeoffenders from society. This preventscrimes that they would have committed ifthey were not locked up.

Though police often supplement these crimecontrol strategies with other programmes, theseadd-ons play a secondary role. This is unfortu-nate, as there is strong and convincing evidencethat conventional policing is not very effective,while some of these add-ons are highly effec-tive. Indeed, a comprehensive examination ofscientific research on policing by the UnitedStates National Academy of Sciences under-scores the limited crime reduction utility ofconventional policing and the greater effective-ness of a problem-oriented policing.

Conventional policing relies heavily onpatrolling, rapid response, and follow-upinvestigations. Considerable research hasbeen conducted on the effectiveness of thesetactics. Though they can be effective undersome restricted circumstances, the evidenceshows that they are not particularly effectiveas general, all-purpose procedures. Policingthat relies largely on these tactics is less effec-tive than it could be if it used a broader set oftools to achieve a greater range of crime pre-vention effects.

The famous Kansas City patrol experiment, aswell as some early Home Office research,

showed that random patrolling has little or noeffect on crime. This is because crime is still arelatively rare event and the chances of randompatrols being in the right place at the right timeto prevent a crime are very small. However,focused patrolling of hot spots can have a largeimpact on crime for short periods of timebecause hot spot patrols apply policing towhere it is most needed. It is important to notethat these patrols cannot be maintained forlong periods and hot spot patrolling acts as ashort-term palliative, not a long-term solution.

Rapid response to reports of crime has a negli-gible effect on arrests. That is because citizensoften delay reporting crimes to the police, ifthey report them at all. Most property crimesare discovered long after offenders have left,so rapid response is extremely unlikely toresult in an apprehension. Further, the firstimpulse on being victimised or witnessing acrime is not to call the police, but to seek com-fort or advice from a friend or relation.Research in the 1980s in the United Statesfound that people involved in a crime delayedfive minutes or more in half the cases beforecalling the police. These delays in reportinggive offenders a head start in escaping. So,only when offenders are still very near crimescenes will rapid police response have a mean-ingful chance of securing an arrest. Thesecircumstances are rare.

Follow-up investigations by detectives are notmuch more effective at producing arrests.When there are no witnesses, as is usually thecase with burglaries, car thefts and most otherproperty crimes, the chances of detectingoffenders are very small. Even when offendersconfront victims, they often protect theiranonymity so that the usefulness of the victimreports is often limited. By focusing on caseswith considerable evidence, or on highlyactive offenders, police may be more effectiveat producing arrests, but again, most cases andmost offenders do not fit into these categories.

Overall, there are four main reasons why con-ventional policing is so ineffective:

4 Know the limits of conventionalpolicing

1. Police are spread thinly and would be even if their numbers were substantiallyincreased. To be effective police need tostimulate crime prevention by others.Unfortunately, conventional policing doeslittle to stimulate protective actions by citi-zens or other organisations.

2. As discussed, police are given little usefulinformation in most cases. Many criminalevents are not reported and those that arehave very little information useful for iden-tifying suspects. However, this same datacan be used to identify crime patterns andidentify situations that provoke crime.

3. Penalties imposed through the criminal jus-tice system are not immediate. Offenderspay more attention to the risks closelyattached to crime opportunities, than torisks that will take months to become appar-ent. Offenders usually live in the ‘here andnow’ rather than in the ‘there and then’.Offenders are more likely to consider risk atthe moment they are deciding to commit acrime, based on the circumstances immedi-ately in front of them. But they consider anumber of other things as well: How diffi-cult is it to commit this crime? What will Iget from it? And can this behaviour beexcused? (See Steps 28 to 33.)

4. Over-reliance on the criminal justice sys-tems clogs it and makes it less effective. Inaddition, there is a large-scale attrition ofcases through the system. It therefore doesnot present a credible threat to determinedoffenders.

The limitations on conventional policing stemfrom the overuse of enforcement and the neg-lect of other effective approaches – in short,conventional policing is too narrowly basedand inflexible. It operates too much like a fac-tory production line making a standardproduct rather than like a professional servicethat tailors its product to the particular needsof clients. Problem-oriented policing suppliespolice with a method of responding to thediverse nature of crime problems and helpsthem to become more effective at preventingcrime than they have in the past.

PREPARE YOURSELF

PR

EP

Lawrence Sherman and John Eck (2002). Policing forCrime Prevention. In Evidence-Based CrimePrevention, edited by Lawrence Sherman andcolleagues. London: Routledge.

National Research Council (2003). Fairness andEffectiveness in Policing: The Evidence. Committee toReview Research on Police Policy and Practices.Edited by Wesley Skogan and Kathleen Frydl.Committee on Law and Justice, the Division onBehavioral and Social Sciences and Education.Washington DC: The National Acadamies Press.

Read more

When a serious crime occurs, the police areexpected to react immediately. They must pro-vide help and reassurance to victims and movefast to arrest offenders. Yet we have seen thatmany times the police are not able to arrestthe culprits and may not be able to secure aconviction when they do. We have also seenthat random patrolling, which the publicexpects, is not an efficient way to apprehendcriminals, even when guided by crime analysisto focus on high-risk times and places. Thismeans that much police work done to meetpublic expectations is of limited value in con-trolling crime.

If they knew these facts, people would not becontent for police to abandon patrol or down-grade their response to serious crimes. Rather,they would expect the police to find new andbetter ways to control crime, while continuingtheir traditional work. In fact, this is what thepolice leadership has been trying to do byexperimenting with Compstat, ‘zero tolerance’,community policing and problem-orientedpolicing (or problem solving as it is oftencalled). While crime analysts have a role in allthese innovations, problem-oriented policing(POP) thrusts them into the limelight and givesthem an important team function. That’s whyyou should learn about it.

Herman Goldstein originated the concept ofproblem-oriented policing in a paper pub-lished in 1979. His idea was simple. It is thatpolicing should fundamentally be aboutchanging the conditions that give rise to recur-ring crime problems, and should not simply beabout responding to incidents as they occur orforestalling them through preventive patrols.Police find it demoralising to return repeatedlyto the same place or to deal repeatedly withproblems caused by the same small group ofoffenders. They feel overwhelmed by thevolume of calls and rush around in a futileeffort to deal with them all. To escape fromthis trap, Goldstein said the police must adopta problem-solving approach in which theywork through the following four stages:

1. Scan data to identify patterns in the inci-dents they routinely handle.

2. Subject these patterns (labelled problems)to in-depth analysis of causes.

3. Find new ways of intervening earlier in thecausal chain so that these problems are lesslikely to occur in the future. These newstrategies are not limited to efforts to iden-tify, arrest and prosecute offenders. Rather,without abandoning the use of the criminallaw when it is likely to be the most effectiveresponse, problem-oriented policing seeksto find other potentially effective responses,alone or in partnership with others, with ahigh priority on prevention.

4. Assess the impact of the interventions and,if they have not worked, start the processall over again.

SARA is the acronym used to refer to thesefour stages of problem solving – Scanning,Analysis, Response and Assessment. Latersections of this manual will discuss these indetail, but you can already see why you have acentral role in POP. You are the person mostfamiliar with police data and you know howbest to analyse and map that data to identifyunderlying patterns. You may know better thananyone else in the department how to use datain evaluating new initiatives. If you make ityour business to become the local crimeexpert, you will also know where to find otherrelevant information about problems; where tofind information on the Internet and in special-ist literature about successful responses usedelsewhere; how to use insights from environ-mental criminology in developing a problemanalysis; and how to anticipate and measureany possible displacement. Without your day-to-day involvement at all four stages, the POPproject will not achieve a substantial and sus-tained reduction in the problem.

Many commentators have criticised the qualityof POP projects undertaken to date, eventhough the concept has been widely wel-comed by police. The greatest problems are

5 Become a POP expert

found at analysis and assessment, preciselywhere you could make your greatest contribu-tion. Indeed, from the very first, Goldstein hasargued that problem-oriented policingdepends crucially on the availability of high-level analytic capacity in the department – anargument repeated in his most recent publica-tions. In fact, he has been very supportive ofthe idea of writing this manual that isaddressed directly to the role of the crime ana-lyst in problem-oriented policing.

You might agree that you have a substantialrole in problem-oriented projects, but youmight ask how you could ever succeed in thatrole given the realities of your job. How couldyou devote the time needed for the kind ofcareful analyses required? How could youmake a long-term commitment to a project,when you are continually being asked to pro-duce statistical reports and maps immediately,if not before? How would you ever beaccepted as an equal member of the team,when you are a mere civilian? How could youfunction as an equal member when your bosswants to approve every analysis you suggestand wants to see all your work before it leavesthe unit? How could you restrain the naturalimpatience of officers to move to a solutionbefore the analysis is complete? How couldyou persuade them to consider solutionsother than identifying and arresting offenders?How would you deal with criticisms that youare more interested in ‘research’ than practicalaction? In short, you may be wondering whatplanet we are living on because it certainlyresembles nothing you have seen.

These are good questions, but we believe thatpolicing is changing and that you can helpspeed up these changes. There is increasingpressure on police to become more effectiveand the time is long past when chief consta-bles could say they would cut crime if onlythey had more resources. Now, they mustmake a detailed ‘evidence-based’ case forthese resources and must explain preciselyhow they would use them. Their performanceis being monitored more closely every day,and crime reductions achieved in New Yorkand elsewhere have undermined excuses forfailure. In short, there is no doubt that policewill become increasingly reliant on data toacquire resources and manage them effec-tively. By providing these data you can ride thistide of change to a more rewarding career inpolicing, though you will have to workpatiently to supply timely information in aform that is helpful to the organisation. If youdo this, and you remain firmly focused oncrime reduction, you and your profession willgradually move into a more central policingrole – and problem-oriented policing providesyou with the perfect vehicle.

LEARN ABOUT PROBLEM-ORIENTED POLICING

PO

P

Herman Goldstein (1979). Improving Policing: AProblem-Oriented Approach. Crime and Delinquency,April: 234–58.

Herman Goldstein (1990). Problem-Oriented Policing.New York: McGraw Hill.

Read more

Some police managers attracted to problem-oriented policing hedge their bets bycombining it with other strategies that enlistcommunity help in dealing with problems,such as community policing, crime reductionpartnerships and broken windows policing (ororder maintenance). These managers arelikely to claim that they are implementing‘community-oriented problem solving’ (COPS)or ‘problem-solving partnerships’, or that theirorder maintenance activities are a form ofcommunity policing. In fact, these strategiesare not really compatible and the attempt tocombine them can produce a mess.

As explained, problem-oriented policing is amethod for analysing and solving recurrentcrime problems, while community policingrepresents a solution to what is defined as thecentral difficulty of conducting police business– gaining the support of the local communityin helping to prevent crime and disorder.Community policing is therefore focused onthe means not the ends of policing, and itsstarting point is a single highly general ‘prob-lem’ of conducting police business. This‘problem’ is defined a priori rather thanemerging from a careful analysis of the every-day business of individual departments. Eventhe emphasis on working with communities,which the two approaches (and crime reduc-tion partnerships) share, is not reallysomething they have in common. This isbecause problem-oriented policing rarelyseeks partners among the community at large.Rather, it identifies specific partners whosehelp is needed in dealing with the problem inquestion. If this is a problem of assaultsaround bus stops, a necessary partner indeveloping a response will be the local buscompany. If the problem is shoplifting, thenthe cooperation of local shopkeepers will beneeded. Sometimes the community’s helpmay be needed in implementing solutions (forexample, in fitting deadbolts or not givingmoney to beggars), but rarely can the commu-nity help in analysing specific problems or indeveloping solutions.

These distinctions are most easily confusedwhen the focus of a problem-oriented projectis a dilapidated neighbourhood. In this case,the project should proceed by identifying thecollection of individual problems that togethermake up the greater one. Rather than attempt-ing to build a relationship with the communityat large, which would be the objective of acommunity policing project, the problem-oriented project should focus on solving thespecific problems of, say, drug houses, com-mercial burglaries, and pub fights. To theextent that members of the communitybecome productively involved in solving thesediscrete problems, they may be a rather differ-ent group of individuals in each case.

It is also important to understand the differ-ence between problem-oriented policing andbroken windows. Under the former, specificsolutions to the variety of problems con-fronting the police emerge from careful anddetailed analysis of the contributory causes ofeach. By contrast, ‘broken windows’ advocatesthe same general solution – policing incivili-ties and maintaining order – whenever crimeshows signs of becoming out of hand. Thisapproach is based on two principles, the firstof which is that small offences add up todestroy community life. Thus, a large numberof less serious offences, each of which is aminor irritant, together become a major one.For example, littering one piece of paper isnothing terrible, but if everybody does it theneighbourhood becomes a dump. The secondprinciple of broken windows is that smalloffences encourage larger ones. For example,abandoned and boarded up properties oftenbecome the scene for drug dealing and canspawn many other crimes, even murders. Thisimportant insight has led New York City andother places to pay much more attention topolicing against small offences.

All policing requires discretion, and brokenwindows policing requires some very impor-tant decisions to be made by officers on thestreet. (This is why it should not be confusedwith ‘zero tolerance’ which is a political

6 Be true to POP

slogan, impossible for the police to deliverbecause it would soon result in clogged courtsand an alienated population.) One has tofigure out which of the small offences multiplyinto more crimes and which do not. For exam-ple, New York City subway system managerslearned that young men jumping turnstiles totravel free often committed robberies withinthe system. Controlling the minor crimehelped reduce the major one. But the subwaymanagers also learned that those painting graf-fiti did not normally commit more seriouscrimes. Although their efforts to control graf-fiti were very effective (see Step 11), they didnot reduce robbery.

In helping to learn about these multipliers,your analysis can assist in deciding whichminor offences warrant more attention. Youcan play a similar supporting role in attemptsby your force to introduce community policingand crime reduction partnerships. But onlyproblem-oriented policing offers you thechance to play a central role in initiating, imple-menting and assessing projects. This is becauseproblem-oriented policing is data driven – andcollecting, analysing and interpreting data isyour business.

LEARN ABOUT PROBLEM-ORIENTED POLICING

PO

P

Differences between problem-oriented policing and other new strategiesFocus Objective Rationale Methods First steps

Problem- Specific, Remove the Prevention Undertake Identify problems oriented recurring causes of is more focused requiring attentionpolicing crime these problems effective than action-research

problems enforcement (SARA)

Community Minority Enlist local Communities Build trust Identify a minoritypolicing neighbourhoods communities can be the through contacts neighbourhood

in the fight eyes and ears with residents and appointagainst crime of the police and regular community police

community officersmeetings

Crime Local areas Harness Coordinated Form partnerships Form a multi-reduction resources of multi-agency with businesses, agency partnershippartnerships entire action is the community

community in most effective groups and localreducing way to deal governmentcrime with crime

Broken Deteriorating Halt slide of Nip trouble Policing Identify a windows neighbourhoods neighbourhood in the bud incivilities/order deteriorating

into serious maintenance neighbourhoodcrime

James Q. Wilson and George Kelling (1982). BrokenWindows. The Atlantic Monthly, March: 29–38.

Read more

Your force will sometimes have a blitz on a par-ticular crime such as car thefts or residentialburglaries, and you may be asked to map theseoffences or provide other data to support theoperation. But these categories are too broadfor problem-oriented policing. They includetoo many different kinds of crimes, all ofwhich need to be separately analysed. Forexample, ‘car thefts’ could include (in roughorder of seriousness):

l Stealing hubcaps for resale or badges forcollections.

l Breaking into cars to steal items left inside.

l Breaking into cars and stealing radios andother fittings.

l Joyriding by juveniles.

l Taking a car for temporary transport.

l Stealing a car for use in another crime.

l Stealing and keeping a car.

l Stealing cars for ‘chopping’ and sale of theirparts.

l Stealing cars for resale.

l Stealing cars for export overseas.

l Car-jacking.

You can see these crimes are committed for avariety of motives, by different offenders, withvarying degrees of organisation, knowledge andskills. Stealing hubcaps is the least difficult anddaring and is committed by juvenile wanabees.Joyriding requires more courage and some basicknowledge about starting and driving cars.Stealing cars for export is a much more compli-cated crime requiring high levels of organisation,with many more stages and people involved.The offenders are as likely to be dishonest busi-nessmen as career criminals. More ruthless,hardened criminals commit car-jackings.

These differences between crimes explain whythe solutions to each cannot be the same.Joyriding can be reduced by better built-in secu-rity, which explains why immobilisers are

beginning to bring down overall levels of cartheft. However, immobilisers cannot preventcar-jacking because victims can be forced tohand over the keys if these are not already in theignition. In fact, some commentators believethat car-jackings have increased as thieves arenow breaking into houses to steal car keysbecause newer cars with immobilisers are diffi-cult to steal in the usual way. Immobilisers canalso be overcome by those with sufficient tech-nical skill and they may do little to reduce theftof cars for export. The solution to this problemmay lie in better port and border controls anddocuments that are harder to forge.

Breaking down a larger problem of crime intosmaller categories is merely the first step intightening the focus of a POP project. Forexample, a recent POP project in Charlotte,North Carolina, originally focused on down-town thefts from cars, became progressivelymore specific as the analysis of the problemunfolded. First, it became clear that the prob-lem was concentrated in the car parks. Only17% involved cars parked in residences or onthe streets. Then it was found, after countingparking spaces, that cars in surface car parkswere six times more at risk than those in multi-storey car parks, which were generally moresecure. This meant the project could focus onimproving security in the surface car parksthrough better lighting and fencing, and moresupervision by attendants. This would bemuch easier than trying to reduce the alreadylow levels of theft in multi-storey car parks.Tightening the focus of a POP project in thisway increases the probability of success anduses resources effectively.

There are few rules for determining preciselythe level of specificity needed for a successfulPOP project. Tightening the focus too muchcould result in too few crimes being addressedto justify the expenditure of resources, thoughthis depends on the nature and seriousness ofthe crimes. If only a few hubcaps are beingstolen, then this problem would not merit afull-blown POP project. On the other hand, aPOP project to reduce corner store robberies

7 Be very crime specific

could be worth undertaking, even if only a fewsuch robberies occur each year, because thesecan escalate into worse crimes such as murder,and because they increase public fear.

‘Because so much effort has beenconcentrated on crude groupings of crimetypes, such as burglary, robbery or auto theft,it has been virtually impossible to find trulycommon facts about the conditions whichlead to each of these groups of crimes. Thisimplies that we have to be very patient and tryto solve the problems of crime gradually andprogressively, piece by piece.’

Source: Barry Poyner (1986). A Model for Action. InSituational Crime Prevention, edited by Gloria Laycockand Kevin Heal. London: HMSO.

Some serious crimes, such as school shootings,are so rare that they cannot be properlyaddressed at the local level by problem-orientedpolicing. This is because the methodologydepends upon a certain level of repetition topermit underlying causes to be identified. Forthese kinds of crimes, police forces must ensurethat routine security measures are in place andthat they have a well-worked out plan forresponding to incidents.

While one should avoid beginning with a solu-tion, some solutions for specific crimes are sopromising that they might help define thefocus of a POP project. To return to the exam-ple of robbery at corner stores, there is goodresearch showing that having at least twomembers of staff on duty can reduce late nightrobberies of these stores. You could thereforetake a look at how many corner store robberiesoccur late at night in your area. If there wereenough of them, you might persuade yourdepartment to mount a POP project focusedon these late night robberies simply becauseyou know that an effective solution exists.

Finally, as you learn more about a problem inthe analysis stage, you might decide that it isso similar to a related problem that it is worthaddressing the two together. For instance,when working on a problem of assaults on taxidrivers, you might discover that many of theseare related to robbery attempts and that itwould be more economical to focus your proj-ect on both robberies and assaults. In this wayyou may identify a package of measures thatwould reduce the two problems together.

Being more specific aboutresidential burglary

Barry Poyner and Barry Webb have arguedthat preventing residential burglaries targetedon electronic goods requires quite differentmeasures from those to prevent burglariestargeted on cash or jewellery. This is becausethey found many differences between thesetwo sorts of burglaries in the city they studied.When the targets were cash or jewellery,burglaries occurred mostly in older homesnear to the city centre and were apparentlycommitted by offenders on foot. When thetargets were electronic goods such as TVs andVCRs, the burglaries generally took place innewer, more distant suburbs and werecommitted by offenders with cars. The carswere needed to transport the stolen goodsand had to be parked near to the house, butnot so close as to attract attention. The layoutof housing in the newer suburbs allowed theseconditions to be met, and Poyner and Webb’spreventive suggestions consisted principally ofmeans to counter the lack of naturalsurveillance of parking places and roadways.Their suggestions to prevent inner cityburglaries focused more on improving securityand surveillance at the point of entry.

Source: Barry Poyner and Barry Webb (1991). CrimeFree Housing. Oxford: Butterworth-Architecture.

LEARN ABOUT PROBLEM-ORIENTED POLICING

PO

P

Using POP police are required to: (1) closelydefine a specific, recurring problem, (2) conduct an in-depth analysis for a clear under-standing of the contributory causes, (3)undertake a broad search for solutions toremove these causes and bring about a lastingreduction in the problem, and (4) evaluatehow successful they have been. This is a formof ‘action research’, a well-established socialscience method in which researchers workalongside practitioners, helping to formulateand refine interventions until success isachieved. This can be contrasted with theusual research role, which is to work one stepremoved from the practitioners, collectingbackground information about problems andconducting independent evaluations of theirwork. In action research, however, theresearcher is an integral member of the prob-lem-solving team. This is identical to your rolein a POP project because your analyses mustinform and guide action at every stage.

You will find that SARA will help you and yourteam keep on track. This is the acronym for-mulated by John Eck and Bill Spelman to referto the four problem-solving stages ofScanning, Analysis, Response and Assessment.By dividing up the overall project into theseseparate stages, SARA helps to ensure that thenecessary steps are undertaken in propersequence – for example, that solutions are notadopted before an analysis of the problem hasbeen undertaken. This is a useful check on thetendency of police to jump straight to theresponse stage, while skimping on definitionof the problem and analysis.

Some commentators have criticised SARA foroversimplifying the POP process and forencouraging the idea that crime and disorderproblems are rather easily solved if SARA isapplied. For these critics, SARA masks somereal issues about the complexity of crimeproblems and the amount of thought andnegotiation that typically goes into developingand implementing new responses to prob-lems. We agree that many if not most POPprojects reported by police lack rigour andthat the analysis and assessment stages areoften perfunctory. But these failures cannot be

laid at SARA’s door. Rather, they reflect thepolice lack of analytic and evaluative skills –which is exactly why your contribution to POPis so badly needed.

Nor do we agree with a second criticism,which is that SARA fails to make clear that ifthe response is found ineffective, the processmust begin again. In action research the teamis expected to persist until success is achieved,refining and improving an intervention in thelight of what is learned from earlier failures.According to some critics, however, SARA sug-gests that the process is complete once theassessment has been made – and that thefunction of assessment is merely to documentthe successes likely to result from followingthe problem-solving process. This criticismcan be rather easily answered by always indi-cating feedback from assessment wheneverSARA is represented diagrammatically.

áScanning á Analysis á Response á Assessment

However, SARA can mislead in suggesting thatthe four problem-solving stages follow oneanother in a strictly linear fashion. The policeoften think that once a stage is completed –usually the analysis stage – it can be putbehind them and need not be revisited. Infact, projects rarely follow a linear path fromthe initial scanning and analysis stages throughthe stages of response and assessment. Rather,the process is iterative so that an unfoldinganalysis can result in refocusing of the project,and questions about possible responses canlead to the need for fresh analyses. The longerand more complicated the project, the moreiterations of this kind are likely to occur.

One of us (Ronald Clarke) recently workedwith Herman Goldstein on a project to reducethefts of kitchen appliances from housesunder construction in Charlotte, NorthCarolina. The housing developments wereoften in fairly isolated rural areas and wereimpossible to patrol effectively because therewere so many of them. They were difficult tosecure because the builders wanted to encour-age prospective buyers to tour the sites in theevenings and weekends. Because so few

8 Be guided by SARA – but not ledastray!

offenders were ever caught, we could discoverlittle about them and about how they disposedof the appliances. We considered a wide rangeof possible solutions, including the use ofportable alarms and CCTV cameras, storingappliances in secure containers on site andconcealing GPS tracking devices in them.Quite soon we hit on a solution being used bysome small builders, which was to delay instal-lation of the appliances until the day that thebuyer took possession. In the language of rou-tine activity theory, this would mean that thetargets of theft would no longer lackguardians.

Many builders were hostile at first to the idea.Sales staff believed that having the appliances in place made a home more saleable, and thatthe absence of appliances, if attributed to theft,might alarm purchasers about the area theywere moving into. Site supervisors felt that thelogistics of delivering and installing appliancesindividually as houses were occupied were considerably more difficult than batch deliveryand installation. Some erroneously believedthat building inspectors would not certify the houses as suitable for occupancy unless appliances were in place. Others (again erroneously) believed this was a mortgagerequirement. Finally, individual installationwould mean that builders could no longerarrange for building inspectors to visit a site andissue certificates of occupancy wholesale.

At first, these objections prevailed, but wesoon decided that the solution had so manyadvantages that we should see if answerscould be found to the builders’ objections.Here is the point of story, which is that seekinganswers meant that we had to begin docu-menting more carefully the objections raisedand the likely benefits of the solution we pro-posed. In other words, we had to revisit theanalysis stage to find detailed informationneeded for implementation of the response.This information was useful in persuadingbuilders to adopt the solution and thus reduc-ing the number of appliance thefts.

This shows how problem-oriented policing is aniterative process, in which the gradual acquisi-tion of data and information informs the project,

leading to more questions, to redefinition, andeven to changes in focus as it moves along. Assoon as a promising response is identified, itscosts and benefits need to be analysed in depth.The alternative of comprehensively exploring allavailable response options runs the risk that theproject will lose momentum and the support ofthose involved.

SARA and the ‘5Is’

Paul Ekblom of the Home Office has recentlyproposed the ‘5Is’, a development of SARA,which aims to capture, organise, and transferknowledge of good practice:

1. Intelligence – gathering and analysinginformation on crime problems and theirconsequences, and diagnosing their causes.

2. Intervention – considering the full rangeof possible interventions that could beapplied to block, disrupt or weaken thosecauses and manipulate the risk andprotective factors.

3. Implementation – converting potentialinterventions into practical methods,putting them into effect in ways that areappropriate for the local context, andmonitoring the actions undertaken.

4. Involvement – mobilising other agencies,companies and individuals to play theirpart in implementing the intervention.

5. Impact and process evaluation –assessment, feedback and adjustment.

The 5Is are supported by a wealth of otherpractical concepts and tools developed byEkblom including his ‘Conjunction of CriminalOpportunity’ framework, a development ofroutine activity theory. A summary of the 5Is isat www.crimereduction.gov.uk/learningzone/5is.htm

LEARN ABOUT PROBLEM-ORIENTED POLICING

PO

P

Ronald Clarke and Herman Goldstein (2002).Reducing Theft at Construction Sites: Lessons from aProblem-oriented Project. In Analysis for CrimePrevention. Crime Prevention Studies, vol. 13, editedby Nick Tilley. Monsey, New York: Criminal JusticePress (and Willan Publishing, UK). (Download from:www.cops.usdoj.gov)

Read more

Criminological theory is of little help in deal-ing with crime in the real world because itfinds causes in distant factors, such as child-rearing practices, genetic makeup, andpsychological or social processes. These aremostly beyond the reach of everyday practice,and their combination is extremely compli-cated for those who want to understandcrime, and do something about it. But you willfind that the theories and concepts of environ-mental criminology (and of the new disciplineof crime science) are much more helpful toeveryday police work. This is because theydeal with the immediate situational causes ofcrime events, including temptations andopportunities and inadequate protection oftargets. You will be a stronger member of theproblem-oriented team if you are familiar withthese concepts.

The crime triangle (also known as the problemanalysis triangle) comes straight out of one ofthe main theories of environmental criminology– routine activity theory. This theory formulatedby Lawrence Cohen and Marcus Felson statesthat predatory crime occurs when a likelyoffender and suitable target come together intime and space, without a capable guardianpresent. It takes the existence of a likelyoffender for granted since normal human greedand selfishness are sufficient explanations ofcriminal motivation. It makes no distinctionbetween a human victim and an inanimatetarget since both can meet the offender’s pur-pose. And it defines a capable guardian in termsof both human actors and security devices. Thisformulation led to the original crime trianglewith three sides representing the offender, thetarget and the location, or place (see darkshaded area in the diagram).

By directing attention to the three major com-ponents of any problem, the crime trianglehelps to ensure that your analysis covers allthree. Police are used to thinking about aproblem in terms of the offenders involved –indeed, their usual focus is almost exclusivelyon how to identify and arrest them. But POP

requires that a broader range of solutions isexplored and this requires information aboutthe victims and the places involved.

The crime triangle is the basis for anotheruseful analytic tool – a classification of thethree main kinds of recurring problems thatconfront police:

1. Repeat offending problems involve offend-ers attacking different targets at differentplaces. These are ravenous WOLF prob-lems. An armed robber who attacks a seriesof different post offices is an example of apure wolf problem.

2. Repeat victimization problems involve vic-tims repeatedly attacked by differentoffenders. These are sitting DUCK prob-lems. Taxi drivers repeatedly robbed indifferent locations by different people is anexample of a pure duck problem.

3. Repeat location problems involve differentoffenders and different targets interactingat the same place. These are DEN of iniq-uity problems or hot spots. A drinkingestablishment that has many fights, butalways among different people, is an exam-ple of a pure den problem.

Note that pure wolf, duck, and den problemsare rare. Most problems involve a mixture. Thequestion is, which is most dominant in a givenproblem, the wolf, duck or den?

The latest formulation of the crime triangle(see light-shaded area in the diagram) willhelp you to think about the response as wellas the analysis. This adds an outer level of ‘controller’ for each of the three originalelements:

l For the target/victim, this is the capableguardian of the original formulation ofroutine activity theory – usually people pro-tecting their own belongings or those offamily members, friends, neighbours andco-workers.

9 Use the crime triangle

l For the offender, this is the handler, some-one who knows the offender well and whois in a position to exert some control overhis or her actions. Handlers include parents,siblings, teachers, friends and spouses.

l For the place, the controller is the placemanager, a person who has some respon-sibility for controlling behaviour in thespecific location such as a bus conductor orteacher in a school.

The addition of the outer level of controllersturns the Wolf/Duck/Den classification into atheory of how these recurring problems arise:

l WOLF problems occur when offenders areable to locate temporarily vulnerable targetsand places. The controllers for these targetsand places may act to prevent futureattacks, but the offenders move on to othertargets and places. It is the offender-handlerbreakdown that facilitates wolf problems.

l DUCK problems occur when victims con-tinually interact with potential offenders atdifferent places, but the victims do notincrease their precautionary measures andtheir guardians are either absent or ineffec-tive. The handlers may prevent theoffenders from engaging in more of theseevents, and the managers may improvehow they regulate conduct at their places,but the victim encounters other offendersat other places.

l DEN problems occur when new potentialoffenders and new potential targetsencounter each other in a place wheremanagement is weak. The setting contin-ues to facilitate the problem events, eventhough handlers suppress offending, andguardians suppress victimisation.

Understanding how these recurring problemsarise will help you think about what might bedone not just to arrest offenders, but also toprevent them from re-offending by makingbetter use of handlers; what victims can do toreduce the probability of being targets; andwhat changes could be made to the placeswhere problems occur, be these schools, tav-erns or parking lots. In short, right from thebeginning, it helps you to avoid collecting dataabout every conceivable aspect of the prob-lem, but to focus instead on those aspectsmost likely to lead to practical solutions.

STUDY ENVIRONMENTAL CRIMINOLOGY

STU

DY

John Eck (2003). Police Problems: The Complexity ofProblem Theory, Research and Evaluation. In Problem-Oriented Policing: From Innovation to Mainstream.Crime Prevention Studies, vol. 15, edited by JohannesKnutsson. Monsey, New York: Criminal Justice Press(and Willan Publishing, UK).

Read more

Guardian

Place

Offe

nder

ManagerHa

ndle

r

CRIME

Target/victim

The crime triangle

For environmental criminologists, ‘opportu-nity makes the thief ’ is more than just apopular saying; it is the cornerstone of theirapproach. They believe that if opportunityincreases so will crime. More important, theyalso believe that if opportunity is reducedcrime will decline, which is why they advocatethe situational prevention measures discussedlater in this manual. To see if you agree thatopportunity (and temptation) is a cause ofcrime, consider the situation suggested byGloria Laycock and Nick Tilley:

Suppose all situational controls were to beabandoned: no locks, no custom controls,cash left for parking in an open pot foroccasional collection, no library check-outs, no baggage screening at airports, noticket checks at train stations, no trafficlights, etc., would there be no change in thevolume of crime and disorder?

If you answer that, of course, crime and disor-der would increase, then you too thinkopportunity is a cause of crime. Incredibly,most criminologists would not agree. Theybelieve that opportunity can only determinewhen and where crime occurs, not whether itoccurs. In their view, whether crime occurs iswholly dependent on offenders’ propensitiesand these propensities collectively determinethe volume of crime in society.

In fact, crime levels are as much determined bythe opportunities afforded by the physical andsocial arrangements of society as by the atti-tudes and dispositions of the population. Thisis difficult to prove without conducting anexperiment, but it would be unethical to createnew opportunities for burglary or robbery,then sit back to see what happens. However,experiments have been undertaken with moreminor transgressions. In America in the 1920s,researchers gave schoolchildren the opportu-nity to cheat on tests, to lie about cheating, andto steal coins from puzzles used. Otherresearchers have scattered stamped/addressedletters in the streets, some containing money,to see if these were posted. In a third group oflaboratory experiments, subjects have beeninstructed to ‘punish’ others for disobeying

test instructions by delivering severe electricshocks through the test apparatus. (In fact the‘victims’ were part of the research team and noshocks were delivered.)

The results of these experiments support thecausal role of opportunity. Most of the sub-jects, even those who generally resistedtemptation, took some opportunities tobehave dishonestly or aggressively – opportu-nities they would not have encountered butfor their participation in the studies. However,the transgressions studied were relativelyminor and you cannot generalise from them tocrimes of robbery or car theft. We thereforemust turn to some less rigorous but still con-vincing studies to show the importance ofopportunity in causing crime.

Suicide and opportunity. Suicide is not acrime, but like much crime it is generallythought to be a deeply motivated act. However,suicide trends in this country show a strongand surprising opportunity component. In the1950s, almost 50% of people killing themselvesdid so by domestic gas, which contained lethalamounts of carbon monoxide. In popular parl-ance, they put their heads in the gas oven.During the 1960s, gas began to be made fromoil not coal. The new gas had less carbonmonoxide and the number of gas suicidesbegan to decline. By 1968, only about 20% ofsuicides involved gas. This is when a secondchange began: the replacement of manufac-tured gas with natural gas from the North Sea.Natural gas contains no carbon monoxide andis almost impossible to use for suicide. By themid-1970s, less than 1% of suicides used thismethod.

What is deeply surprising is that gassing sui-cides did not displace wholesale to othermethods. Between 1968 and 1975, total sui-cides dropped by one third from 5,298 to 3,693(see figure). (This was during an economicdepression when suicide could have beenexpected to increase and, indeed, was increas-ing in other European countries.) People didnot turn to other methods because these allhave significant drawbacks. Overdoses requiresufficient pills to be amassed and, in any case,

10 Never forget opportunity makes thethief

these are much less lethal than carbon monox-ide. Hanging requires more knowledge as wellas courage. Not everyone has a gun, and thesecan result in disfiguring injuries not death.Domestic gas, on the other hand, is readilyavailable in most homes. It is bloodless andpainless and was highly lethal. It is not surpris-ing that it was the preferred method for solong. Nor is it so surprising that when theopportunity to use it was removed, the overallsuicide rate declined.

Murder and opportunity. Opportunity playsan important causal role in murder, one of themost serious crimes, as shown by a compari-son made a few years ago of homicide rateshere and in the United States. Crime rates,including those for assaults, differ littlebetween the countries, with the glaring excep-tion of murder. For 1980–84, the periodcovered by the study, the overall homicide ratein the United States was 8.5 times greater thanhere. The gun homicide rate was 63 times asgreat and the handgun homicide rate was 175times as great. In the whole of England andWales in this period (with about 50 millionpeople), only 57 murders were committedwith a handgun. In the United States, with a

population of about 230 million (less than fivetimes greater) a total of 46,553 people weremurdered with a handgun.

The difference in the homicide rates betweenthe two countries has narrowed during thepast few years, as their overall crime rates haveconverged, but there is still a much higher rateof murder in the United States. This is becausemany more people there own guns, especiallyhandguns. When they fight someone is morelikely to be shot. In other words, gun availabil-ity, an opportunity variable, plays an importantcausal role in murder.

Understanding the arguments in this section,and accepting that opportunity causes crime,does not mean you must deny the impor-tance of other causes, such as broken homesand inconsistent discipline. But it will helpdirect your attention to practical means ofpreventing crime, and help you defend themfrom criticism.

STUDY ENVIRONMENTAL CRIMINOLOGY

STU

DY

6,000

5,500

5,000

4,500

4,000

3,500

3,000

2,500

2,000

1,500

1,000

500

0

Num

ber

of s

uici

des

1958 1959 1960 1961 1962 1963 19641965 1966 19671968 1969 1970 1971 1972 1973 1974 1975 1976 1977

Year

Total suicides

Suicides by domestic gas

Suicides in England and Wales 1958–1977

Marcus Felson and Ronald Clarke (1998). OpportunityMakes the Thief. Police Research Series, Paper 98.London: Home Office.

Read more

Whenever you analyse a problem or thinkabout solutions, try to discover the reasonswhy the crimes are committed – not the dis-tant social or psychological causes, but theimmediate benefits of the crime for theoffenders involved. A radical critique of crimi-nology pointed out 30 years ago that bankrobbers are not propelled through the door ofthe bank by their genes; they rob banksbecause they want to get rich.

In many cases of theft and robbery the bene-fits will be obvious, but they may not be clearfor gang violence or so-called ‘senseless’ van-dalism and graffiti. In fact, graffiti can mark theterritory of a juvenile gang, can indicate wheredrugs can be purchased or can simply be a wayto show off. Knowing which of these reasons isdominant can help to define the focus of aproblem-solving project and unravel the con-tributory factors. It can also help the projectteam identify solutions. Thus, the New YorkCity subway only succeeded in eradicatinggraffiti when it understood the motivation ofthe ‘taggers’: in the words of George Kellingand Maryalice Sloan-Howitt who helped solvethe problem, this was to ‘get up’, to see theirhandiwork displayed day after day as the trainstravelled around the system. This insight ledto a programme of immediate cleaning of graf-fiti, which brought the problem under control.

Even in the case of theft the benefits are notalways obvious. For example, the most com-monly stolen items from American drug storesinclude painkillers, decongestants and antihista-mines. These are taken because theiringredients can be used to obtain or enhance a‘high’. Knowing that addicts are responsible formuch of the shoplifting in drug stores is helpfulin crafting a response, and your problem-solvingwork will always be helped by understandingthe motives of the offenders involved.

As important as knowing why offenderscommit the crimes is knowing how theycommit them. Rational choice theory, anothertool of environmental criminology, can behelpful in thinking about these questions. Thename is misleading because the theory doesnot assume that offenders plan their crimescarefully; it assumes only that they are seeking

to benefit themselves by their crimes, which isrational enough. The theory does not evenassume that offenders succeed in obtainingthe benefits they seek. This is because theyrarely have all the information they need, theydo not devote enough time to planning theiractions, they take risks and they make mis-takes. This is how we all behave in everydaydecision-making and is what theorists call lim-ited or bounded rationality.

Offenders must often decide quickly abouthow to accomplish their goals and how to getaway without being caught. Interviewing themcan help you understand how they make thesedecisions (see the Boxes). So long as you con-fine yourself to the general nature of theproblem you are trying to solve, and avoidspecific questions about crimes they havecommitted, you will be surprised how freelythey will talk. After all, we all enjoy talkingabout ourselves and about the work we do.

Martin Gill of Leicester University tells a story ofinterviewing an experienced offender in prison.When dealing with the crime that had led to hisarrest, Gill asked: ‘Did you think you’d get caught?’The prisoner leaned back in his chair and gave hima long look before saying: ‘I never expected to hearsomeone from a university ask such a stupidquestion. Do you think I’d have done it, if I thoughtI’d get caught?’

Quite often, however, you can get a long waywith your own imaginative construction of thecourse of a crime. What must be done at eachstage? How are targets selected? Victims subduedor tricked? Witnesses avoided? The policeescaped? The goods disposed of? Even if youcannot answer all these questions about modusoperandi, your attempt to enter the offender’smind can help you think about responses. This iswhat Paul Ekblom means when he advises prob-lem solvers to ‘think thief ’.

Pickpockets on the Underground told Paul Ekblomthat they would stand near signs warningpassengers that pickpockets were operating. Thesigns caused passengers to reassuringly patwhichever pockets contained their wallets. Thisconsiderably simplified the task for pickpockets.

11 Always ‘think thief’

Another rational choice theorist, DerekCornish of the London School of Economics,has developed a second concept that will helpyou unravel the sequence of decisionsinvolved in any crime. This is the ‘crimescript’. The underlying idea is that any particu-lar category of crime requires a set of standardactions to be performed in a particular order,just as in the script of a play. The scenes arethe sequential stages of the crime, the crimi-nals are the actors and the tools they use arethe props. Using this framework shows thateven an apparently simple crime, like stealinga car for temporary use from a parking lot,requires the offender to make decisions ateach step about how to proceed. Studyingthis sequence reveals many possible points ofintervention and the more completely youunderstand the crime script, the more helpyou can provide in identifying possibleresponses.

As well as conducting your own interviews, youcan also search the literature for reports ofinterviews with similar groups of offenders.Environmental criminologists have greatlyexpanded our knowledge about the methodscriminals use by interviewing car thieves, mug-gers, commercial robbers, residential andcommercial burglars, shoplifters and even EUfraudsters. Even though the offenders may notbe quite the same group as your own, carefullylooking at the results of published interviewstudies can help you understand your prob-lem, especially if you round out the picture byexamining the details of crime reports or thedistribution of the crimes concerned in timeand space. Knowing that residential burglarsgenerally enter houses through rear doors orwindows, and that they often return quite soonto burgle the same house again, not only givesyou insight into their decision making, but alsoimmediately suggests some interventions.

STUDY ENVIRONMENTAL CRIMINOLOGY

STU

DY

Armed robbers talkingMotives‘You are sitting there alone and you feeling light in yourpocket, your rent is due, light and gas bill, you gotthese bill collectors sending you letters all the time,and you say, “I wish I had some money. I need somemoney.” Those are the haints. [You haint got this andyou haint got that.] Your mind starts tripping cause youain’t got no money and the wolves are at the door…[After my last stickup] I gave my landlord some moneyand sent a little money off to the electric company, alittle bit off to the gas company. I still had like twenty orthirty dollars in my pocket. I got me some beer, somecigarettes, and [spent] some on a stone [of crackcocaine]; enjoy myself for a minute. I let the peopleknow I’m trying to pay you and they ain’t gonna beknocking on my door. Now I can do me legitimatehustles until the crunch comes again.’ (pp.43–44)

Advantages of robbery‘Robbery is the quickest money. Robbery is the mostmoney you gonna get fast… Burglary, you gonna haveto sell the merchandise and get the money. Drugs, yougonna have to deal with too many people, [a] bunch ofpeople. You gonna sell a fifty-dollar or hundred dollarbag to him, a fifty-dollar or hundred-dollar bag to him,it takes too long. But if you find where the cash moneyis and just go take it, you get it all in one wad. Noproblem. I’ve tried burglary, I’ve tried drug selling… themoney is too slow.’ (pp.51–52)

Choosing the victim‘See, I know the places to go [to locate good robberytargets]. Usually I go to all the places where dope menhang out… but I [also have] done some peoplecoming out of those instant tellers.’ (p.78)

‘That’s all I done robbed is drug dealers…they notgonna call the police. What they gonna tell the police?He robbed me for my dope? They is the easiest bait tome. I don’t want to harm no innocent people, I justdeal basically with drug dealers.’ (p.64)

Violence‘Well, if [the victim] hesitates like that, undecided, youget a little aggressive and you push them…I mighttake [the] pistol and crack their head with it. “Comeon with that money and quit bullcrapping or else yougonna get into some real trouble!” Normally when theysee you mean that kind of business they…come onout with it.’ (p.109)

Source: Richard Wright and Scott Decker (1997). Armed Robbers inAction. Boston: Northeastern University Press.

Read more

Derek Cornish (1994). The Procedural Analysis ofOffending and its Relevance for Situational Prevention. InCrime Prevention Studies, vol. 3, edited by RonaldClarke. Monsey, New York: Criminal Justice Press.(Download from: www.popcenter.org)

Almost all crime prevention involves changingoffenders’ perceptions of the opportunity forcrime. As the figure shows, prevention schemessometimes work directly on perceptions, aswhen police inform offenders that they arebeing closely watched. But most preventionschemes work through one or more intermedi-ate steps as in property marking schemes, forexample, where residents apply window stick-ers showing participation. Changes in offenderperception influence offenders’ behavioursthat, in turn, alter crime patterns.

In many cases, the preventive measures willdeter offenders from further criminal activity.They can also have the unintended effects of (1)reducing crime beyond the focus of the meas-ures, which is known as ‘diffusion of benefits’(see Steps 14 and 41) and (2) reducing crimebefore they have actually been implemented,known as ‘anticipatory benefits’ (Step 46).However, preventive measures do not alwaysachieve the desired effects, sometimes becauseoffenders are quite unaware of the interven-tions in place. For example, covert enforcementmay change the risks to offenders, withoutoffenders being aware of it. Consequently, they

will continue to offend. In other cases, offend-ers may adjust negatively to the preventivemeasures. These negative adjustments includedefiance, displacement and adaptation.

l Defiance occurs when offenders challengethe legitimacy of prevention efforts andcommit more offences rather than fewer. Ithas been suggested that some offenders actthis way in response to being arrested fordomestic violence. Defiance is more likelywhen the police are perceived to be unfairand heavy handed and there is evidencethat people are more law abiding whenpolice treat them fairly, even if the outcomeis not what people desire.

l Displacement occurs when offenderschange their behaviour to thwart preven-tive actions. Displacement is the oppositeof diffusion of benefits and both come infive forms, as the table illustrates.Displacement is a serious threat, but it is farfrom inevitable. Reviews show that manysituational prevention programmes showlittle or no evidence of displacement, andwhen displacement is found, it seldom fullyoffsets the prevention benefits (Step 14).

l Adaptation refers to a longer-term processwhereby the offender population as awhole discovers new crime vulnerabilitiesafter preventive measures have been inplace for a while. Paul Ekblom, Ken Peaseand other researchers often use the anal-ogy of an ‘arms race’ between preventersand offenders when discussing thisprocess. So, in time, we can expect manycrimes that have been reduced by preven-tive measures to reappear as criminalsdiscover new ways to commit them. A per-fect example is credit card fraud (see theBox). But not all preventive measures areso vulnerable to criminal ingenuity. Forexample, Neal Shover has argued that tech-nology has brought a lasting respite fromsafecracking, which is now very rare thoughit was once quite common.

12 Expect offenders to react negatively

PreventionScheme

IntermediateSteps

OtherFactors

Offenders’Perceptions

Offenders’Behaviours

Change inCrime

All prevention schemes work through offender perceptions

STUDY ENVIRONMENTAL CRIMINOLOGY

STU

DY

Displacement and diffusion of benefits: burglary of flatsType Definition Diffusion Displacement

Geographical (1) Geographic change Reduction in targeted building Switch to another buildingand in nearby buildings

Temporal (1) Time switch Reduced burglaries during Switch from day to eveningday and evening

Target (2) Switch object Reduced burglaries in flats Switch from flats to housesof offending and houses

Tactical (2) Change in procedures Reduction in attacks on Switch from unlocked doorsfor offending locked and unlocked doors to picking locks

Crime type (2) Switch crimes Reduction in burglary Switch from burglary to theftand theft

(1) See Step 42 (2) See Step 43

Paul Ekblom (1997). Gearing up Against Crime: a Dynamic Framework to Help Designers Keep up with the AdaptiveCriminal in a Changing World. International Journal of Risk, Security and Crime Prevention, 2: 249-265. Alsowww.homeoffice.gov.uk/rds/pdfs/risk.pdf

Read more

Other Card not Application Counterfeit Mail non- Lost and Totalpresent fraud receipt stolen

1991 1.6 0.4 2.0 4.6 32.9 124.1 165.61992 1.0 1.3 1.4 8.4 29.6 123.2 165.01993 0.8 1.6 0.9 9.9 18.2 98.5 129.91994 0.5 2.5 0.7 9.6 12.6 71.1 96.91995 0.3 4.6 1.5 7.7 9.1 60.1 83.31996 0.5 6.5 6.7 13.3 10.0 60.0 97.11997 1.2 12.5 11.9 20.3 12.5 66.2 122.01998 2.3 13.6 14.5 26.8 12.0 65.8 135.01999 3.0 29.3 11.4 50.3 14.6 79.7 188.32000 6.5 56.8 10.2 102.8 17.3 98.9 292.5

Offender adaptation and credit card fraud

In a series of papers, Michael Levi and his colleagues have described how a partnership betweenthe police, the Home Office and the credit card issuers led to successful action in the mid-1990s to reduce credit card frauds. The measures introduced include new lower limits forretailers seeking authorisation of transactions and greatly improved methods of delivering newcredit cards to consumers via the mail. As the table shows there was a resulting markedreduction in fraud losses. In recent years, however, credit card losses have begun to climb again.This is due principally to a growth in losses resulting from ‘card not present frauds’ (due to therapid expansion of Internet sales) and in counterfeiting of cards (said to be the work of organisedgangs in East Asia).

Credit card fraud losses, UK, £ millions

Problem-oriented policing often tries toreduce opportunities for crime. For example,window locks may be fitted to prevent bur-glary in a block of flats, or CCTV camerasinstalled to prevent thefts in parking lots.These ways of reducing opportunities forcrime often meet the same objection: all theydo is move crime around, not prevent it. Thistheory of ‘displacement’ sees crime as beingshifted around in five main ways:

1. Crime is moved from one place to another(geographical).

2. Crime is moved from one time to another(temporal).

3. Crime is directed away from one target toanother (target).

4. One method of committing crime isreplaced by another (tactical).

5. One kind of crime is substituted foranother (crime type).

In each case, the theory assumes that offendersmust commit crime, whatever impedimentsthey face. The basis for the assumption is eitherthat the propensity to commit crime builds upand must be discharged in the same way thatsexual release is sought, or that ‘professional’criminals or drug addicts must obtain a certainincome from crime to maintain their lifestyles.Whatever its basis, the displacement theoryneglects the important causal role of tempta-tion and opportunity in crime (Step 10).

Even in the case of more committed offend-ers, the displacement theory fails to giveenough importance to opportunity. Thus,research on drug addicts has shown that theyadapt to variations in the supply of drugs. Noris there any simple progression in drug use.Rather, addicts might be forced to use smalleramounts or less agreeable drugs because thesupply of drugs has been cut.

As for professional criminals like bank rob-bers, there is no reason to assume that theymust obtain a fixed amount of money fromcrime. They would surely commit fewer rob-beries if these became difficult and risky, justas they would commit more robberies if thesebecame easy. Bank robbers, like everyone else,may sometimes have to adjust to reduced cir-cumstances and be content with lower levelsof income.

This does not mean that we can ignore dis-placement. Indeed, rational choice theorypredicts that offenders will displace when thebenefits for doing so outweigh the costs. Forexample, once steering locks were introducedfor all new cars sold in Britain from 1971, oldercars without these locks were increasinglystolen. Since these cars were easy for offend-ers to find, this displacement was not asurprising outcome. But numerous other stud-ies have found that displacement did notoccur at all, or only to a limited extent. Forexample:

l New identification procedures greatlyreduced cheque frauds in Sweden, with noevidence of displacement to a range of‘conceivable’ alternative crimes.

l Extensive target hardening undertaken inbanks in Australia lowered robbery rates,but there was no sign that corner stores,petrol stations, betting shops, motels, orpeople in the street began to experiencemore robberies.

l Crime was not displaced to a nearby estatewhen the council improved street lightingfor a run-down housing estate in theMidlands.

l When a package of security improvementsreduced thefts in a multi-storey car park inDover, there was no evidence that theftswere displaced to other nearby car parks.

13 Don’t be ground down by thedisplacement pessimists

l When streets were closed in Finsbury Parkand policing was intensified, there was littleevidence that prostitutes simply moved toother nearby locations. According to theresearchers, many of the women workingthe streets in Finsbury Park were notdeeply committed to prostitution, but sawit as a relatively easy way to make a living.When conditions changed so did theirinvolvement and many seem to have givenup ‘the game’ (Step 44).

In these examples, the offenders’ costs of dis-placing seemed to have outweighed thebenefits and the examples bear out the argu-ment that displacement occurs much less thancommonly believed. This is the consensus ofthree different reviews of the displacement lit-erature undertaken in Canada, the UnitedStates and the Netherlands. The Dutch review(the most recent one) reports that in 22 of 55studies in which displacement was examined,no evidence of it was found. In the remaining33 studies in which evidence of displacementwas found, only some of the crime seems tohave been displaced. In no case was theamount of crime displaced equal to theamount prevented.

To sum up, displacement is always a threat,but there are strong theoretical reasons forbelieving that it is far from inevitable. In addi-tion, the studies of displacement show thateven when it does occur, it may be far fromcomplete and that important net reductionsin crime can be achieved by opportunity-reducing measures.

Claims of displacement oftenevaporate under closer scrutiny

London Underground officials believed thattheir success in modifying new ticketmachines to eliminate 50p slugs had simplydisplaced the problem to £ slugs, whichbegan to appear as soon as the 50p oneswere eliminated. However, analysis showedthat:

1. The scale of the £ slug problem (less than3,500 per month) never approached thatof the 50p slugs (95,000 per month attheir height).

2. The £ slugs were found in stations notpreviously affected by 50p slugs.

3. Any schoolboy could make a 50p slug bywrapping a 10p coin in silver foil. Onlypeople with the right equipment couldmake £ slugs by filling copper pipes withsolder and then slicing them carefully.

We can see that the two problems involveddifferent stations and offenders and that theclaim of displacement is dubious.

Source: Ronald Clarke, Ronald Cody and MangaiNatarajan (1994). Subway Slugs: Tracking Displacementon the London Underground. British Journal ofCriminology, 34: 122–138.

STUDY ENVIRONMENTAL CRIMINOLOGY

STU

DY

Rene Hesseling (1994). Displacement: A Review of theEmpirical Literature. In Crime Prevention Studies, vol. 3,edited by Ronald Clarke. Monsey, New York: CriminalJustice Press. (Download from: www.popcenter.org)

Read more

Researchers looking for displacement havesometimes found precisely its reverse. Ratherthan finding that crime has been pushed tosome other place or time, they have foundthat crime has been reduced more widelythan expected, beyond the intended focus ofthe measures. This is a relatively recent dis-covery, but already many examples exist:

l As expected, electronic tagging of books ina University of Wisconsin library resulted inreduced book thefts. However, thefts alsodeclined of video-cassettes and other mate-rials that had not been tagged.

l When a New Jersey discount electronicretailer introduced a regime of daily count-ing of valuable merchandise in thewarehouse, thefts of these items plum-meted – but thefts also plummeted of itemsnot repeatedly counted.

l When ‘red light’ cameras were installed atcertain junctions in Strathclyde, not onlydid fewer people ‘run the lights’ at theselocations, but also at other traffic lightsnearby.

l The implementation of added security forhouses that had been repeatedly burgledon the Kirkholt housing estate reducedburglaries for the whole of Kirkholt, notjust for those houses given additional protection.

l When street lighting was improved in alarge housing estate in England, crimedeclined in both that estate and a nearbyone where the lights were not changed.

l When vehicle tracking systems were intro-duced in six large American cities, rates oftheft declined citywide, not just for carowners who purchased the devices.

These are all examples of the ‘diffusion of benefits’ of crime prevention measures. Itappears that potential offenders may be awarethat new prevention measures have beenintroduced, but they are often unsure of their

precise scope. They may believe the measureshave been implemented more widely thanthey really have, and that the effort needed tocommit crime, or the risks incurred, havebeen increased for a wider range of places,times or targets than in fact is the case.

Diffusion of benefits is a windfall that greatlyincreases the practical appeal of situationalcrime prevention, but we do not yet knowhow to deliberately enhance it. One importantmethod may be through publicity. A publicitycampaign helped to spread the benefits ofCCTV cameras across an entire fleet of 80buses in the North of England, although thesewere installed on just a few of the buses. Oneof the buses with the cameras was takenaround to schools in the area and the firstarrests resulting from the cameras were givenwide publicity in the news media.

We should expect the diffusion of benefits todecay when offenders discover that the risksand effort of committing crime have notincreased as much as they had thought. Thisoccurred in the early days of the breathalyserwhich had a much greater immediate impacton drunken driving than expected given theactual increase in the risk of getting caught.However, as drivers learned that the risks ofbeing stopped were still quite small, drunkendriving began to increase again. This maymean that ways will have to be found of keep-ing offenders guessing about the precise levelsof threat, or quite how much extra effort isneeded if they are to continue with crime.

At a practical level, diffusion is important as acounter argument about displacement fromthose resisting the introduction of preven-tative measures. And you will certainlyencounter many of those! Secondly, it isimportant that you plan your evaluation totake account of diffusion. Otherwise, youmight find that people question the effec-tiveness of the preventive measures ongrounds that crime fell more dramaticallythan expected.

14 Expect diffusion of benefits

Diffusion of benefits and CCTV in auniversity parking lot

A new head of security at the University ofSurrey decided to deal with a plague of theftsin the university’s car parks by introducingCCTV. He installed a camera on a mast toprovide surveillance of the car parks. Asshown by the diagram, the camera could notprovide surveillance equally for all four carparks because its view of car park 1 wasobscured by buildings. It might have beenexpected, therefore, that if the CCTV had anyvalue in preventing crime this would only befor the car parks it covered adequately. Itmight also have been expected that crimewould be displaced by the camera from thesecar parks to the one not given propersurveillance. In fact, in the year following theintroduction of the CCTV, incidents of theftand vandalism in the lots were cut in half,from 138 in the year prior to 65 in the yearafter. Incidents declined just as much in carpark 1, not covered by the cameras, as in theother three car parks. This diffusion of thebenefits of CCTV probably resulted frompotential offenders being aware that it hadbeen introduced at the University, but notknowing its limitations. Many probablydecided that it was no longer worth the riskand effort of going to the university car parksto commit crime.

Source: Barry Poyner (1997). Situational Prevention inTwo Parking Facilities. In Situational Crime Prevention:Successful Case Studies, edited by Ronald V. Clarke.Monsey, New York: Criminal Justice Press.

STUDY ENVIRONMENTAL CRIMINOLOGY

STU

DY

Ronald Clarke and David Weisburd (1994). Diffusion ofCrime Control Benefits: Observations on the Reverse ofDisplacement. In Crime Prevention Studies, vol. 2,edited by Ronald Clarke. Monsey, New York: CriminalJustice Press. (Download from: www.popcenter.org)

Read more

CCTV

Security gate

University buildings

Parking lots

1

234

CCTV system

A problem is a recurring set of related harmfulevents in a community that members of thepublic expect the police to address. This defi-nition draws attention to six defining elementsof a problem: Community; Harm; Expectation;Events; Recurring; and Similarity. These ele-ments are captured by CHEERS:

l Community. Problems are experienced bymembers of the public. This includes indi-viduals, businesses, government agencies,and other groups. Troublesome eventswithin police agencies, not directly impact-ing members of the public, are notincluded. Police vehicle accidents, forexample, occurring on police property arenot ‘problems’ (though many of the tech-niques described in this manual could beapplied to police vehicle accidents). Notethat this element does not require thateveryone or even most members of a com-munity experience a problem, only thatsome community members do.

l Harmful. People or institutions mustsuffer harm. The harm can involve prop-erty loss or damage, injury or death, orserious mental anguish. Most events ofthis sort are violations of the law, but ille-gality is not a defining characteristic ofproblems. There are problems involvinglegal behaviour that the police must still address. Noise complaints arisingfrom the impact of legitimate commercialactivity on neighbouring residents is acommon example. Some problems arefirst reported as involving illegal behaviourthat on closer examination do not involveillegalities. Nevertheless, if they meet theall the criteria of the definition, they arestill problems.

l Expectation. Some members of the publicexpect the police to do something aboutthe causes of the harm. Again, the numberof people who expect the police to addressthe problem need not be large. Events thatare annoying only to police officials are notproblems, in the technical sense of thisterm. Though public complaints to thepolice are important indicators of expecta-tion, sometimes citizens have trouble

communicating with the police, or do notrealise the police are willing to and capableof addressing their concerns so problemsare hidden from the police. If the publicunderstood police capacity, their expecta-tions might change. Nevertheless, publicexpectation should never be presumed, butmust be evident.

l Events. Problems are comprised of dis-crete incidents, such a break-in to a home,one person striking another, two peopleexchanging money and sex, or a burst ofnoise. These are events. Most events arebrief, though some may involve a great dealof time – some frauds, for example. A prob-lem must have more than one event.

l Recurring. Having more than one eventimplies that events must recur. They maybe symptoms of an acute or a chronic prob-lem. An acute problem suddenly appears,as in the case of a neighbourhood with fewvehicle break-ins suddenly having manysuch break-ins. Chronic problems arearound for a long time, as in the case of aprostitution stroll that has been locatedalong one street for many years. Whetheracute or chronic, unless something is done,these events will continue to occur. Ifrecurrence is not anticipated, problem solv-ing may not be necessary.

l Similarity. Implied in the idea of recurringis that these events are similar or related.They may all be committed by the sameperson, happen to the same type of victim,occur in the same types of locations, takeplace in similar circumstances, involve thesame type of weapon, or have one or moreother factors in common. Without commonfeatures, we have a random collection ofevents instead of a problem. With commonfeatures, we have a pattern of events. Crimeand disorder patterns are often symptomsof problems.

Problems need to be defined with great speci-ficity because small details can make adifference between a set of circumstances thatgives rise to harmful events, and a set of circumstances producing harmless events.Solving problems involves changing one or

15 Say Cheers! when defining a problem

more of these small details. CHEERS suggestssix basic questions that need examination inthe scanning stage:

l Who in the community is affected by theproblem?

l What specifically are the harms created bythe problem?

l What are the expectations for the policeresponse?

l What types of events contribute to theproblem?

l Where and when do these events recur?l How are the events similar?

Not everything the police are asked to addressqualifies as a problem. The CHEERS conceptscan help identify demands that are not prob-lems. Again, we are using the term ‘problem’ inthe technical, POP sense, not as we would ineveryday speech. So things we define as ‘notproblems’ still may be troublesome, and still mayrequire police attention. These are as follows:

Single events. Isolated crimes, acts of disor-der or related phenomena, regardless of howserious, are not problems. These may deserveinvestigation or some other police action, butproblem solving cannot be applied to isolatedevents because they are not similar nor dothey recur.Neighbourhoods. Small areas, such as citycentres or particular housing estates, some-times get reputations as ‘problems’.Neighbourhoods are seldom problems, how-ever. Rather they are geographic areas thatmay contain multiple problems. These individ-ual problems might be related, but not always.Tackling an entire area as a single problemincreases the complexity of the effort andreduces the likelihood of finding effectiveresponses. Instead, you should identify spe-cific problems within the neighbourhood andtackle them individually. In some cases, ofcourse, there may be common solutions todistinct problems (see box). Status conditions. Truant schoolchildren,bored teenagers, vagrant adults and convictedcriminals are not problems because of theirstatus of not being in school, having nothing todo, not being employed or having been found

guilty of an offence. A community might expectthe police to do something about them butstatus conditions lack the characteristics ofharm and events. Some of these people mayplay a role in problems, as targets, offenders orin some other capacity, but that does not makethem a problem. Defining a problem by statusconditions is evidence of lack of precision and aneed to examine the issue in greater depth.Status conditions may, at best, point to pieces ofa larger problem, but they are not the problem.

Always define a problem using all elements ofCHEERS!

Separate problems, common solutions

Specific problems in a dilapidatedneighbourhood or council estate shouldalways be separately analysed, but, for cost-effectiveness reasons, solutions ought to beconsidered together. In the hypotheticalexample below, the last identified solution, aconcierge scheme and CCTV system, is themost costly of all those listed. But it is alsopredicted to be the most effective solution foreach problem. It might therefore be chosen asa solution to all three problems when costsmight have ruled out its selection for just oneof the problems.

SCAN FOR CRIME PROBLEMS

SC

AN

Identified Vandalism Thefts Burglariessolutions to lifts of/from of flats(from least carscostly to most)

Trim bushes to improve ** **surveillance (£)

Block watch scheme (£) *** * **

Alarms for lifts (££) ****

Electronic access to ****car park (££)

Installation of entry ** ****phone (£££)

Window locks and strengthened ****doors for flats (££££)

Security patrol (££££) * ** ***

Concierge system and ***** ***** *****estate-wide CCTV cameras (£££££)

£ Predicted costs * Predicted effectiveness

Local police have to deal with a wide range ofproblems that meet the CHEERS definition(Step 15) and we have developed a scheme toclassify these problems. Classification is impor-tant because it allows comparison of a newproblem to similar problems that have alreadybeen addressed, and it helps identify importantfeatures for examination. The scheme is basedon two criteria: the environments within whichproblems arise and the behaviours in whichthe participants engage. (The scheme is differ-ent from the wolf/duck/den classification that isa classification of persistent problems, Step 9.)

Environments regulate the targets available,the activities people can engage in and whocontrols the location. Specifying an environ-ment allows comparisons of environmentswith and without the problem. It also helpsidentify potential stakeholders and partnersfor addressing the problem. There are elevendistinct environments for most commonpolice problems:

l Residential – Locations where people dwell.Houses, flats, and hotel rooms are exam-ples. Though most are in fixed locations, afew are mobile, such as caravans.

l Recreational – Places where people go tohave a good time. Pubs, nightclubs, restau-rants, cinemas, playgrounds, and parks areexamples.

l Offices – Locations of white-collar workwhere there is little face-to-face interactionbetween the workers and the generalpublic. Government and business facilitiesare often of this type. Access to these loca-tions is often restricted.

l Retail – Places for walk-in or drive-upcustomer traffic involving monetary trans-actions. Stores, branch banks, and postoffice branches are examples.

l Industrial – Locations for processing ofgoods. Cash transactions are not importantactivities in these environments and the publicis seldom invited. Factories, warehouses, pack-age-sorting facilities are examples.

l Agricultural – Locations for growing cropsand animals.

l Education – Places of learning or study,including day care centres, schools, univer-sities, libraries and churches.

l Human service – Places where people gowhen something is wrong. Courts, jails,prisons, police stations, hospitals and somedrug treatment centres are examples.

l Public ways – Routes connecting all otherenvironments. Roads and highways, foot-paths and bike trails, and drives andparking facilities are examples.

l Transport – Locations for the mass move-ment of people. These include buses, busstations and bus stops, airplanes and air-ports, trains and train stations, ferries andferry terminals, and ocean liners and piers.

l Open/transitional – Areas without consis-tent or regular designated uses. Thesediffer from parks in that they have not beendesignated for recreation, though peoplemay use them for this. Transitional areasinclude abandoned properties and con-struction sites.

Behaviour is the second crucial dimensionfor classifying a problem. Specifying behav-iours helps pinpoint important aspects ofharm, intent, and offender–target relation-ships. There are six types of behaviour:

l Predatory – The offender is clearly distinctfrom the victim and the victim objects tothe offender’s actions. Most commoncrimes are of this type. Examples includerobbery, child abuse, burglary, and theft.

l Consensual – The parties involved know-ingly and willingly interact. This typicallyinvolves some form of transaction. Examplesinclude drug sales, prostitution and stolengoods sales. Note, however, that assaults onprostitutes are predatory behaviours.

l Conflicts – Violent interactions involvingroughly coequal people who have somepre-existing relationship. Domestic vio-lence among adults usually involves thistype of behaviour, though domestic vio-lence against children and the elderly isclassified as predatory because the partiesinvolved are not roughly coequal.

16 Know what kind of problem you have

l Incivilities – Offenders are distinguishablefrom victims, as in predatory events, but thevictims are spread over a number of individu-als and the harms are not serious. Manyconcerns that are annoying, unsightly, noisy ordisturbing, but do not involve serious prop-erty damage or injury fall into this category.Some incivilities are troublesome regardless ofthe environment, while others are only trou-blesome in specific environments.

l Endangerment – The offender and thevictim are the same person or the offenderhad no intent to harm the victim. Suicideattempts, drug overdoses, and motor vehi-cle accidents are examples.

l Misuse of police – A category reserved forunwarranted demands on the police service.False reporting of crimes and repeated call-ing about issues citizens can handlethemselves are examples. This is a categoryof last resort — for use when the sole harmstemming from the behaviour is the expendi-ture of police resources and when none ofthe other types fit.

The table shows the full classification. A problemis classified by putting it in the cell where theappropriate column intersects with the appro-priate row. So, for example, the 2001 TilleyAward winner dealt with glass bottle injuriesaround pubs, a conflict-recreational problem(A). The 2002 Tilley Award winner dealt withmotorcycle accidents along a scenic road, anendangerment-public ways problem (B).

Though most problems fit into a single cell, onoccasion a problem might involve multiplebehaviours or environments. For example, theStaffordshire Police had a problem createdwhen protesters occupied abandoned build-ings along a construction right of way. Thesewere open/transitional environments. Theprotests involved incivilities, but the tactics foroccupying these buildings also posed a dangerto the protesters. Thus, endangerment wasanother relevant behaviour (C in the table).Though multiple types of behaviours or envi-ronments are sometimes needed, excessiveuse of multiple types can lead to imprecision.

SCAN FOR CRIME PROBLEMS

SC

AN

John Eck and Ronald Clarke (2003). Classifying Common Police Problems: A Routine Activity Approach. In CrimePrevention Studies, vol. 16, edited by Martha Smith and Derek Cornish. Monsey, New York: Criminal Justice Press(and Willan Publishing, UK).

Read more

A classification scheme for common problems facing local policeEnvironments Behaviours

Predatory Consensual Conflicts Incivilities Endangerment Misuse of police

Residential

Recreational A

Offices

Retail

Industrial

Agricultural

Educational

Human service

Public ways B

Transport

Open/transitional C C

The crime triangle (Step 9) identifies the threeessential elements of crime, but does notexplain how the offender finds a suitablevictim and place. This task is left to crime pat-tern theory, which was developed by theenvironmental criminologists Pat and PaulBrantingham, working from a background insocial geography.

The Brantinghams describe offenders’ searchpatterns in terms of their personal activityspaces. Starting with a triangle, they consideroffenders going from home to work to recre-ation. Around each of these three nodes andalong each of these three paths (excepting abuffer zone where they might be recognised)offenders look around for crime opportunities.They may find these a little way off the path,but they usually do not go far beyond the areathey know. This is because it is easier tocommit crimes in the course of their daily rou-tine than by making a special journey to do so.

The Brantinghams also use another importantconcept: edges, which refers to the bound-aries of areas where people live, work, shop orseek entertainment. Some crimes are morelikely to occur at these edges – such as racialattacks, robberies, or shoplifting – becausethis is where people from different neighbour-hoods who do not know each other cometogether. In an early study, the Brantinghamsfound that residential burglaries in Tallahasseetended to cluster where affluent areas bor-dered on poor areas. Their explanation wasthat the affluent areas provided attractive tar-gets to burglars from the poorer areas.However, the burglars preferred not to ven-ture too far into the affluent areas. They wereunfamiliar with them and thought they mightbe recognised as not belonging there.

The paths that people take in their everydayactivities and the nodes they inhabit explain risksof victimisation as much as patterns of offend-ing. This is why the Brantinghams and othercrime pattern theorists pay so much attention tothe geographical distribution of crime and thedaily rhythm of activity. For example, they gener-ate crime maps for different hours of the day anddays of the week, linking specific kinds of crimesto commuter flows, school children being let

out, pubs closing, or any other process thatmoves people among nodes and along paths.Pickpockets and bag snatchers seek crowds,while other offenders pay closer attention to theabsence of people. For example, the flow ofpeople to work generates a counter flow of bur-glars to residential areas, taking advantage oftheir absence. The flow of workers home atnight and at weekends produces a counter flowa few hours later of commercial and industrialburglars to take advantage of the situation.

You can use the concepts of crime patterntheory to understand crime in your force area.You should try to piece together offender andoffence patterns by finding nodes, paths andedges. You can begin to distinguish betweenhow offenders search for crime and when theyfind it by accident. You can find where offendersare absent and where they congregate in ‘hotspots’ and think about the reasons for this (Step18). Unfortunately, ‘spots’ have no clear andfinal size, but are defined by the parameters youenter into your analysis and you may sometimesdo better to focus on parks, schools, housingestates, street segments and other identifiableplaces. You will find that very local crime pat-terns tell the story. Thus a high crime districtwill have some streets with no crime at all andsome addresses which generate most of theproblem. Residents will know it is fairly safe towalk down one street but very unsafe to walkdown another. They will even choose one sideof the street over the other. If residents knowtheir local turf this well, what’s to stop you fromfinding out about it? Crime pattern theory helpsyou do just that, and it will help to define a spe-cific problem at the scanning stage andunderstand the contributory causes at analysis.

17 Study the journey to crime

Brantinghams (1993) Environment, Routine, andSituation: Toward a Pattern Theory of Crime. In RoutineActivity and Rational Choice, Advances in CriminologicalTheory, Vol. 5, edited by Ronald Clarke and MarcusFelson. New Brunswick, NJ: Transaction Publishers. Marcus Felson (2002) Crime and Everyday Life.Thousand Oaks, CA: Sage.Paul Wiles and Andrew Costello (2000) The ‘Road toNowhere’: The Evidence for Travelling Criminals. HomeOffice Research Study 207. London: Home Office

Read more

Kim Rossmo prepared this diagram to representthe Brantinghams’ theory. It shows an offender’sactivity space (residence, work, recreation, and thetravel routes between them), the buffer zone closeto the home in which offenders do not usuallycommit crimes and five potential target areas (forexample, car parks). Where an offender's activityspace intersects a target area, this is where crimeshappen (blue crosses). Note that in this exampleno crimes occur around the offender's work place,because there are no suitable targets there. Also,there are two target areas with no crimes in thembecause this offender is not aware of those places.

SCAN FOR CRIME PROBLEMS

SC

AN

Work

Target areas

Recreation

Activityspace

Crime sitesResidence

Bufferzone

Brantingham crime pattern theory

Source: Kim Rossmo (2000). Geographic Profiling. Boca Raton,FL: CRC Press.

The journey to crime and the ‘self-containment index’Andy Brumwell, crime analyst with the WestMidlands Police, has recently completed an analysisof the distance travelled to crime from home usingforce data for AY2000/01–2002/03. He included258,074 crime trips in his analysis. He found:

l Just over 50% of all journeys were less than onemile.

l Distance travelled varies with the offence. Forexample, 50% of arsonists travelled less thanquarter of a mile, whereas only 13% of shoplifterscommitted their offences this close to home.

l Females travel further than males, possiblybecause many committed shopliftings.

l There is considerable variation among individualoffenders in crime trips. Some usually commitcrime in their local neighbourhoods. Others travelfurther particularly when working with co-

offenders. In some cases, co-offenders may travelconsiderable distances from where they now live,to an area where they all grew up together.

l Younger criminals do not travel as far as oldercriminals to commit crime, as shown here in hisgraph.

This study has led him to develop what he calls the‘self-containment index’ that looks at thepercentage of crimes in an area that is committedby offenders who also live in that area. A value of100 indicates that local offenders are responsiblefor all the crimes, whereas a value of zero indicatesthat local offenders commit none of them. Thisvalue should be calculated when analysing a localproblem. Whether predators are local or come froma distance will have an influence on the type ofsituational crime prevention measures that could besuccessfully introduced. For example, alley-gating(Step 33) will be less successful if many of theoffenders live within the gating scheme itself.

2.50

2.00

1.50

1.00

0.50

0.005 10 15 20 25 30 35 40 45 50 55 60 65 70

Average miles travelled to commit crime by age

Hot spots are geographic concentrations ofcrime. The Brantinghams have distinguishedbetween three kinds of hot spots and underly-ing causal mechanisms:

1. Crime generators are places to whichlarge numbers of people are attracted forreasons unrelated to criminal motivation.Providing large numbers of opportunitiesfor offenders and targets to come togetherin time and place produces crime or disor-der. Examples of generators includeshopping areas, transportation hubs, festi-vals, and sporting events. The large numberof crime or disorder events is due princi-pally to the large number of place users. Sothese types of problems grow as the use ofthe area grows.

2. Crime attractors are places affording manycriminal opportunities that are well knownto offenders. People with criminal motiva-tion are drawn to such locales. In the shortrun, offenders may come from outside thearea, but over longer time periods, andunder some circumstances, offenders mayrelocate to these areas. Prostitution anddrug areas are examples. Some entertain-ment spots are also well known for allowingdeviant activity. Such places might start offbeing known only to locals, but as their rep-utation spreads increasing numbers ofoffenders are drawn in, thus increasing thenumber of crime and disorder events.

3. Crime enablers occur when there is littleregulation of behaviour at places: rules ofconduct are absent or are not enforced.The removal of a car park attendant, forexample, allows people to loiter in theparking area. This results in an increase inthefts from vehicles. This is an example ofan abrupt change in place management.Sometimes place management erodesslowly over time, leading to problemgrowth. Crime enablers also occur with theerosion of guardianship and handling. Forexample, if parents attend a play area withtheir children they simultaneously protectthe children (guardianship) and keep their

children from misbehaving (handling). Ifparenting styles slowly change so that thechildren are increasingly left to themselves,they can become at increased risk of victim-isation and of becoming offenders.

The Brantinghams also suggest that areas canbe crime neutral. Crime-neutral areas attractneither offenders nor targets, and controls onbehaviours are adequate. These areas tend tohave relatively few crimes, and the crimes tendto be relatively unpatterned. For this reason,crime-neutral areas seldom draw police atten-tion. Though they seldom require crimeanalysis, they are important because they pro-vide a useful comparison to the other types ofareas. Comparing crime-neutral areas, forexample, to a hot spot can help identify the dif-ferences that create the troubles in the crimegenerator, crime attractor, or crime enabler.

In summary, the development and growth ofcrime and disorder hot spots involves threedifferent mechanisms: increasing targets,increasing offenders, and decreasing controls.To varying extents all three are at work in mostproblems. Shoppers might increase in an area,for example, due to new roads or shoppingopportunities. This might increase thefts asoffenders, also shopping in the area, takeadvantage of the new theft opportunities. Newoffenders might be attracted to the areabecause of the success of the early offenders.Increasing offending causes the number ofshoppers to decline. This removes guardian-ship (shoppers). But it has another effect. Itreduces the resources of the shopkeepers tomanage their stores and the surrounding area,a reduction in place management. So, a prob-lem that started out as a crime generatorevolved into a crime attractor and then into acrime enabler.

Crime generators have many crimes, but astheir number of targets is high, each may havelow crime rates. Crime attractors also havemany crimes, but as they have relatively few tar-gets, their crime rates may be high. Crimeenablers, with their weakened behaviour

18 Know how hot spots develop

controls, tend to be unattractive to targets.However, those few available targets have highrisks. So an area with relatively few crimes but ahigh crime rate suggests a crime enabler. Finally,the number of crimes at crime-neutral locationswill be low, so even if the number of targets isnot particularly great, their crime rate will alsobe low. Table 1 summarises these relationships.

To diagnose which process is operating, firstrank the areas using numbers of crimes, thenlook at their rates. Table 2 shows a hypotheti-cal example. Area A has a relatively highnumber of crimes (column 2) and a high rate(column 4) so it may be a crime attractor. Area

D has relatively few crimes and a low rate, so itmay be crime neutral. Area B has a largenumber of crimes, but a relatively low rate, soit appears to be a crime generator. Andbecause Area C has a relatively low number ofcrimes, but a high crime rate, we would expectit to be a crime enabler.

What are the practical consequences of know-ing how your hot spot arose? The answer is summarised in Table 3. Knowledge of theunderlying causes (column 2) suggests possi-ble responses (column 3).

SCAN FOR CRIME PROBLEMS

SC

AN

Table 3: What to do about worsening hot spots Hot spot type Cause Type of response Questions to answer

Crime generator Many unprotected Increase protection In what circumstances targets are targets vulnerable?

How can vulnerability be changed?

Crime attractor Attracts offenders Discourage offenders What is attracting offenders?from coming How can this be changed?

Crime enabler Erosion of controls Restore guardianship, Who could control behaviour?handling or place How can they be encouragedmanagement to exert controls?

Table 1: Diagnosing hot spotsNumber Rateof crimes of crime

Crime attractors High High

Crime generators High Low

Crime enablers Low (High) High

Crime neutral Low Low

Table 2: Using numbers and rates to diagnose a problemArea Crimes Targets Rate per Type

100 targets

A 391 898 44 Attractor

B 148 1,795 8 Generator

C 84 243 35 Enabler

D 28 638 4 Neutral

Patricia and Paul Brantingham (1995). Criminality of Place: Crime Generators and Crime Attractors. EuropeanJournal on Criminal Policy and Research, 3(3): 1–26.

Read more

One of the most important principles of crimeanalysis is that a few people and places areinvolved in most of the criminal events. This isthe heart of the Wolf/Duck/Den phenomenondescribed in Step 9. This rule can also befound at work with hot products – a few prod-uct types are disproportionately the targets ofthieves (see Step 29). This type of concentra-tion is not peculiar to crime and disorder; it isa practically universal law. A small portion ofthe earth’s surface holds the majority of life onearth. Only a small proportion of earthquakescause most of the earthquake damage. A smallportion of the population holds most of thewealth. A small proportion of police officialsproduce most of the arrests.

This phenomenon is commonly referred to as the 80–20 rule: 20% of some things areresponsible for 80% of the outcomes. In practice, it is seldom exactly 80–20, but it isalways a small percentage of something orsome group involved in a large percentage ofsome result.

One of the most important questions in theinvestigation of any problem is to ask if the80–20 rule applies. A simple 7-stage procedureshows how to answer this:

1. Identify the people or places the rule mightapply to. The problem of assaults outsidepubs, for example, suggests that a few pubsmight be responsible for most of the trouble.

2. Get a list of these people or places with acount of the number of events associatedwith each person or place.

3. Rank order the people or places accordingto the number of events associated witheach – most to least (the table is a hypo-thetical list of pubs along with the numberof reported assaults associated with each).You should ask whether there is somethingdifferent about the people or places at thetop of the list, compared to those in themiddle or at the bottom. Perhaps the pubsat the bottom of the list are popularevening entertainment spots for young

people, or are all located in the city centre,or are owned by the same company. If so,then these differences might be related tothe source of the problem. If there are clearand obvious differences, then divide thislist into meaningful categories, with sepa-rate ranked lists for each. Potentially, eachcategory may be a distinct problem. Foreach separate category, continue with Stage4. We will assume that in our example thereare no important differences.

4. Calculate the percentages of the eventseach person or place contributes. Thereare 121 assaults. The first pub, the WhiteHart, contributed 31 of these. So it has25.6% of the problem. The third columnshows the percentage.

5. Cumulate the percentages starting with themost involved person or place. This showsthe proportion of the events that is associ-ated with each percentile (e.g. worst 10%,worst 20%, and so on to 100%). The fourthcolumn shows what is called the cumulativepercentage (percentages from the thirdcolumn are added starting with the WhiteHart and going up). The shading separateseach 10 percentile. With only 30 pubs, 10%is about as small a percentile as is practical.But with a much longer list, it may makesense to look at smaller gradations, such as5 or even 1%.

6. Calculate the proportion of the people orplaces each single person or place repre-sents. In our example, there are 30 pubs soeach represents 3.3% of the pubs. Thencumulate these percentages in the samedirection as you followed in Stage 5 (topdown in column 5).

7. Compare the cumulative percentage ofpeople or places (column 5) to the cumula-tive percentage of outcomes (column 4).This shows how much the most involvedpeople or places contribute to the problem.

This type of analysis can be used during scan-ning to detect offenders most in need ofattention, places most in need of intervention,

19 Learn if the 80–20 rule applies

and victims most in need of assistance. It canalso be used in the analysis stage to determineif there are important differences between

people and places at the top and those at thebottom of the list.

SCAN FOR CRIME PROBLEMS

SC

AN

The concentration of 121 assaults in 30 pubs 1 2 3 4 5

No. % Cum. Cum.assaults assaults % assaults % pubs

White Hart 31 25.6 25.6 3.3

Union 17 14.0 39.7 6.7

Feathers 13 10.7 50.4 10.0

Wellington 11 9.1 59.5 13.3

Black Prince 8 6.6 66.1 16.7

Angel 7 5.8 71.9 20.0

George & Dragon 6 5.0 76.9 23.3

Cross Keys 6 5.0 81.8 26.7

Saracen’s Head 4 3.3 85.1 30.0

White Bear 4 3.3 88.4 33.3

Mason’s Arms 3 2.5 90.9 36.7

Cock 3 2.5 93.4 40.0

Badger 3 2.5 95.9 43.3

Hare & Hounds 1 0.8 96.7 46.7

Red Lion 1 0.8 97.5 50.0

Royal Oak 1 0.8 98.3 53.3

George 1 0.8 99.2 56.7

Cross Hands 1 0.8 100 60.0

Rose & Crown 0 0 100 63.3

King’s Arms 0 0 100 66.7

Star 0 0 100 70.0

Mitre 0 0 100 73.3

Dog & Fox 0 0 100 76.7

Griffin 0 0 100 80.0

Plough 0 0 100 83.3

Queen’s Head 0 0 100 86.7

White Horse 0 0 100 90.0

Bull 0 0 100 93.3

Swan 0 0 100 96.7

Black Bear 0 0 100 100

A hypothesis is an answer to a question abouta problem, and can be true or false.Hypotheses come from experience and theory.The table provides three examples of hypothe-ses, the questions they answer, and possibleways of testing their validity.

A set of hypotheses is a roadmap for investi-gating a problem. Hypotheses suggest types ofdata to collect, how this data should beanalysed, and how to interpret analysis results.Consider a problem involving neighbourhooddisturbances, for example. Based on the ideaspresented in this manual, you should hypothe-size that: some times of the year have moredisturbances than others; there are a few loca-tions with many disturbances and manylocations with few or no disturbances; thereare a few people who are routinely involved increating disturbances, but most people whoare involved are only involved on rare occa-sions. By determining the truth or fallacy ofstatements like these, you describe the prob-lem and reveal solution options.

Hypotheses suggest the type of data to collect.If you want to test the hypothesis that thereare a few people who are often involved in cre-ating disturbances, but a larger number of

people who are only involved on occasion,you must find data that describes the numberof disturbances offenders are involved in. Ifyou wanted to test the hypothesis about timeyou would have to find data that gave the dateand time of incidents.

Paralysis by analysis

The lack of explicit hypotheses can lead to‘paralysis by analysis’, collecting too muchdata, conducting too much analysis, and notcoming to any useful conclusion.

Hypotheses direct the analysis of data. Step 18describes several ways problems grow:increasing targets, increasing offenders, ordecreasing controls on behaviour. Step 25shows how to compare numbers and rates todetermine which of these processes is at work.You might formulate a hypothesis such as,‘This problem is due to an increasing numberof targets becoming available’. When youexamine the number and rates of events youcan determine if this hypothesis is reasonable.Another example can be found in Step 19,which describes the 80–20 rule. A hypothesisclaiming that the 80–20 rule is at work can be

20 Formulate hypotheses

Questions, hypotheses and testsQuestion Example hypothesis Possible test

Why does this This hot spot is due to a Count the number of targets in the hot spot and hot spot occur? large number of targets calculate the crime rate. Compare this to rates

being available. for the surrounding area. If the hot spot rate is higher, the hypothesis is false, but if it is aboutthe same or lower then the hypothesis is true.

Why are there more Problem area residents are If the problem area has similar or lower on-street car thefts in the more likely to park their parking rates as the others, reject the hypothesis.problem area than cars on the street than If higher, accept it.in nearby areas? residents of the other areas.

Why did the theft It increased when a scrap Compare the thefts of copper piping for periods ofof copper piping metal dealership was sold time before and after the change in owners. If thefrom new to a new owner. theft rate is the same before and after, or theconstruction sites trend in thefts was already going up before thesuddenly increase? change, then the hypothesis is probably false.

If otherwise, the hypothesis appears reasonable.

tested using the procedures described in thisstep. Not all hypotheses are closely associatedwith specific analytical procedures. Thesecond two examples in the table describetests that do not directly come from any spe-cific theory, but instead are logically related tothe question and hypothesis.

Hypotheses help interpret the analysis results.Test results can suggest useful solutions toproblems. If you are examining vehicle thefts,you might ask, ‘Why are there more carsstolen from this car park than nearby carparks?’ From your knowledge of the problemcar park, you might hypothesise that the facil-ity has many users who leave their vehicles forlong periods and that because users pay whenthey enter, no one watches who leaves withthe car. Various theories suggest that whenthere is no one watching, crime is more likely.So, if the hypothesis is correct, then aresponse involving someone watching parkedor exiting cars might be effective. Comparingthis parking facility to ones nearby can helptest the hypothesis. If parking facilities withmore surveillance have fewer thefts than thosewith less surveillance, then the hypothesis issupported. But if you find nearby parking facil-ities with the same level of surveillance, butmuch lower theft rates, you will have to con-sider alternative responses.

To formulate hypotheses you need to askimportant questions. Then create a simple and

direct speculative answer to the questionbased on experience and theory. This answeris your hypothesis. The statement must bebold enough that it could be wrong, and theremust be a way of showing whether it is right orwrong. If possible, create two or more com-peting hypotheses. If each hypothesis is linkedto a potential solution, the test of thesehypotheses simultaneously directs your atten-tion to feasible responses and rules outineffective approaches.

If you cannot test a hypothesis, you cannotanswer the question, and any response basedon the hypothesis is of little use. If offenderdata are unavailable, for example, then creatinga response directed toward reducing repeatoffending has a high likelihood of failurebecause you do not know if repeat offending ispart of this problem. Consequently, it is some-times useful to list hypotheses that you cannottest as these may be linked to responses thatcannot be supported by your analysis. Avoidusing these responses!

Finally, make sure that your test results make adifference. That is, if the hypothesis is true youwill take a different decision than if it is false. Ifyou will take the same decision regardless ofthe test results, then the hypothesis and itstest are irrelevant.

ANALYSE IN DEPTH

AN

ALY

SE

It is helpful to distinguish between acute andchronic hot spots. Acute hot spots showabnormal spikes in crime, which may declineautomatically, while chronic hot spots havepersistently higher crime levels than otherareas. There are three basic forms of chronichot spots, each of them linked to particulartheories and types of responses.

l Hot dots are locations with high crimelevels. These show crime concentrated atfacilities or at addresses of repeat victims(see Steps 26 and 27). Multiple crimeevents at places are represented by dots.

l Hot lines are street segments where crimeis concentrated. These might occur, forexample, if vehicles parked along particularstreets suffer high rates of break-ins.Multiple crimes along street segments areshown with lines.

l Hot areas are neighbourhoods wherecrime is concentrated. Hot areas arise for avariety or reasons. Area characteristics maygive rise to crime. Or an area may be hotbecause it contains many separate and dis-crete problems. On maps, hot areas aretypically shown as shaded areas, contourlines, or gradients depicting crime levels.

The figure depicts these three forms of hotspots. Troublesome entertainment locationsare shown as dots because the assaults arelocated at addresses. Vehicle break-ins, how-ever, are along continuous street segments, sothis concentration is shown as two intersect-ing lines. Finally, the graduated contours forthe residential hot spot suggests that risk forbreak-ins is highest in one small area butdeclines as one goes away from the centre.The dots within this graduated area depictrepeat burglary locations.

Clarifying the nature of your hot spot gives aninkling of response:

l Hot dots suggest changing the physicalenvironment of particular places or chang-ing their management. They also suggestintervening with high-risk victims.

l Hot lines suggest changing streets, pathsand other routes, or the environmentsalong them.

l Hot areas suggest large-scale partnershipsto change neighbourhoods. The tableshows how crime concentration is relatedto the way it is mapped, and where theresponse is focused.

21 Diagnose your hot spot

Hot lines representingthefts from vehicles along

a side street

A hot area representedas a gradient of risk

for residential burglary

One of five hot dotsshowing repeat

burglary locations

One of four hot dotsrepresenting

entertainmentvenues with a highnumber of assaults

Types of hotspots

Analysis of hot spots should begin with places,then move to streets, and finally to areas.Consider, for example, the problem of burned-out cars. Are they repeatedly found at specificaddresses? If ‘yes’ then you should ask whythese places are chosen instead of other nearbysites. If ‘no’ you should move on to examinestreets. If you find street-level concentration,you should compare streets to find out whysome attract burned-out cars and others do not.If there is little street-level concentration, thenyou should consider community concentrationand make the relevant comparisons. Thisapproach assures a highly focused response.

In the figure, the hot burglary dots indicaterepeat victims within the overall neighbour-hood problem. An area hot spot alone wouldnot reveal this. Before proceeding further, youshould determine if the area hot spot is largelydue to the few repeat burglary spots. You cando this by treating each crime location as if ithad only a single event, and then looking at thearea hot spot. If it is, your problem-solvingefforts should focus on these locations, not theneighbourhood. On the other hand, if theserepeat burglary dots are fragments of a largerconcentration of burglaries, the scope of youranalysis must expand beyond these places.

Hot spot analysis provides early indicators ofwhere to focus attention, but further analysisis always required. It is only useful if there maybe a geographic component to the problem,or the concentration of crime is related to geo-graphical features like roads or land use. It canbe a valuable tool early in the problem-solvingprocess, but it only produces suspicions thatmust be investigated further. For some prob-lems, hot spot mapping has little utility andyou must use other analytical approaches.Over-reliance on hot spots can result in super-ficial analysis and the implementation ofineffective responses.

ANALYSE IN DEPTH

AN

ALY

SE

Two articles in Crime Prevention Studies, vol. 13,edited by Nick Tilley (2002). Monsey, New York:Criminal Justice Press (and Willan Publishing, UK):

1. Elizabeth Groff and Nancy LaVigne. Forecasting theFuture of Predictive Crime Mapping.

2. Michael Townsley and Ken Pease. Hot Spots andCold Comfort: The Importance of Having a WorkingThermometer.

Read more

Concentration, mapping and actionConcentration Hot spots shown as: Action level Action examples

Places – at specific Dots Facility, corner, address CCTV in a parking deck, addresses, corners, changing the way alcoholor facilities is served in pubs.

Victims Dots Victims’ addresses Repeat victimisation programmes.

Streets – along streets Lines Along paths, streets, Creating cul-de-sacs, or block faces and highways changing traffic patterns,

altering parking regulation.

Area – neighbourhoods Shaded areas Neighbourhoods, Community partnerships, regions and other neighbourhoodareas redevelopment.

Conventional software is of little use whenmapping crime in a city centre, a universitycampus, a council estate or any site with manylarge buildings. This is because most buildings,however large, have only one street addressand crimes occurring anywhere in the buildingare assigned to that address. Mapping mighttherefore suggest that a particular building orfacility has a crime problem, but this may onlybe because it is so large. When account istaken of the many people working in thebuilding or using the facility, it could prove tobe relatively safe. For example, GeorgeRengert showed that a multi-storey car park incentral Philadelphia identified as a car crimehot spot actually had a lower rate of car crimethan the surrounding streets, once accountwas taken of the large number of cars thatcould be parked in the facility.

In fact, many large buildings are not safe. Inhis devastating critique of 1960s Americanpublic housing, Oscar Newman showed thatthe taller a tower block, the higher the rate ofcrime per 100 residents. He argued that verylarge blocks invited crime because residentsdid not know their neighbours and the designand layout of the buildings made it difficult forthem to exercise any supervision of the publicspaces, including corridors, lifts and playareas. His ideas have since been developedinto a set of principles – Crime PreventionThrough Environmental Design (or CPTED) –for designing and laying-out secure buildingsand public spaces.

To understand why a particular building isinsecure, crimes need to be divided into spe-cific categories and their locations within thebuilding need to be charted. This is wherehigh-definition or ‘3-D’ mapping comes intoplay. Unfortunately, high-definition mapping isdifficult and time consuming. It suffers fromtwo principal problems:

1. Police records of crime rarely give the pre-cise location of incidents within thebuilding, though building managers orsecurity departments can sometimes

supply this information. When they cannot,special crime recording procedures mayhave to be established for a period of timein order to obtain this information.

2. For new buildings, it may be possible toobtain plans in digitised format, which canmake mapping easier. But when the build-ing is old, it may be difficult to obtainup-to-date plans and you may have to getthese drawn.

In many cases, these problems will simply ruleout high-definition mapping, but they can beovercome as George Rengert and his col-leagues showed in their study of crime onTemple University’s campus in Philadelphia.They developed a high-definition GIS by com-bining mapping software with AutoCADdrawings of the campus. Features, such aswater pipes and electrical wiring, were elimi-nated and the maps were altered so that streetswere represented as lines (with lines on eitherside representing pavements), while polygonswere used to represent the footprints of build-ings and the shapes of athletic fields andparking lots. Shrubbery, fences, lighting andother physical features were also representedon the maps. Crimes recorded by the campuspolice were then plotted exactly where theyoccurred, allowing them to be related to envi-ronmental features such as poor lighting or ablind corner allowing the attacker to lie in wait.

Crime was mapped for the floors of eachbuilding and a picture of the horizontalarrangement of crime within the building wasprojected onto its ground floor footprint.Figure 1 shows the result for one of the build-ings – Gladfelter Hall. With the exception ofthe first floor, it shows that crime was mostlikely to occur on the upper floors. It alsoshows some clear clusters of crime, the largestof which was in the area closest to the bank offour elevators near the centre of the building.This is where each department’s ‘fishbowl’offices for secretaries and receptionists arelocated – fishbowls because they are sur-rounded by glass windows, which allow

22 Know when to use high-definitionmaps

thieves to look into them to see if anyone isthere and if anything is worth taking. TheDepartment of Criminal Justice – Rengert’sown department – has now installed blinds,which can be lowered in the evenings to pre-vent people seeing into the fishbowl.

Commercial software is already available thatwill produce photo-realistic city models andtechnological developments, such as 3-D laserimaging, will simplify the production of com-puter maps like those of Gladfelter Hall.Meanwhile, when the number of incidents issmall, good clear drawings with the location ofcrimes clearly indicated can sometimes do justas well. Figure 2 is a plan drawn by BarryPoyner of the Lisson Green council estate inLondon showing the locations of robberiesand snatches on the walkways connecting thebuildings for two six-month periods: beforeany preventive changes were made and afterfour of the blocks were fitted with entryphones. In effect, the entry phones closedaccess to the walkway system from the main

street entrance. In this case, high-definitionmapping assisted with the assessment of pre-ventive action, but Figure 2 also helped withdiagnosis of the problem because it showedrobberies and snatches tended to occur onthose parts of the walkway system that lackedsurveillance from neighbouring buildings orground level.

ANALYSE IN DEPTH

AN

ALY

SE

Tim Crowe (1991). Crime Prevention Through Environmental Design. Applications of Architectural Design and SpaceManagement Concepts. Boston: Butterworth-Heinemann.

George Rengert, Mark Mattson and Kristin Henderson (2001). Campus Security. Situational Crime Prevention inHigh-Density Environments. Monsey, New York: Criminal Justice Press.

Read more

High definitiongeographicinformationsystem

Crime projected onbuilding footprint

Crime on building floors

10

9

8

7

6

5

4

3

2

1

Figure 1: 3-D map of GladfelterHall, Temple University Campus

SIX MONTHS BEFORE CHANGES

Housing blocks

Blocks fitted withentry phones

Walkways

Robbery or snatch

Elevator andstair towers

KEY

SIX MONTHS AFTER FOUR BLOCKS WERE FITTED WITH ENTRY PHONES

Figure 2: Location of robberies and snatches on the walkwaysystem of the Lisson Green Estate

Source: Barry Poyner (1997) An Evaluation of Walkway Demolition ona British Housing Estate. In Situatonal Crime Prevention. SuccessfulCase Studies (2nd edn.), edited by Ronald V. Clarke. Monsey, NY:Criminal Justice Press.

Cycles of activities have tremendous influenceson problems. The ebb and flow of vehiclescaused by commuting and shopping rhythms,for example, changes the number of targets andguardians in parking facilities. This in turn influ-ences when vehicle thefts and break-ins aremost frequent. Robberies of intoxicated revellersmay be more likely around pub closing time onFridays and Saturdays, because the number oftargets is higher. In this example, two importantrhythms concentrate problem activities. Thefirst is the workday/weekend cycle that makesFriday and Saturday nights so popular for enter-tainment and recreation. The second involvesthe daily cycle of opening and closing of drink-ing establishments. In this step we will discussshort-term fluctuations occurring over hoursand days. In Step 24, we will look at longer timeperiods covering months and years.

Different facilities have different cycles ofactivities that can contribute to its associatedproblems. School rhythms are similar, but dis-tinct from job rhythms. Bus stops areinfluenced by the rhythm of commuting andshopping, but also by the more frequentcoming and going of buses.

Charting the rhythm of crime or disorderevents helps identify important activity cyclesthat may contribute to a problem. As shown inFigure 1, calculate the average of the numberof events occurring in each hour (or other

time interval) over several days. Then plot theresults. The more days you can average over,the clearer the pattern. In general, weekendsand holidays should be analysed separatelyfrom weekdays.

Temporal analysis is easiest when problemevents are frequent. So temporal analysis will bemost useful for common minor events, like noisecomplaints and minor traffic accidents, than foruncommon serious events, like murder. If thereare few events, then you can look at a longerperiod to collect more events. But if the problemchanges in the longer period, the picture thatemerges may be distorted or out of date.

Having reasonably exact times of occurrencehelps temporal analysis. Contact crimes, suchas robbery, rape and assault, can be easily pin-pointed as victims can often describe whenthese crimes took place. Property crimes, suchas vehicle crimes, burglary and vandalism, aremuch harder to pin down because victims canusually provide only a range during whichsuch crimes could have occurred. You can usethe midpoints in these ranges to estimate thetimes discovery crimes are most likely to haveoccurred. This can lead to some distortion.(Jerry Ratcliffe proposes the use of ‘aoristicanalysis’. The limitation of this approach isthat it is more complex than using the mid-point, and the software to make suchcomputations is not commercially available.)

23 Pay attention to daily and weeklyrhythms

Hours1

20

15

10

5

0Aver

age

num

ber

of e

vent

s

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

AVERAGE DAILY RHYTHM OF EVENTS

Days12345

Average

17

10483

6.4

239

118

128.6

36

1012

76

8.2

49

11697

8.4

597849

7.4

611

63

10128.4

71613181713

15.4

81723172119

19.4

91615242020

19.0

101718141119

15.8

1156

1010

36.8

126

12754

6.8

1312

744

107.4

1477

1134

6.4

1598

1199

9.2

1655963

5.6

172016231715

18.2

181818242416

20.0

191617241817

18.4

208

12748

7.8

2173

115

107.2

221011

337

6.8

238

103

106

7.4

2474466

5.4

Number of events each hour for five days

Hours

Figure 1: Examining a 24-hour rhythm of crime

Jerry Ratcliffe has identified three forms oftemporal clustering (Figure 2). Events may berelatively evenly spread over the entire day. Hecalls this a diffused pattern. Focused pat-terns show clustering within distinct timeranges. Events clustered around rush hoursfollow focused patterns. Acute patterns aretightly packed within small periods.Disturbances immediately following pub clos-ing time might be an example. Focused andacute patterns immediately suggest temporalcycles that should be investigated.

Though Ratcliffe developed his typology fordaily patterns, the basic idea can be applied toweekly cycles. If no particular day of the weekis routinely troublesome, this indicates a dif-fused weekly pattern. A cluster of daysshowing a marked increase in troublesomeevents indicates a focused pattern. Finally, ifone or two days have a marked concentrationof events, this indicates an acute pattern.

Hot spots can be classified into one of threetypes: area concentrations, line concentrations,and point concentrations (Step 21). Along withthe absence of any hot spot, we have fourincreasingly precise forms of concentration.Ratcliffe’s types of temporal concentration alsovary from least to most precise. These are com-bined in Figure 3. Problems toward the upper

right can be addressed more readily than prob-lems to the lower left.

ANALYSE IN DEPTH

AN

ALY

SE

Jerry Ratcliffe (2002). Aoristic Signatures and theSpatio-Temporal Analysis of High Volume Crime Patterns.Journal of Quantitative Criminology, 18(1): 23–43.

Jerry Ratcliffe (2003) The Hotspot Matrix: A Frameworkfor the Spatio-temporal Targeting of Crime Reduction.Police Practice and Research. Forthcoming.

Read more

Incre

asing

pre

cision

in und

erstan

ding o

f pro

blem

Incre

asing

focu

s of r

espo

nse

Incre

asing

likeli

hood

of r

espo

nse be

ing effe

ctive

Decre

asing

pot

entia

l sta

keho

lders

in pro

blem

Less

ening

pot

entia

l disr

uptio

n fro

m resp

onse

Less

ening

pot

entia

l cos

ts fr

om re

spon

se

Tem

pora

l con

cent

ratio

n

Spatial hot spot type

Acute

Focused

Diffused

None Area Line Dot

Figure 3: Combining temporalconcentration and hot spots

Hours1

Even

ts

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

DIFFUSED

Hours1

Even

ts

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

FOCUSED

Hours1

Even

ts

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ACUTE

0

5

10

15

20

25

30

Within a spatial hot spot,a diffused temporal patternindicates no particulartime concentration

Focused temporalpatterns suggest generaltime periods of high activity

Acute temporalconcentration of eventssuggest very specific timeperiods of hot spot activity

Figure 2: Ratcliffe’s typology of temporal concentration

Study the problem by putting data in a graphof either the number of events or a rate plot-ted against time. A rate is typically the numberof crime or disorder events divided by thenumber of targets at risk (Step 25). If both thenumber and the rate are plotted, and theyhave the same shape, then change in thenumber of available targets does not play animportant role in changing the problem. If thetwo graphs look different, then targets are animportant consideration.

The time course of a problem can be dividedinto three basic components:

l The overall trend, which may be obviousfrom visual inspection, and which showswhether the problem is getting worse,better or staying the same over a longperiod.

l Seasonal, daily and weekly cycles.

l Random fluctuations that are caused bya large number of minor influences.

Figure 1 shows the 26-month time frame(black line) of a hypothetical commercial bur-glary problem, from May 2001 to June 2003.Early in this time frame there are burglaries,and some months have none. After October

2001 the problem begins to grow and thenafter April 2002 it oscillates around threeevents per month. The period ends with arecord of seven burglaries. We do not knowwhether this portends a new growth spurt ormerely random variation.

Throughout the time frame, there is consider-able monthly variation, shown by the jaggedpeaks and valleys. Low intensity problems, likethis one, tend to have this characteristicwhereas problems with many events per timeperiod often show smoother changes. Theserandom fluctuations can hide systematic varia-tion. One method for revealing a trendobscured by random variation is to use amoving average (blue line in Figure 1). This iscalled ‘smoothing’. A three-month movingaverage was used in this example. A July value,for example, is the average of May, June andJuly, while the August value is the average ofJune, July and August. Notice that there is nodata for the first two months of the seriesbecause we do not have three months of datafor these months. Moving averages fill in thevalleys and knock off the peaks. Longermoving averages produce smoother graphsthan shorter ones, but they can also hideuseful information by making the graph toosmooth.

24 Take account of long-term change

Even

ts

8

7

6

5

4

3

2

1

0M J J A S O N D J F M A M J J A S O N D J F M A M J

3-month moving average

Actual events

2001 2002 2003

Figure 1: Looking for trends

Cycles can be detected by comparing the samemonths of the year (or same weeks of themonth, or same days of a week, or same hoursof a day, depending on the time periods you areexamining). Its important to note that monthsare of different lengths (and do not forgetFebruary in leap years) as this might influencethe number of problem events. Figure 2 plotseach year separately, rather than in a continuousstring. Two things are readily apparent. First, aswe saw in Figure 1, the problem increased: bur-glaries in 2002 and 2003 are higher thanburglaries in the same months of 2001. Second,some months appear to be consistently high andothers consistently low, relative to other months:July, November, December, and February arepeaks in 2001 and 2002, while August,September and January are valleys. As we arelooking at only two years of data, this is not solidevidence of seasonal fluctuation. Nevertheless, itdoes suggest that seasonal effects might play arole in this problem.

The average for each month produces a linethat gives a clearer picture of the seasonal fluc-tuations (solid blue line). This is another formof smoothing and also reduces the confusioncaused by random fluctuations.

Decomposing a time series into componentparts can reveal possible causes of a problem.

The commercial burglary trend, for example,could be decomposed into two charts showingthefts of computer equipment and thefts ofother things. So if the thefts of computer equip-ment were trending upward while thefts ofother things were remaining stable, this wouldsuggest that attention should be focused onstores selling computer equipment.

Time frame analysis is a powerful tool for eval-uating the effectiveness of a response. Thebasic principle is to obtain a good idea of aproblem’s natural trends, cycles, and variationbefore the response is implemented, using thetechniques just discussed. This tells you whatyou can expect from the problem in thefuture, if you did nothing about the problem.This provides a basis for examining timeframes after the response. Changes in thetrend, cycles, or even the random fluctuationsuggest the response had an impact. Thelonger the time frames before and after, thegreater the confidence you can have in yourconclusions.

Time frame analysis can also be very complex,so if there is a great deal depending on a pre-cise answer to a time frame analysis, it may beuseful to seek the help of a statistician special-ising in this area.

ANALYSE IN DEPTH

AN

ALY

SE

Even

ts8

7

6

5

4

3

2

1

0May June July Aug Sept Oct Nov Dec Jan Feb Mar Apr

Yr 2001

Yr 2002

Yr 2003

Average

Figure 2: Looking for seasons

Rates describe the number of crimes per targetat risk, during a period of time. A target rate, forexample, might be one burglary for every 1,000households during 2002. Target rates describethe risk the average target has of being involvedin a crime during the time period.

Calculating rates can be very helpful in diag-nosing a problem. This is illustrated by someanalyses undertaken by Nanci Plouffe, KarinSchmerler and Rana Sampson to support aPOP project on vehicle thefts in Chula Vista,California. Officers believed that the city’s closeproximity to the Mexican border exacerbatedits vehicle theft problem. The fact that ChulaVista’s theft rates, and those of other citiesclosest to the border, were two to six timeshigher than ones further north in San DiegoCounty supported this hypothesis (Figure 1).Many vehicles were stolen from parking lotsand driven across the border before ownershad even discovered the thefts. Analysis had

also suggested that pick-up trucks – highlyprized in Mexico – were at greater risk of theftin Chula Vista and in the two other cities clos-est to the border. Figure 2 confirms that therecovery rate of these trucks when stolen inChula Vista was generally lower than in citiesfurther from the border. Other analysesshowed that some lots had much higher theftrates than others and, ultimately, the analysisshowed that border-point interventions wereless valuable in preventing auto theft thanefforts to improve the security of lots.

The value of calculating rates is also illustratedby a project in Charlotte, North Carolina onwhich one of us (Ronald Clarke) worked withHerman Goldstein. Assisted by local analysts andpolice officers, we examined thefts from cars inparking facilities in the downtown area of thecity (locally known as ‘Uptown’). Hot spot analy-sis had shown a large undifferentiated cluster ofthese thefts centred in the middle of the area,

25 Know how to use rates anddenominators

San Diego County

San Diego Southern Div

EscondidoSan MarcosCarisbad

Vista

Oceanside

Encinitas

Solana Beach

Del Mar

City of San Diego

Poway

Santee

El Cajon

La Mesa

Lemon Grove

National City

Chula Vista

Mexico Border

Coronado

Approx. 60 m

iles - Oceanside to M

exico Border

23%

43%

73%

58%

52%

72%

68%

79%74%

77%

N

Figure 2: Recovery rates for stolenpick-up trucks, San DiegoCounty, 2001

San Diego Southern Div

Escondido

San Marcos

Carisbad

Oceanside

City of San Diego

Poway

Santee

El Cajon

La Mesa

National City

Chula Vista

Mexico Border

Coronado

Approx. 60 m

iles - Oceanside to M

exico Border

12.57

8.803.61

2.45

4.17

N

3.84 7.05

1.67

9.10

6.78

9.84

15.89

Imperial Beach

6.00

1.91

Figure 1: Vehicle theft rates per 1,000 residents, San Diego County, 2001

but a map based on rates of theft was far morerevealing (Figure 3). This map was producedby one of the crime analysts on the project,Matt White, who enlisted the help of precinctofficers in counting the number of parkingspaces in each facility. He then calculated theftrates for each lot and parking deck. The result-ing map revealed a much more detailedpicture of risk. Further analysis showed thatcars parked in lots were six times more at riskthan ones in decks and that some lots werecrime enablers due to inadequate security.

Using rates to determine crime risksin a large hotel

Lawrence Sherman examined a large Dallashotel with 1,245 reports of crime over a two-year period. To determine the crime rate heused information about the number of rooms(1,620), average room occupancy rate (1.8guests per room), number of employees(1,000), and the number of other patrons andemployees on the premises. All these peoplecould be considered at risk of being a victim.The robbery rate for them was 1.2 per 1,000(compared to 4.9 for Dallas as a whole). Thetheft rate, however, exceeded the city theftrate (10.1 compared with 6.4 per 1,000).

Source: Lawrence Sherman (1989). Violent StrangerCrime at a Large Hotel: A Case Study in RiskAssessment Methods. Security Journal, 1(1): 40–46.

ANALYSE IN DEPTH

AN

ALY

SE

Ronald Clarke and Herman Goldstein (2003). Theftsfrom Cars in Center-City Parking Facilities: A CaseStudy in Implementing Problem-Oriented Policing. InProblem-oriented Policing. From Innovation toMainstream. Crime Prevention Studies, vol. 15, editedby Johannes Knutsson. Monsey, New York: CriminalJustice Press (and Willan Publishing, UK). (Downloadfrom: www.cops.usdoj.gov)

Read moreLegend

Theft from cars

Lower

Higher

InterstatesMajorstreetsStreetsRail Lines

Charlotte Uptown

Figure 3: Rates of theft from carsby block, Charlotte NC, 1999

Facilities are environments with special func-tions (Step 16). Educational facilities involveteaching and study. Industrial facilities pro-duce and process materials. Office facilitiesprocess information. Retail facilities involvesales and monetary transactions. Some facili-ties are frequent sites for crime and disorder.Examples include pubs, car parks, railway sta-tions and, in America, convenience stores andpublic housing projects. These relatively fewsites make a disproportionate contribution tocrime and disorder – they are ‘risky facilities’.

But the term has also a more precise meaning.It refers to the fact within each type of facility,a few of them are especially risky. In discussingthe 80–20 rule (Step 19) we gave an exampleof pubs, which varied greatly in their risks ofassault. Here are some more examples:

l Banks – Four per cent of UK bankbranches have rates of robbery four to sixtimes higher than other bank branches.

l Businesses – Surveys of businesses inScotland show that 10% of them accountfor 40% of the business burglaries andalmost three-quarters of the non-employeethefts.

l Shops – 1.6% of shops in Australia experi-ence 70% of shoplifting.

l Convenience stores – 6.5% of conven-ience stores in America experience 65% ofall robberies.

l Bus stops – about 7% of Wirral bus stopsexperience 70% of vandal attacks.

l Schools – Eighteen per cent of Merseysideschools reported 50% of the burglary andcriminal damage.

26 Identify risky facilities

Reported car crime in Nottingham city centre car parks 2001Car park Spaces Type* Theft Theft Other car Crimes per

from of crimes 1,000 spaces

Victoria Centre (White Zone) 1,066 M-S 0 0 0 0.0

Talbot Street 590 M-S 0 0 0 0.0

Forest Park and Ride 3,000 S 8 0 3 3.7

Victoria Centre (Main Zone) 1,700 M-S 7 0 2 5.3

Trinity Square 335 M-S 0 1 2 9.0

Broadmarsh Centre 1,200 M-S 19 1 3 19.2

Wollaton Street GNCS 125 M-S 1 0 2 24.0

Stoney Street 600 M-S 13 0 2 25.0

Mount Street 425 M-S 13 0 4 40.0

Fletcher Gate 550 M-S 19 1 6 47.3

Curzon Street 167 S 6 1 2 53.9

Arndale 412 M-S 25 0 1 63.1

St James Street 475 M-S 31 1 5 77.9

Sneiton Market 50 S 6 0 1 140.0

Royal Moat House Hotel 625 M-S 78 0 25 164.8

EuropaPS 225 M-S 37 0 2 173.3

Brook Street 56 S 16 0 7 410.7

Gill Street 49 S 14 2 7 469.4

Huntingdon Street 75 S 34 2 5 546.7

*M-S = Multi-storey; S = Surface Source: David G. Smith et al. (2003) Between the Lines, Home Office Research Study 266.

l Parking facilities – In Basingstoke, half of the thefts of, from and damage to vehiclescame from only five car parks. Similar resultswere found in downtown parking lots inCharlotte, North Carolina (see previousstep). The concentration of risk inNottingham was even greater (see table).Just one car park (Royal Moat House Hotel)of the twenty in the city centre accounted forabout 25% (103) of the 415 crimes reportedfor all the car parks in 2001. Even so, thislarge car park did not have the highest risk ofcrime per 1,000 parking spaces. The verywide variation among car parks in these risks(0.0 to 546.7) underlines the degree of crimeconcentration in risky facilities.

There are at least seven reasons for risky facil-ities and different analysis procedures canhelp determine which is operating in particu-lar circumstances:

1. Many targets. Some facilities containmany targets. The Royal Moat House carpark in Nottingham accounted for so manycrimes partly because it was so large. Butthis was not the whole story because, whenaccount is taken of its size by calculatingcrimes per parking space, it is still one ofthe riskiest facilities.

2. Hot products. A risky facility may not havea large number of targets, but it might havethe types of targets that are particularly‘hot’ (see Step 29). Compare the thingsbeing stolen in risky and non-risky facilities.If the things taken are different, and thethings being taken from the risky facilitiesmeet the CRAVED criteria described in Step29, then hot products are the likely cause ofthe elevated risk.

3. Location. Facilities located in high crimeareas, perhaps where many habitual offend-ers live, are more likely to be crime risks.This is because offenders prefer not totravel far to commit crime (Step 17).

4. Repeat victimisation. Some places attractpeople who are particularly vulnerable to

crime. Compare the people being vic-timised in risky and non-risky facilities. Ifthe re-victimisation rates are different, thenrepeat victimisation may be the cause ofthe elevated risk.

5. Crime attractors. Facilities that draw inlarge numbers of offenders are crime attrac-tors (Step 17). Crime attractors have highnumbers of offences and high offencerates. Additional diagnostic checks involveanalysis of arrest records and other infor-mation containing offender names.

6. Weak controls. Owners of most facilitiesregulate conduct. When conduct is not reg-ulated a risky facility can develop. These arecalled crime enablers (Step 18). Risky facili-ties that are crime enablers can have lownumbers of offences, but high rates ofoffending. However, more definitive testsfor weak controls include observations ofsimilar high and low crime facilities, exami-nations of employee rules and procedures,and interviews of workers and patrons.

7. Provocations. The physical design or theway a place is managed can provoke mis-conduct (Step 36). Provocative facilities aresimilar to weak control facilities, but whereweak control sites simply permit misbehav-iour provocative sites stimulate it. Use thesame analysis procedures as for weak con-trols but also look for circumstances thatexcite misbehaviour.

Facilities become risky for many reasons. RossHomel of Griffiths University found that someAustralian problem pubs marketed themselvesto heavy drinking males. This attracted bothlikely offenders and victims. He also found thatthe behaviour of staff exacerbated the problemby the way they sold drinks and the manner inwhich they handled disruptive clients. In thisexample we see repeat offending, repeat vic-timisation, weak controls and provocations.

ANALYSE IN DEPTH

AN

ALY

SE

Forthcoming article on ‘Risky facilities’ by the authors.

Read more

Some people are repeatedly victimised and arather small proportion of victims account for alarge proportion of all victimisations. In fact,according to the British Crime Survey, about4% of people experience about 40% of all vic-timisations (see the table). Repeat victimisationhas been found in domestic violence, burglary,racial attacks and sexual assault, to name but afew examples. Like risky facilities and repeatoffending, repeat victimisation is a variation onthe 80–20 rule (Step 19).

It is easy to miss the extent of repeat victimisa-tion for several reasons:

l Many victims do not report crimes to thepolice, which means that repeat victimisationis undercounted in official police records.This is why researchers try to use surveys.With surveys, people can be asked aboutcrimes they did not report to the police.

l Analysts often look for repeat victimisation bycounting the number of crimes at addresses,but police data often contains inaccurateaddress information. This leads to higher esti-mates of one-time only victimisations than isactually the case. This difficulty can bereduced by increasing the accuracy of policedata and through the use of address match-ing in mapping software (i.e. geocoding).

l Repeat victimisation can be underestimatedbecause of the ‘time-window effect’. If onlyvictimisations during a specific time periodare counted – a time window of January2002 through June 2002, for example –then someone who had been victimised inDecember 2001 and once during the six-month window would not be counted as arepeat victim. If they had the misfortune tobe victimised in July 2002, we would notknow that this person had three victimisa-tions. The Home Office recommends usinga rolling window where each new victim isfollowed for a year after the first event.

Knowledge of repeat victimisation is useful forpredicting which people are most at risk andwhen they are at most risk. People who havebeen victimised once have a greater chance ofbeing victimised in the future than people whohave not been victimised. People victimisedtwice have a greater chance of another victimi-sation than people only victimised once. Thetime between repeat victimisations is often

rather short. Knowing the average timebetween crimes makes it possible to temporar-ily deploy crime prevention for short periodswhen the risk of crime is the greatest.

Crime prevention resources can be focused onpeople who have the highest risk when theyare at most risk rather than spread over a largenumber of people, most of whom have a verylow risk of crime. Many forces now use a‘graded response’ when dealing with repeatvictims. This means that the more often some-one has been victimised the more intensivethe preventive action taken by the police.

In explaining repeat victimisation, Ken Peasedistinguishes two kinds of accounts:

1. Boost accounts explain repetitions interms of positive experiences at the initialoffence. A burglar, for example, learns agreat deal during a break-in. This know-ledge may encourage him to come back foranother break-in. A burglar may also tellothers about goods he left behind, leadingto subsequent attacks by other burglars.

2. Flag accounts explain repetitions in termsof the unusual attractiveness or vulnerabilityof particular targets that result in their vic-timisation by a variety of offenders. Someprofessions have much higher victimisationrates than others (taxi drivers for example)and people who spend time in risky facili-ties are also more prone to repeatedvictimisation. Finally, the ownership of hotproducts, such as cars attractive to joyriders(Step 29), will also increase the probabilityof repeat victimisation.

27 Be ready for repeat victimisation

About 4% of people experience about 40% of all crimesCrimes Per cent of Per cent of reported respondents incidents

0 59.5 0.01 20.3 18.72 9.0 16.53 4.5 12.44 2.4 8.85+ 4.3 43.5

Source: British Crime Survey, 1992 all offences

Virtual repeats (also referred to as nearrepeats) involve victims who have characteris-tics like those of the first victim. Aftersuccessfully attacking the first target, offendersgeneralise to targets with similar characteris-

tics. Houses with the same lay-out and in thesame neighbourhood as the first burglary, forexample, can be expected to have higher risksbecause the offender has learned somethingabout them from breaking in before.

ANALYSE IN DEPTH

AN

ALY

SE

Graham Farrell and colleagues (2002). The Time-Window Effect in the Measurement of RepeatVictimisation. In Analysis for Crime Prevention, CrimePrevention Studies, vol. 13, edited by Nick Tilley.Monsey, New York: Criminal Justice Press (and WillanPublishing, UK).

Home Office (2002). Crime Reduction Tool Kits: RepeatVictimisation – Timescale.http://www.crimereduction.gov.uk/toolkits/rv030104.htm.

Ken Pease (1998). Repeat Victimisation: Taking Stock.Crime Prevention and Detection Paper 90. London:Home Office.

Read more

Neighbour beware!

All crime analysis involves saying things aboutthe future on the basis of what has happenedin the past. Repeat victimisation tells us of anelevated risk that the same victim will sufferagain, most often in the immediate days orweeks following the preceding crime. But acrime communicates more than that. Itillustrates how risk will be communicated tonearby places. Kate Bowers and ShaneJohnson have shown how burglary risk iscommunicated down a street. This is illustratedin the graph. A home is burgled. Let us callthat the reference burglary. The numbers at thebottom are a measure of distance from thelocation of the reference burglary. A distance ofone tells of a home next to a burglary locationon the same side of the street, or the home

immediately opposite. A distance of two refersto homes two doors down on the same side ofthe street, or diagonally opposite, and so on.The ordinate shows the number of burglariesfollowing reference burglaries. The data comefrom Merseyside Police. It will be seen that therisk of another burglary declines the further thedistance from the reference burglary. For anygiven distance, the risk is greater for homes onthe same side of the street. This shows whichhomes one should seek to protect in the wakeof a burglary. Priority should be given to homesclose to the burgled home and especially onthe same side of the street.

Source: Two forthcoming articles in the European Journal of Criminology by Kate Bowers and Shane Johnson: (1)‘The Burglary as Clue to the Future: the Beginnings ofProspective Hot-spotting’, and (2) ‘Domestic BurglaryRepeats and Space-time Clusters: the Dimensions of Risk.’

Num

ber

of b

urgl

arie

s

4,500

4,000

3,500

3,000

2,500

2,0001

Same side equivalent

Opposite side

2 3 4 5 6 7 8 9 10

Number of houses apart

One of the fundamental facts of criminology isthat a small proportion of individuals commita large proportion of crime. Data from MarvinWolfgang’s famous Philadelphia cohort sug-gested that around 5% of offenders accountfor 40% of crimes. There are two explanationsfor repeat offending, the first of which is thatcareless individuals, with weak social attach-ments to others, tend to get into trouble morefrequently than less impulsive and moreattached individuals. The second explanationis that people adapt to crime and disorderopportunities (see Step 10). Both of these the-ories can be true.

Repeat offending can be detected by testingfor the presence of the 80–20 rule (Step 19).This can be difficult in practice becauseoffenders try to remain anonymous so the datais seldom comprehensive, and may not exist.Intelligence information can also provide evi-dence of repeat offending, but thisinformation is often more suggestive thandefinitive. Nevertheless, systematic interviewswith offenders and their associates sometimescan reveal networks that are major contribu-tors to problems.

Understanding repeat offenders’ objectivesand motives can help create prevention strate-gies. It makes a difference to the solution to avehicle theft problem if the thieves want tohave a good time riding around in a fancy car,to obtain transportation home after a latenight of partying, or to sell it for cash to sup-port a drug habit.

Successful offending can lead to more offend-ing. This occurs in three ways:

l Offenders, like others, learn from doing. Asuccessful crime teaches important lessons.This can lead to the offender attacking thesame target again (see box). But offenders,like everyone else, can generalise. So theylearn that they may be successful if theyattack similar targets (see Step 27).

l Offenders learn from each other.Information can spread through individualsworking in small groups, group breakupand new group formation. This under-scores the need to understand offendernetworks. Police can use networks tospread information that enhances offend-ers’ perception of risks or lack ofdesirability of the target or place. Part of theeffort to reduce glass bottle injuries by theMerseyside Police, for example, involvedhighly targeted advertising to potentialoffenders and victims about proper dis-posal of glass beer bottles.

l Successful offending can erode prevention,thus making subsequent offending easier. Asmall break in a fence, for example, willbecome larger with use. If the influx ofoffenders and offensive behaviours is fasterthan the responses of guardians and placemanagers, then a small problem willbecome worse.

Many crime prevention techniques rest on theassumption of a credible threat (Step 34).CCTV provides a deterrent threat to the extentthat potential offenders believe either thatsomeone is watching who will take actionshould they see misbehaviour, or that offend-ers can be identified and arrested later basedon CCTV recordings. This does not mean thatthere have to be many arrests, but a few well-publicised arrests can reinforce an importantmessage. And the message can be made morepowerful if it is communicated throughoffender networks rather than relying onnormal publicity.

When there is specific information that a fewpeople are responsible for most of a problem,it can be productive to focus on these individ-uals. The Boston Police Department, in theUnited States, reduced homicides amongyoung males by monitoring a relatively fewgang members. Environmental correctionsuggests that probation and parole authoritiesshould learn the specific circumstances under

28 Consider repeat offending

which each offender gets into trouble, thenhelp offenders develop plans to avoid thesecircumstances, and finally monitor compliancewith these plans.

Tackling repeat offending through removingfacilitating environments can sometimes bevery effective. For example, in Staining, a vil-lage in Lancashire, a scrap yard served as areceiver for stolen vehicles, parts, and otherloot from thefts. Many of the associated offend-ers were known. But despite policeenforcement efforts this problem could not beresolved. Constable Farrand was able to closethe site using laws governing pollution andother environmental hazards. This substantiallyreduced crime in the village. Similarly, police inthe United States often use civil laws to closedown facilities that foster drug dealing, prosti-tution, and other crimes and disorder.

Catch prolific criminals by focusingon repeat victimisation

Ken Pease has recently written about thebenefits for detection resulting from a focuson repeat victimisation. Evidence isaccumulating that repeat victimisations are

the work of the most committed offenders. Hepoints out that this raises the intriguingpossibility that offender targeting may beachieved simply by detecting repeatedoffences against the same household orperson, since these offences are committedby offenders whom one would in any casewish to target. This kind of offender targetingwould avoid all the aspects of an approachderiving from the harassment of knownoffenders, since it would focus not on people,but on the subset of acts that prolificoffenders habitually commit.

ANALYSE IN DEPTH

AN

ALY

SE

Francis Cullen, John Eck and Christopher Lowenkamp(2002). Environmental Corrections: A New Frameworkfor Effective Probation and Parole Supervision. FederalProbation, 66

John Eck (2002). Preventing Crime at Places. InEvidence-Based Crime Prevention, edited by LawrenceSherman, David Farrington, Brandon Welsh and DorisLayton MacKenzie. New York: Routledge.

David Kennedy, Anthony Braga, Anne Piehl and ElinWaring (2001). Reducing Gun Violence: The BostonGun Project’s Operation Ceasefire. Research Report.Washington, DC: National Institute of Justice.

Read more

Repeat offenders, repeat victims, hot spotsand risky facilities are all concepts thatdescribe important ways in which crime isconcentrated. Crime is also concentrated onparticular ‘hot products’, which are likely to bestolen, including cars, bicycles, videomachines and mobile phones. The hottestproduct of all is cash, which Marcus Felsondescribes as the mother’s milk of crime. Asshown by the British Crime Survey, it is themost frequently stolen item in larcenies, bur-glaries and robberies. It fuels robberies ofbanks and betting shops, attacks on phoneboxes, and muggings near bank machines.

The things people own can help explain theirvictimisation risks. For example, owning a cardoubles the risk of becoming a crime victim,even when account is taken of relevant demo-graphic and social variables. And the particularmodel of car owned can raise this risk manytimes over. This is why the Home Office pub-lishes the Car Theft Index that gives risks oftheft for different models. The Index helpspeople avoid purchasing a theft-prone car andputs pressure on manufacturers to improvethe security of cars.

Useful as it is, the Index does not show whichcars are most at risk from specific forms oftheft. American research in the mid 1980sfound that the models preferred by joyriderswere American-made ‘muscle’ cars with pow-erful acceleration. Those stolen and neverrecovered were expensive cars such asMercedes, and those broken into and strippedof contents were European models with goodradios, such as VWs. American-made stationwagons were not at risk of any form of theft.These were inexpensive, had terrible radios,and joyriders wouldn’t be seen dead in them.

Shops carrying hot products (such as ciga-rettes, videocassettes, CDs, brand-nameclothing and footwear) are also more vulnera-ble to shoplifting and burglary. Many of theseitems can readily be sold on the street.Knowing what is ‘hot’ in your area can helpyou explain patterns of theft and help you

think about how stolen goods are sold andhow to disrupt the market. Police have gener-ally paid little attention to fencing because it isdifficult to prove and often attracts relativelylight sentences. But if thieves found it harderto fence goods, the incentive for ‘volume’thefts would be reduced.

The acronym CRAVED will help you rememberwhich goods are most stolen. In general theseare Concealable, Removable, Available,Valuable, Enjoyable and Disposable.

l Concealable. Things that can be hidden inpockets or bags are more vulnerable toshoplifters and other sneak thieves. Thingsthat are difficult to identify or can easily beconcealed after being stolen are also moreat risk. This explains why we write ournames in books and why car thieves do notgenerally steal Rolls Royces for their ownuse. Instead, they steal less valuable carsthat merge into the surroundings. In somecases, thefts may even be concealed fromthe owners of goods, as when timber orbricks left lying around on building sites arestolen.

l Removable. The fact that cars and bikesare mobile helps explain why they are sooften stolen. Nor is it surprising that laptopcomputers and VCRs are often stolen sincethese are desirable and easy to carry. Theimportance of these factors depends on thecircumstances of theft, as borne out byAmerican data on targets of theft fromsupermarkets. Both burglars and shop-lifters target cigarettes, liquor, medicinesand beauty aids, but burglars take them inmuch larger quantities.

l Available. Desirable objects that are widelyavailable and easy to find are at higher risk.This explains why householders try to hidejewellery and cash from burglars. It mayalso help explain why cars become more atrisk of theft as they get older. They becomeincreasingly likely to be owned by peopleliving in poor neighbourhoods with less off-street parking and more offenders living

29 Know which products are CRAVED bythieves

nearby. Finally, theft waves can result fromthe availability of a new attractive product,such as the mobile phone, which quicklyestablishes its own illegal market.

l Valuable. Thieves will generally choose themore expensive goods, particularly whenthey are stealing to sell. But value is notsimply defined in terms of resale value.Thus, when stealing for their own use, juve-nile shoplifters may select goods thatconfer status among their peers. Similarly,joyriders are more interested in a car’s per-formance than its financial value.

l Enjoyable. Hot products tend to be enjoy-able things to own or consume, such asliquor, tobacco and cassettes. Thus, resi-dential burglars are more likely to takevideos and televisions than equally valuableelectronic goods, such as microwave ovensor food processors. This may reflect thepleasure-loving lifestyle of many thieves(and their customers). Fashion items suchas trainers or designer jeans are much morelikely to be stolen than ordinary shoes ortrousers.

l Disposable. Only recently has systematicresearch begun on the relationshipbetween hot products and theft markets,but it is clear that thieves will tend to selectthings that are easy to sell. This helpsexplain why batteries and disposable razorsare among the most frequently stolen itemsfrom American drug stores.

Which lorries are stolen?

Home Office research conducted by RickBrown shows that body-type significantlydetermines which lorries are stolen.

The highest risk of theft was found forlivestock carriers, many of which are privatehorseboxes. These were 56 times as likely tobe stolen as refuse carriers! There is a thrivingsecond-hand market in horseboxes, whichwould make it easier for thieves to sell them.Tippers and drop-side lorries used by theconstruction industry are also at high risk,probably for the same reason.

Heavy goods vehicles stolen, England and Wales 1994Body Number Per cent of Theft rate type stolen incidents per 1,000

registered

Livestock carrier 156 5.1 56

Drop-side lorry 582 19.1 27

Tipper 920 30.2 16

Flat-bed lorry 565 18.5 14

Skip loader 86 2.8 13

Goods lorry 349 11.5 9

Insulated van 88 2.9 7

Bottle float 12 0.4 3

Tanker 29 1.0 2

Refuse disposal 10 0.3 1

Other 248 8.1 1

TOTAL 3,047 99.9 6

ANALYSE IN DEPTH

AN

ALY

SE

Ronald Clarke (1999). Hot Products. Understanding,Anticipating and Reducing the Demand for StolenGoods. Police Research Series. Paper 112. London:Home Office.

Read more

Crime facilitators help offenders commitcrimes or acts of disorder. There are threetypes of facilitators:

l Physical facilitators are things that aug-ment offenders’ capabilities or help toovercome prevention measures. Lorriesextend offenders’ capacity to move stolengoods, telephones allow people to makeobscene phone calls, and firearms helpovercome resistance to robberies. Somephysical facilitators are tools, but others arepart of the physical environment. BarryPoyner and Barry Webb describe how thearrangement of stalls in a Birmingham retailmarket, for example, facilitated thefts fromwomen’s handbags.

l Social facilitators stimulate crime or disor-der by enhancing rewards from crime,legitimating excuses to offend, or byencouraging offending. Groups of youngmen, for example, can provide the socialatmosphere that encourages rowdy behav-iour at sporting events.

l Chemical facilitators increase offenders’abilities to ignore risks or moral prohibi-tions. Some offenders, for example, drinkheavily or use drugs before a crime in orderto decrease their nervousness.

Each type of facilitator acts against particularforms of situational crime prevention (Steps33–37), as shown in the table. Physical facilita-tors help offenders overcome preventivemeasures that increase risk or effort. They canalso act as provocations to deviancy. Socialfacilitators can increase the perceived rewardor the acceptable excuses for committing acrime, and they can provoke crime or disorderthrough encouragement. Chemical facilitatorsallow offenders to ignore the risk and effortinvolved in committing a crime. And theyallow offenders to make unacceptable excuses.

Because of their capacity to blunt crime pre-vention, it is important to identify the role offacilitators in a problem. Evidence about facili-tators can be found in investigative reports

and from investigators, by interviewing victimsand offenders, and by observing social situa-tions. Statistical analysis of crime reports canbe used to determine the association betweencrimes and various facilitators.

If facilitators do play a role in the problem,then the next step is to find the sources of thefacilitators. Sources will, of course, vary by typeof facilitator. Physical facilitators might be read-ily available, as in the case of paving stones forrioters or public phones for drug dealers. Orthey may be purchased legitimately, as is thecase for many burglary tools. Or they may bestolen, as is sometimes the case with vehiclesused in serious crimes. Having found theirsource it may be possible to do somethingabout them. The list opposite shows measurestaken to address the use of public phones indrug dealing, and facilitating environmentsaround cash machines.

Social facilitators depend heavily on whomoffenders associate with, and the settings forthe association. Risky facilities (Step 26), forexample, can provide settings for social facili-tation. Gangs provide the social support forcrime. But even legitimate activity can on occa-sion spark social facilitation, as in the case withsome politically motivated violence.

Chemical facilitators are abundant and fre-quently associated with crime and disorder.Alcohol is particularly implicated as a

30 Look for crime facilitators

Crime facilitators versus situational prevention methodsSituational Type of facilitatorprevention method

Physical Social Chemical

Increase risk ← ←

Increase effort ← ←

Reduce rewards ←

Remove excuses ← ←

Reduce provocations ← ←

facilitator. Various mixtures of facilitators arecommon, particularly social and chemical inentertainment venues. Several of the 25 tech-niques of situational crime prevention aredesigned to reduce the effect of the threekinds of crime facilitators (Steps 33–37).

Measures to prevent use of publicphones by drug dealers in U.S. cities

Before cell phones became widely available,drug dealers in the US often relied on the useof public phones to make contact withsuppliers and customers. Many ways to stopthem were tried, including:

1. City ordinances to license public phonesand ban them or limit their number atspecific locations or categories of location.

2. Installation of rotary dials that do notpermit outgoing calls to pagers.

3. Modification of phones to block incomingcalls.

4. Community pressure on local phonecompanies or the city government toremove public phones or relocate them inbetter lit or supervised areas.

5. Permitting only operator-assisted calls oremergency calls during night hours byblocking coin operation of the phones.

6. Removal or modification of public phonesby businesses such as convenience storesand petrol stations.

7. Other types of intervention such asincreased police patrols, warning labels onphones, and ‘hotlines’ to report problems.

Source: Mangai Natarajan and colleagues (1996). DrugDealing and Pay Phones: The Scope for Intervention.Security Journal, 7: 245–251.

ANALYSE IN DEPTH

AN

ALY

SE

Security provisions for bank cash machines in New York City and Los Angeles

New York LosCity Angeles

Enclosed cash machine l

vestibule with secured entry door

Increased lighting l l

Transparent windows in facility enclosure l

Elevated mirrors for users l ll

Reduced vegetation near l

machine

Surveillance cameras l l

Safety reminders to users l ll

Security provisions notice l l

to potential offenders

Crime assessment prior to l

installation of cash machine

Security guards l*

Reduced cash machine ll

operational hours based on temporal crime patterns in area

l Required by legislation.ll Not required under legislation, but commonlyimplemented at bank’s volition.* Required only during non-banking hours for ATMs locatedinside bank buildings open for customer use. Source: Rob Guerette and Ronald Clarke (2003). ProductLife Cycles and Crime: Automated Teller Machines and Robbery. Security Journal, 16: 7–18.

When you have completed your analysis(using concepts discussed in the previous tensteps), you should ask whether it meets thetest of a good newspaper story. Does it ade-quately answer the 5 W and one H questions:what, where, when, who, why and how?

These same questions structure Barry Poyner’smethod of crime analysis by breaking up alarger problem into its constituent parts. Forexample, when he was asked by the HomeOffice to study ‘street attacks’ in the city cen-tres of Coventry and Birmingham, he foundthat the police classified these as robberies andthefts from the person, but he found that themajority of incidents fell into a number of quitedistinct problems:

l Robbery from street kiosks.

l Robbery of drunks.

l Money snatched while being taken to thebank.

l Handbag snatches.

l Purse/money snatched from hand afterverbal ploy.

l Thefts from shopping bags.

l Pickpocketing at bus stops.

This was a much more meaningful characteri-sation of ‘street attacks’ and was an importantfirst step in understanding the events. He thenbegan to sort through the incident reports,trying to arrive at a picture of each problemthat would help to find a response.

Incident reports are quite variable in the infor-mation recorded, especially when the victim isnot present and there are no witnesses.However, Poyner tries to piece the reportstogether to get a picture of the particularproblem (see box). For each incident he triesto discover:

l What happened? This entails spelling outthe sequence of events and the actions ofthose involved.

l Where did it happen? Sometimes thesequence of events takes place in severallocations. For example, a car might be stolenfrom a car park, moved to a garage for strip-ping of valuable parts, and then dumped ona piece of waste land. Information may onlybe available about the first and last locations.Visiting these can help explain why theoffender selected them.

l When did it happen? Householders orcar owners might know only that their carwas stolen or their house burgled ‘some-time during the weekend’. For manyinterpersonal crimes, however, the victimwill be able to report precisely when thecrime occurred, which may permit infer-ences about such matters as whether thestreets were deserted.

l Who was involved? There is always atleast one offender; there may be one ormore victims even if they have no directcontact with the offender; there may be wit-nesses and other third parties. Statementsin police records made by witnesses and vic-tims can provide much useful information,but it might sometimes be important toquestion a sample about the event.

l Why did they act as they did? It isimportant to understand the specific bene-fits that a particular kind of crime brings tothe offender. In many cases of theft, themotive will be obvious, but for interpersonalcrimes and for vandalism the motives mayonly emerge from interviewing offenders(Step 11). Equally important for preventionmay be to understand why victims and wit-nesses behaved as they did and to answersuch questions as ‘What causes some vic-tims to “have a go” at the offender?’ and‘Why do witnesses often fail to intervene?’

l How did the offender carry out thecrime? Crime can be thought of as a process,with several steps from initiation to comple-tion, rather than a circumscribed actoccurring at a specific point in time. At eachstep the offender must make decisions, mightneed to work with others and might need to

31 Check you have answered the five ‘W’(and one ‘H’) questions

employ specific knowledge and tools. This isessentially the idea underlying Cornish’s‘script’ approach discussed in Step 11. It maynot always be possible to develop detailedscripts, but the analysis should give a clearpicture of how the crime was accomplished.

Poyner’s analysis of pickpocketing at bus stopsillustrates the approach. He was able to con-struct a detailed description of the crime bysupplementing the rather sketchy incidentreport with observations of the queues. Hefound that the peak time for pickpocketingwas the afternoon rush hour, particularly onFridays when queues were long. Groups ofthree or four youths would hang about nearthe queues, looking in the windows of nearbystores to avoid arousing suspicion, whilewatching at the same time for suitable victims.These were invariably middle-aged or oldermen who kept their wallets in the back pock-ets of their trousers. (Younger men woretighter-fitting trousers and did not keep theirwallets in their back pockets.)

As their victims began to board the bus, whichused a pay-as-you-enter system, the youthswould run to the front of the queue, jump onthe boarding platform of the bus and jostle thequeue. They would ask the driver some irrele-vant question about the destination of the bus.Meanwhile one of the youths would pick thepocket of the victim. The victim would be irateat being jostled and would not realise what washappening. The driver would shout at theyouths to get off his bus and other passengerswould be complaining. The youths would stepoff the bus and slip away into the crowd. Thevictim would only find out later that his walletwas gone. None of the youths were ever caught.

This analysis suggested four possible responses:

1. Instead of pay-as-you-enter, tickets mightbe sold in advance for use at these stops.

2. Bus stops might be re-sited away from mainpavements and organised in bays more likea bus station. This would make it difficultfor offenders to lie in wait.

3. Screen the queues by a bus shelter so theoffenders would be unable to identifypotential victims in advance.

4. Construct queue-marshalling barriers at theboarding point so offenders cannot jumponto the bus platform.

Working like an archaeologist

‘There is a further advantage of combiningincidents. In some cases we may have muchless detail than other cases but otherwise thefacts we do know about are the same. It maybe possible to reconstruct the missing data inthese less well reported incidents in much thesame way as the archaeologist reconstructsbroken pottery from an excavation. He mayonly have a few pieces of the broken pot butfrom knowledge of other similar pots he canbe reasonably sure about the form of thewhole pot. This archaeological approach isquite helpful when, for example, we may havesome detailed accounts of what offenders doin a few cases where they have been caught.It seems reasonable to believe that similarbehaviour occurred in similar crimes eventhough the offenders were not caught.’

Source: Barry Poyner (1986). A Model for Action. InSituational Crime Prevention, edited by Gloria Laycockand Kevin Heal. London: HMSO.

ANALYSE IN DEPTH

AN

ALY

SE

The first instinct of police, even when theyhave been involved in a detailed analysis of aproblem, is to try to solve it by beefing upenforcement. You should expect this and notoppose it, even if the impact is usually short-lived. But from the beginning you should behelping your police colleagues find more per-manent solutions. Local community partnerscould provide this help, but instead they oftenpromote their own agendas and push solutionswith limited impact. The result is often a com-promise package of measures, none of which iseffective, but each of which may satisfy one orother of the parties. In fact, the disagreementsover solutions may lead to a loss of momentumand nothing may be properly implemented.

You can help to stop this happening, but youmust first become an expert on solutions. Forexample, if the problem is one of car theft, youmust be able to speak authoritatively aboutthe ineffectiveness of decoy vehicles or ‘lockyour car’ campaigns. If it is a burglary prob-lem, you must know the results of research onburglar alarms or improved street lighting,both of which may be suggested as solutions.You should also be thoroughly familiar withfindings on displacement, since innovativesolutions are often blocked by knee-jerk invo-cations of this theory.

To become expert on solutions you shouldbecome familiar with the many relevant studies on the Home Office website(www.homeoffice.gov.uk) and you must knowhow to find out more about particular responsesby undertaking rapid literature searches. You willneed access to Criminal Justice Abstracts,Security Journal, Crime Prevention andCommunity Safety, Crime Prevention Studiesand other relevant journals. Because evaluativeresearch is scarce, you should look for studiesundertaken overseas by the Australian Instituteof Criminology (www.aic.gov.au) or the NationalInstitute of Justice (www.ncjrs.org) and theOffice of Community-Oriented Policing Servicesin the United States. The latter publishes a seriesof Problem-Solving Guides for Police that arefocused on specific problems. Twenty guides are

currently available and many more are planned(www.popcenter.org) When using research fromoverseas, you may have to explain that, becausecriminals everywhere tend to have similarmotives and use the same methods, the findingsare likely to apply to your local situation.

You must also become an expert on situationalcrime prevention, the science of reducingopportunities for crime. Situational preventionuses the same action-research methodology asproblem-oriented policing and has dozensof evaluated successes to its credit. Much ofthe knowledge about displacement, diffusionof benefits, repeat victimisation, and manyother concepts discussed in this manual havebeen developed by situational preventionresearchers. The next five steps discuss the 25techniques of situational crime prevention,which fall into five main groups (see the box).These are defined by what Nick Tilley calls themechanism through which the techniquesachieve their preventive effect: increasing theeffort of crime, increasing the risks, reducingthe rewards, reducing provocations andremoving excuses.

At this point, you might be asking yourself whyyou should assume this responsibility for iden-tifying solutions. Isn’t it enough that you carrymost of the burden at the scanning, analysisand assessment stages? And even if you didtake on this role, why should anyone pay atten-tion to you? But to become a problem-solvinganalyst you must go beyond your traditionalanalytic function. You must become a full andequal member of the problem-solving team.You may be relatively junior, but your authoritycomes from your expert knowledge, not yourposition. People will listen if you make novelsuggestions, or if you provide supporting evi-dence for other people’s good ideas.

You should always opt for solutions that couldbring a rapid reduction in the problem. Thismeans that you must focus on the immediate,direct causes of a problem rather than the moredistant, indirect ones. This important distinc-tion has been developed by Paul Ekblom and

32 Accept your key role at response

can be illustrated by the problem of bar fightinjuries caused by broken bottles and glasses.Distant ‘root’ causes might include racial discrimination producing a generation of disaffected minority youths, lack of localemployment opportunities resulting in wide-spread social exclusion, and the premiumplaced on a ‘tough’ reputation in a deprivedand lawless community. More immediate, situa-tional causes might include irresponsibleserving practices promoting drunkenness inlocal pubs, and the immediate availability ofbottles and glasses that can easily be used asdeadly weapons.

Rapid and sustained reductions in crime canonly result from addressing situational causes;addressing root causes, even if we knew whatto do about them, can only pay off in the com-paratively distant future – long after the currentstakeholders have any remaining interest in theproblem. Meanwhile, unless the immediatecauses are dealt with, broken glasses and bot-tles will continue to claim victims.

Some situational solutions can also take a longtime to implement. For example, the dangerposed by glasses and bottles could beaddressed by national legislation requiringpubs to use only toughened glasses and bot-tles that disintegrate into crumbs whenbroken. This would probably take years toaccomplish. Much more realistic would be tobring community pressure to bear on localpubs to serve beer only in toughened or plas-tic glasses and to refuse to sell bottles at thebar. This ought to be achievable in a muchshorter time. You might have a particular ana-lytic role in promoting this solution byassembling data about the likely costs for thepubs and the reduced costs of injuries andemergency care. In fact, it will probably fall toyou (who else?) to collect data about the feasi-bility, costs and the public acceptability of anyof the measures that are being seriously con-sidered by the problem-solving partnership.

The bottom line is that unless you acquireknowledge of a broad range of solutions and

unless you are prepared to fight for goodideas, all your careful analytic work may cometo nothing.

FIND A PRACTICAL RESPONSE

RE

SP

ON

D

Ronald Clarke (1997). Situational Crime Prevention:Successful Case Studies. Monsey, New York: CriminalJustice Press.

Derek Cornish and Ronald Clarke (2003).Opportunities, Precipitators and Criminal Decisions:A Reply to Wortley’s Critique of Situational CrimePrevention. In Theory for Situational Crime Prevention.Crime Prevention Studies, vol. 16, edited by MarthaSmith and Derek Cornish. Monsey, New York: CriminalJustice Press (and Willan Publishing, UK).

Read more

Twenty-five techniques of situational crime prevention

Increase the effort1. Harden targets 2. Control access to facilities3. Screen exits4. Deflect offenders5. Control tools/weapons

Increase the risks6. Extend guardianship7. Assist natural surveillance8. Reduce anonymity9. Utilise place managers

10. Strengthen formal surveillance

Reduce the rewards11. Conceal targets12. Remove targets 13. Identify property14. Disrupt markets15. Deny benefits

Reduce provocations16. Reduce frustrations and stress17. Avoid disputes18. Reduce emotional arousal19. Neutralise peer pressure20. Discourage imitation

Remove excuses21. Set rules22. Post instructions23. Alert conscience24. Assist compliance25. Control drugs and alcohol

Here we describe the most basic category of situ-ational measures – those designed to increasethe difficulties of crime – beginning with targethardening. Situational prevention is sometimesdismissed as being nothing more than this,though it is only one of the 25 techniques thatthe approach encompasses. Note that there isconsiderable overlap between the techniques(for example, target hardening can slow upoffenders and increase the risks of their gettingcaught) and that some measures can serve morethan one purpose. When using this classification,do not spend time worrying where a particularmeasure fits. Use the classification only to ensurethat you consider the widest possible repertoireof situational responses to a particular problem.

Harden targets. An obvious, often highly effec-tive way to obstruct the vandal or the thief isthrough physical barriers such as locks, screens,or reinforced materials. Improved design ofLondon Underground ticket machines substan-tially reduced the use of slugs. Screens to shieldbus drivers have significantly reduced assaultsand anti-bandit screens in London post officesreduced robberies by 40%. Strengthened coinboxes have reduced theft and damage to publictelephones in Britain and Australia. The introduc-tion more than 30 years ago of steering locks inGermany, Britain and America produced substan-tial reductions in theft of cars and immobilisersare currently achieving the same result.

Control access to facilities. Stopping peoplefrom entering places where they have no right tobe, such as offices, factories and apartmentblocks, has a long pedigree – think only of theportcullises, moats and drawbridges of medievalcastles. It is also a central component of defensi-ble space, arguably the start of scientific interestin situational prevention. Barry Poyner and BarryWebb found that a combination of entry phones,fencing around apartment blocks and electronicaccess to the parking garage achieved a signifi-cant reduction in vandalism and theft in oneLondon council estate. Poyner has also demon-strated that the installation of entryphones andthe demolition of walkways linking buildings sig-nificantly reduced muggings at another Londonestate (Step 22). In some cases, access controls

are intended to detect prohibited objects or toensure the possession of tickets and documents.Automatic ticket gates on the central zone sta-tions of the London Underground reduced fareevasion by two-thirds throughout the system,while redesigning tickets to make them easier tocheck produced a similar result on Vancouverferries. In the most famous example, baggageand passenger screening at major airports in theearly 1970s contributed to a major reduction inthe number of airline hijackings worldwide fromabout 70 to 15 per year.

Screen exits. The purpose of exit screening isto ensure that those leaving a building, a facilityor some other place have not stolen anything orhave paid all fees and taxes. This techniqueincludes the requirement to surrender ticketson leaving a train, border controls on leaving acountry and the use of electronic tags in librarybooks and shops. These tags activate an alarm ifbooks have not been checked out or if a thieftries to remove a tagged item from the shop.Studies have shown that they significantlyreduce shoplifting and theft of library books.

Deflect offenders. Rival groups of soccerfans are segregated in the stadium to reducefighting, and their arrival and departure hasbeen scheduled to avoid the waiting periodsthat promote trouble. Scheduling the last busto leave immediately after pub closing time isintended to interfere with another of Britain’sless admirable traditions, the closing timebrawl. Reducing congestion around stalls sub-stantially reduced thefts from shopping bags atmarkets in Birmingham because this increasedthe difficulty of pick-pocketing and other‘stealth’ thefts. Roger Matthews has shownthat a road closure scheme to deflect kerbcrawling contributed to the rehabilitation of ared light district in a North London suburb.These are all examples of deflecting offendersaway from crime targets, a situational tech-nique suggested by routine activity theory.

Control tools and weapons. Saloons in theWild West routinely required customers tosurrender their pistols on entry because ofthe risk of drunken gunfights. More recently,

33 Increase the effort of crime

‘safer’ handguns have been developed thatcan only be fired by the owner or which shootwax bullets or tranquillisers. To preventglasses being used as weapons when broken,many pubs in this country now use ‘tough-ened’ beer glasses. The first commercial useof Caller-ID (in New Jersey at the end of the1980s) led to a reduction of one quarter inobscene telephone calls. Step 30 lists the dif-ferent ways in which US cities have attemptedto control the use of public phones in drugdealing, including blocks on incoming callsand banning them from specific locations. Re-programming of public phones at theManhattan bus terminal prevented illegalaccess to the international phone service andwiped out a multi-million dollar scam perpe-trated by hustlers. The introduction ofidentification procedures in Sweden pro-duced a dramatic decline in cheque frauds,while improved security procedures for deliv-ering credit cards produced a substantial dropin credit card frauds in this country in themid-1990s (Step 12).

FIND A PRACTICAL RESPONSE

RE

SP

ON

D

Increase the effort of crimeHarden targets

l Steering column locks, immobilisers and the‘club’.

l Anti-robbery screens at banks and post offices.l Plastic shields for bus drivers.

Control access to facilitiesl Entry phones for blocks of flats.l Electronic card access to garages and offices.l Fenced backyards and the alley-gating

scheme.Screen exits

l Ticket needed to exit.l Export documents.l Electronic tags for shops and libraries.

Deflect offendersl Street closures in red-light districts.l Separate toilets for women. l Disperse pubs from city centre.

Control tools and weaponsl Restrict spray-can sales to juveniles. l Toughened beer glasses.l Stop incoming calls to public phones to foil

drug dealers. l Photos on credit cards and thumbprints on

cheques.

Seven criticisms of situational crime prevention – and rebuttalsCriticism Rebuttal

1 It is atheoretical and simplistic It is based on three crime opportunity theories: routine activity,crime pattern and rational choice.

2 It does not work; it displaces Many dozens of case studies show that it can reduce crime, crime and often makes it worse usually with little displacement.

3 It blames the victim It empowers victims by providing them with information aboutcrime risks and how to avoid them.

4 It promotes a selfish, Defensible space designs for council housing and CCTV on public exclusionary society transport are just two examples of protection given to vulnerable

sections of the population.

5 It promotes Big Brother and is Situational remedies are often barely visible or improve the a servant of repression general quality of life. Prevention is generally less discriminatory

than law enforcement.

6 It diverts attention from the It achieves immediate results and allows time for finding root causes of crime longer-term solutions to crime.

7 It is a conservative, managerial It promises no more than it can deliver. It requires that solutions approach to the crime problem be economic and socially acceptable.

Read more: Andrew von Hirsch, David Garland and Alison Wakefield (2000). Ethical and Social Perspectives on Situational CrimePrevention. Oxford: Hart Publishing.

According to interviews with offenders, theyworry more about the risks of being appre-hended than about the consequences if theyare caught. This makes sense from their pointof view since they can do little to avoid punish-ment if caught, but they can do a lot to reducethe risks of capture by being careful. This iswhy situational prevention seeks to increasethe risks of being caught and makes noattempt to manipulate punishment.

Extend guardianship. Cohen and Felsonshowed that the increase in residential bur-glary in America during the 1960s and 1970swas partly due to the increasing numbers ofwomen working outside the home. Thismeant that for much of the day many houses,if not entire suburbs, had no ‘capableguardians’. Other research has found that bur-glars prefer to commit their crimes onweekday afternoons when people are likely tobe out. This explains why householdersshould cancel newspapers and inform theirneighbours when they go away on holiday.Carrying a cell phone or going out at night in agroup are other ways to extend guardianship.Little is known about the effectiveness of theseroutine precautions and evaluations of neigh-bourhood watch, the only systematic effort toextend guardianship, have not been encourag-ing. However, cocoon neighbourhood watch,under which surrounding homes were alertedafter a burglary, was an important element ofthe successful Kirkholt project.

Assist natural surveillance. Householderstrim bushes outside their homes and banksilluminate their interiors at night to capitaliseupon the ‘natural’ surveillance provided bypeople going about their everyday business.Enhancing natural surveillance is also theprime objective of improved street lightingand defensible space architecture. OscarNewman has recently reported some suc-cesses in reducing crime in American publichousing through the application of defensiblespace principles. Improved lighting in aDudley council estate produced crime reduc-tions with little evidence of displacement.

Robbery rates of convenience stores in Floridawere lower when they were located in areaswith busy evening activity and when the viewof their interiors was unobstructed. Informanthotlines and crime stopper programmes areattempts to capitalise upon the natural surveil-lance provided by the public.

Reduce anonymity. Expanded car owner-ship has allowed people to work in placesdistant from their homes. The development ofout-of-town retail malls has contributed to thedemise of local high-street shopping. Low-costtravel has led to increased tourism both athome and overseas. As a result, people spendincreasing periods of time among anonymousstrangers. The building of large schools is partof this trend because pupils are less wellknown to staff and other pupils. Reducinganonymity is a promising but rarely used situa-tional technique. Some American schools arenow requiring uniforms to reduce theanonymity of pupils on their way to and fromschool. ‘How’s my driving?’ signs with 0800telephone numbers on lorries, and taxi driverIDs displayed for passengers, are two furtherattempts to reduce anonymity.

Utilise place managers. In addition to theirprimary function, some employees also per-form a surveillance role. These ‘placemanagers’ include shop assistants, hotel door-men, car park attendants and bus conductors.Less vandalism has been found on buses withconductors and on council housing estateswith resident caretakers. Public telephonessited in pubs or railway stations where theyreceive some oversight from workers alsosuffer fewer attacks. Canadian research hasfound that apartment blocks with doormenare less vulnerable to burglary. Rewardingcashiers for detection of forged or stolencredit cards helped to reduce annual fraudlosses by nearly $1 million dollars at an elec-tronics retailer in New Jersey. Vandalism on afleet of 80 double-decker buses in northernEngland was substantially reduced when someof the buses were fitted with CCTV for drivers.Having two shop assistants on duty, especially

34 Increase the risks of crime

at night, has been found effective in prevent-ing robbery of convenience stores in theUnited States.

Strengthen formal surveillance. Formalsurveillance is provided by police, securityguards and store detectives whose main func-tion is to furnish a deterrent threat to potentialoffenders. Burglar alarms and CCTV canenhance this surveillance. Vandalism, assaults

and fare dodging were substantially reducedon public transport in Holland when a newforce of inspectors was employed. A bike patrolsucceeded in curbing car thefts from com-muter parking lots in Vancouver. A HomeOffice study found appreciable reductions in avariety of crimes following installation of CCTVin three British cities. New ways of enhancingformal surveillance are provided by linkingdata sets on individuals (see box).

FIND A PRACTICAL RESPONSE

RE

SP

ON

D

Increase the risks of crimeExtend guardianship

l Promote routine precautions such as leavingsigns of occupancy when away from thehouse, carrying a cell phone and going out atnight in a group.

l Cocoon neighbourhood watch.

Assist natural surveillancel Improved street lighting.l Defensible space design.l Neighbourhood watch and informant hotlines.

Reduce anonymityl Taxi driver IDs.l ‘How’s my driving?’ signs.l School uniforms.

Utilise place managers l Train employees to prevent crime.l Reward vigilance.l Support whistleblowers.

Strengthen formal surveillancel Speed cameras and random breath testing.l CCTV in town centres.l Focused bike patrols in parking lots.

A better way to prevent welfarefrauds

Parliamentary rows are periodically sparkedwhen governments try to crack down onfraudulent welfare claimants by asking thepublic to turn them in. A few years ago, anopposition MP denounced thegovernment’s ‘Beat-a-cheat’ campaign. Hesaid that, ‘Encouraging your next-doorneighbour to snoop on you is the sort ofcommunity values we now expect inBritain.’ Eckhart Kuhlhorn has documenteda better way of reducing welfare fraud,which involved a new computer system inSweden that allowed crosschecking ofstatements made about personal income.When people claimed for rent allowancethey were tempted to understate income,but when they claimed sickness benefits,they were tempted to overstate it. Linkingthe two sets of statements reduced thistemptation and Kuhlhorn found two results:(1) a 75% increase in voluntary statementsmade to the rent allowance authoritiesabout raised incomes; and (2) a 55%reduction in the numbers of householdswho lost rent allowances as a result offalse statements about income. Both weredesirable results and they illustrate thevalue of computerised crosschecking ofdata, not just in the welfare system butelsewhere. It is even possible that haddeath certificates been routinely checkedagainst patient data, Harold Shipmanmight never have embarked on hishomicidal career.

Rational choice theory holds that offenders arealways seeking to benefit themselves by theircrimes. These benefits may not simply be mate-rial as in theft because there are many otherrewards of crime, including sexual release, intox-ication, excitement, revenge, respect from peersand so forth. An important strand of situationalcrime prevention is therefore to understand therewards of any particular category of offendingand to find ways of reducing or removing them.

Conceal targets. Householders often try tofoil burglars by hiding jewellery or other valu-ables. They also keep their curtains drawn tostop thieves from looking through the windowsto see what they own. Some people don’t weargold chains in public, and others avoid leavingtheir cars overnight on the streets if these aremodels attractive to joyriders, such as souped-up Fords and Vauxhalls (see box). These are allways to conceal targets and reduce temptation.Some other concealment strategies are lessobvious. For example, gender-neutral phonelists can help protect women from obscenephone calls and ummarked bullion trucks canreduce the risk of in-transit robbery.

Remove targets. The installation of a machinein a Spanish church that accepted plastic cardsbrought several benefits: donors receivedreceipts for tax purposes, the church receivedlarger gifts and, since money was not deposited,the church reduced its theft risk through remov-ing targets. An earlier application of this samesituational technique comes from the days of theCalifornian Gold Rush. Plagued by robberies ofstage coaches, one mine started casting in 400-pound cubes. These were too heavy for robbersto carry away on horseback. More up-to-dateexamples of target removal are provided bychanges made to public telephones. To stoppeople smashing glass, kiosks in high-risk loca-tions have been replaced by booths. Phonecardsthat dispense with the need to store large sumsof cash have removed an important target fortheft. The package of measures that preventedrepeat victimisation of houses on the Kirkholtcouncil estate included the removal of gas andelectric coin meters which were frequently tar-geted by burglars. Perhaps the most strikingsuccess is the introduction of exact fare systemsand safes on buses, which dramatically reduced

bus robberies in New York and in 18 other citiesin the late 1960s.

Identify property. Motor vehicles in devel-oped countries must be registered and mustcarry a unique Vehicle Identification Number(or VIN). This is to assist taxation, but thesemeasures also reduce theft. One of the last USstates to require vehicle registration was Illinoisin 1934, whereupon vehicle thefts declinedfrom 28,000 in the previous year to about13,000. More recently, the US Motor VehicleTheft Law Enforcement Act 1984 has mandatedthe marking of all major body parts of ‘high-risk’automobiles with VINs. Police ‘operation identi-fication’ programmes have had a chequeredhistory in the United States, but Gloria Laycockfound that property marking undertaken inthree small communities in Wales, combinedwith extensive media publicity, halved thenumber of reported domestic burglaries.

Disrupt markets. Criminologists and policehave devoted remarkably little attention tounderstanding and disrupting markets for stolengoods. Criminologists have found it difficult toobtain data about these markets and the policehave preferred to pay more attention to catchingthieves and burglars than fences, partly becausethe penalties for dealing in stolen goods are rela-tively light. However, if there were no market forstolen goods there would be few persistent bur-glars and few thefts of lorries carrying large loadsof tobacco and alcohol. Recent work for theHome Office by Mike Sutton has awakenedinterest in disrupting markets for stolen goods.The disruptive measures need to be tailored tothe nature of the market and include systematicmonitoring of pawn shop transactions by the police, crackdowns on illegal street vendorsand monitoring of small ad sales in newspapers to detect repeat vendors. Police‘sting’ operations – such as bogus second-handshops – should be avoided because research hasfound these may stimulate theft in the areaaround the sting.

Deny benefits. Installing ‘sleeping policemen’or road humps is a sure way to deny the benefitsof speeding. Security-coded car radios and inktags provide further illustrations of crime preven-tion techniques. Security-coded radios cannot be

35 Reduce the rewards of crime

used unless the thief knows the PIN and, accord-ing to studies undertaken in Australia and theUnited States, cars with these radios have lowertheft rates. Ink tags are used in clothing stores toprevent shoplifting. They release ink if tamperedwith and indelibly the stain garment to whichthey are attached. The thief cannot wear the gar-ment or sell it, which removes the incentive totheft. The remarkable success achieved by theNew York Transit Authority in ridding its subwaycars of graffiti was due to a policy of immediatecleansing, which denied offenders the gratifica-tion of seeing their work on public display.

FIND A PRACTICAL RESPONSE

RE

SP

ON

D

Reduce rewardsConceal targets

l Off-street parking for cars attractive tojoyriders.

l Gender-neutral phone directories.l Unmarked bullion trucks.

Remove targetsl Removable car radios.l Women’s refuges.l Pre-paid phonecards for public phones.

Identify propertyl Property marking.l Vehicle licensing and car parts marking.l Cattle branding.

Disrupt marketsl Checks on pawn shops.l Controls on classified ads.l Licensed street vendors.

Deny benefitsl Ink merchandise tags.l Graffiti cleaning.l Disable stolen mobile phones.

Mike Sutton, Jacqueline Schneider and SarahHetherington (2001). Tackling Stolen Goods with theMarket Reduction Approach. Crime ReductionResearch Series Paper 8. London: Home Office.

Read more

Car thefts and parking place, England and Wales, British Crime SurveyWhere parked Car crimes per 100,000 cars per 24 hours

Garage at home 2

Drive/carport 40

Street outside home 117

Housing estate garage 38

Housing estate car park 101

Garage at work 48

Car park at work 37

Street near work 118

Other street 327

Public car park 454

Source: Ronald Clarke and Pat Mayhew (1998). Preventing Crime in Parking Lots: What We Know and What We Need to Know. InReducing Crime through Real Estate Development and Management, edited by Marcus Felson and Richard Peiser. Washington,DC: Urban Land Institute

Locking the stable door

Nick Ross reminds us in the Foreword thatPOP is about locking the stable door ratherthan, as in conventional policing, chasingafter the horse. We may no longer havehorses but most of us have cars. Many ofus also have garages though we do notalways use them. However, if we havemodels that are at high risk of theft (seethe Home Office’s Car Theft Index), wewould be well advised to conceal them inthe garage. As the table shows leavingthem in the driveway or in the streetoutside the house greatly increases therisks of vandalism and theft.

When studying prisons and pubs, RichardWortley noticed that crowding, discomfort andrude treatment provoked violence in both set-tings. This led him to argue that situationalprevention had focused too exclusively onopportunities for crime and had neglected fea-tures of the situation that precipitate or inducecrime. As a result of his work, Clarke andCornish have included five techniques toreduce what they called ‘provocations’ in theirnew classification of situational techniques.These techniques are explained below, draw-ing on Wortley’s examples.

Reduce frustration and stress. Everyone getsangry when treated rudely by waiters, whenpeople jump queues or when trains are delayedwith no explanation. Sometimes they get soangry they become violent. This could beavoided by improved service, which is increas-ingly being demanded and delivered. However,complaints may be ignored when those mis-treated have little power. For example, prisonersare often ignored when they complain that theycannot eat when hungry or choose their TV pro-grammes, even though these complaints couldbe met quite easily by staggering meal times andproviding more TVs. Queuing for phones,another source of frustration for prisoners, canbe reduced by computerised systems to rationphone use (see box). Outbursts of anger andviolence can also result from people being sub-jected to extreme discomfort – too much noise,being jostled and having nowhere to sit. Theseconditions exist in many clubs and pubs andhave consistently been found to induce trouble.More seating, soothing music and muted light-ing are all ways to reduce stress in these settings.

Avoid disputes. Rival groups of fans are seg-regated in football stadiums and their arrivaland departure has been scheduled to avoidthe periods of waiting around that promotetrouble. Taxi fares from Kennedy airport toManhattan are now fixed at a standard $30 toprevent cheating and disputes over fares. In anattempt to produce consensual crowd man-agement at the Australian Motorcycle GrandPrix, riders were permitted to operate camp-sites for their fellow motorcyclists and wereencouraged to develop rules for use of the

facilities. This helped to eliminate the brawlsbetween police and motorcyclists, which hadmarred the event in previous years.

Reduce emotional arousal. Male doctorsshould not conduct detailed physical examina-tions of female patients without a nurse orreceptionist present. This protects the doctorfrom false accusations, but it also reduces thetemptation to sexually abuse the patient ormake inappropriate advances. Laws that pro-hibit convicted paedophiles from taking jobsinvolving contact with children not only pro-tect children, but also help adults to managetheir sexual desires. That the very sight of agun has been found to trigger feelings ofaggression provides one good reason for guncontrol laws. Similarly, the fact that high pro-portions of sex offenders own or use violentpornography provides a rationale for control-ling these materials. Finally, reducingtemptation is the basis for advice about beingcareful with one’s money in public as well asadvice to young women about being carefulwhen out alone at night.

Neutralise peer pressure. Many parentsdiscourage friends who are a ‘bad influence’on their children and schools dispersegroups of troublemakers into differentclasses. But adults as well as children are sub-ject to peer pressure. Existing staff mayinduct new workers into stealing from theiremployers, and young men are often encour-aged to drink too much by friends. Onepublicity campaign mounted in Australia toreinforce the powerful deterrent impact ofrandom breath testing made use of theslogan, ‘Good mates don’t let mates drinkand drive’.

Discourage imitation. All television sets nowsold in America must contain a ‘V-Chip’ so thatparents can programme their TVs to preventchildren from viewing violent programmes.Though the link between violent movies andviolence in society is much disputed, there issome evidence of ‘copycat’ crimes becausemedia reports of unusual crimes sometimesprovoke imitation elsewhere. It has also beenshown, for example, that students who see

36 Reduce provocations

their teachers engaging in illegal computeractivity are more likely to commit computercrimes themselves, and that other pedestrianswill follow someone crossing against a redlight. Indeed, how often have you ‘run the red’only to find when glancing in your rear viewmirror that so has the car behind you? It hasalso been shown that picnic tables in parksthat had been scratched and carved are morethan twice as likely to attract further damage.Findings such as these provide the rationalefor ‘rapid repair’ programmes to deal with van-dalism. Wilson and Kelling extended thisprinciple in their famous ‘broken windows’article by arguing that the failure to dealpromptly with minor signs of decay in a com-munity, such as panhandling or soliciting byprostitutes, can result in a quickly deteriorat-ing situation as hardened offenders move intothe area to exploit the breakdown in control.

FIND A PRACTICAL RESPONSE

RE

SP

ON

D

Reduce provocationsReduce frustration and stress

l Efficient queues and polite service.l Expanded seating capacity.l Soothing music and muted lighting.

Avoid disputesl Separate enclosures for rival soccer fans.l Reduced crowding in pubs.l Fixed cab fares.

Reduce arousal and temptationl Controls on violent pornography.l Prohibitions on paedophiles working with

children.l Advice about avoiding sexual victimisation.

Neutralise peer pressurel ‘Idiots drink and drive’.l ‘It’s OK to say No’.l Disperse troublemakers at school.

Discourage imitationl Rapid repair of vandalism.l V-chips in TVs.l Avoidance of ‘copycat’ crimes by censoring

details of modus operandi.

Richard Wortley (2001). A Classification of Techniquesfor Controlling Situational Precipitators of Crime.Security Journal, 14: 63–82.

Read more

Phone fraud, slot time and VictoriaSecrets at Rikers Island

Rikers Island, a stone’s throw from LaGuardia airport, is a huge system of 10jails. These house different categories ofinmates, whose phone privileges vary withtheir status. Corrections officers were sup-posed to use logbooks to record phoneuse and to regulate the amount of timeeach inmate spent on the phone. In theearly 1990s, this system had brokendown. Inmates had developed their ownsystem, known as ‘slot time’, and theannual cost of calls had escalated to over$3 million. The most powerful inmatescontrolled the phones, which they oftenused to access their beepers and maintaintheir drug businesses in the outside world.Inmates were also accessing ‘sex lines’and were using stolen credit card numbersto make long distance calls and pur-chases. As Nancy La Vigne notes, whostudied this problem as a graduate stu-dent at Rutgers, ‘The female inmates didjust this, accessing the Victoria Secretscatalogue, which resulted in a jail thatcould boast the best-dressed inmates inthe country – until officials caught on.’

The officials introduced a high security com-puterized phone system that put strict limitson phone use, in line with the status of thecaller. Detainees gained access to thephones through bar codes on their ID cardsand by entering a PIN. This system immedi-ately cut phone costs in half, but it was alsonoticed that fewer fights were erupting overaccess to the phones. In fact, Nancy LaVigne’s study showed that the monthly rate ofthese fights dropped from 6.7 per 1,000inmates in the year before the new phonesystem to 3.6 per 1,000 after its introduction.

We don’t know whether the female inmatesremained the best dressed in the country.

Source: Nancy La Vigne (1994). Rational Choice andInmate Disputes over Phone Use on Rikers Island. InCrime Prevention Studies, vol. 3, edited by RonaldClarke. Monsey, New York: Criminal Justice Press.

This fifth category of situational techniquesrecognises that offenders make moral judg-ments about their behaviour and that theyoften rationalise their conduct to ‘neutralise’what would otherwise be incapacitating feel-ings of guilt or shame. They make such excusesas: ‘He deserved it,’ ‘I was just borrowing it’and ‘I only slapped her’. These excuses may beespecially important for ordinary peopleresponding to everyday temptations to evadetaxes, drive when drunk, sexually harass junioremployees and steal employers’ property.

Set rules. All organisations make rules aboutconduct in their fields of governance. For exam-ple, businesses regulate employees’ timekeeping and shops require sales assistants tofollow strict cash-handling procedures.Organisations such as hospitals, transport sys-tems and hotels must, in addition, regulate theconduct of the clients they serve. Any ambiguityin these regulations will be exploited if it bene-fits the client. One important strand ofsituational prevention, therefore, is rule setting– the introduction of new rules or procedures(and the clarification of those already in place)to remove any ambiguity concerning theacceptability of conduct. For example, inattempting to reduce ‘no-shows’, many restau-rants will now only accept reservations if callersleave a telephone number where they can bereached. Some are also requiring that reserva-tions be accompanied by a credit card numberso that a charge can then be made for no-shows. Requiring anglers in California to weartheir licences was successful in getting more ofthem to comply with licence purchase rules.

Post instructions. Work rules are often set outin employment contracts and rules establishedby credit card companies, telephone providersand insurance companies are contained in theservice contracts. Regulations governing publicplaces or facilities may be publicly posted, eitherto prevent people claiming ignorance of therules or to show precisely where these apply.The roads, in particular, make extensive use ofsigns governing driving or parking. Studies havefound that warning signs significantly reduce ille-gal parking in bays reserved for disabled drivers.Many other facilities – parks, libraries, colleges,transport systems, council estates – also post

signs to govern a wide range of behaviours.Despite their wide use, there have been fewevaluations of the preventive effectiveness ofposted instructions – but they are an essentialtool of law enforcement and are often used inproblem-solving efforts.

Alert conscience. This situational techniquediffers from ‘informal social control’ in twoimportant respects. First, the focus is on spe-cific forms of crime occurring in discrete, highlylimited settings and, second, the purpose is toalert conscience at the point of committing aspecific kind of offence rather than attemptingto bring about lasting changes in generalisedattitudes to law breaking. For example, signs atstore entrances announce that ‘Shoplifting isstealing’ and in the Manhattan Bus Terminalsigns proclaim that ‘Smoking here is illegal, self-ish and rude’. Roadside speed boards arewidely used to give immediate feedback (with-out issuing fines) to motorists travelling abovethe speed limit. Finally, government televisioncampaigns that accompany crackdowns on TVlicence evasion show those detected beingtreated by the courts as ‘common criminals’.There is no published evidence on the successof these campaigns but they have beenrepeated for more than three decades.

Assist compliance. When Lombroso suggestedin the 19th century that people should be lockedup for urinating in the streets, his pupil Ferrisuggested a more practical way to solve theproblem: build public loos. This constitutes anexample of facilitating compliance, a techniqueof wide application. It includes subsidised taxirides for those who have been drinking, litterbins and ‘graffiti boards’ (the latter of which aresupplied for people’s public messages), andimproved checkout procedures in libraries,which reduce delay and thus excuses for failingto comply with rules for book borrowing. In aclassic little paper on Disney World, Shearingand Stenning provide a fascinating glimpse intothe ways in which sophisticated crowd controland management – involving the use of pave-ment markings, signs, physical barriers (whichmake it difficult to take a wrong turn) andinstructions from cheerful Disney employees –greatly reduce the potential for crime and incivil-ity in the theme park (see box).

37 Remove excuses for crime

Control drugs and alcohol. Crime is facilitatedby alcohol and drugs, which undermine inhibi-tions, or impair perception and cognition so thatoffenders are less aware of breaking the law. Thevalue of situational controls on drinking has oftenbeen demonstrated. Johannes Knutsson hasshown that limiting the amount of alcohol thatindividuals could bring into a Swedish resort townon Midsummer Eve helped to reduce drunken-ness and disorderly conduct. A local ordinancebanning alcohol consumption in public in thecentre of Coventry achieved large reductions incomplaints of insulting behaviour. Voluntaryagreements reached among local drinking estab-lishments to promote responsible drinking havereduced alcohol-related crime in numerousnightlife areas in Australia. Rutgers University hasdecreed that beer must be served from kegs notcases at dorm parties because cases are easier tohide and, as one student said: ‘If you have one kegand a line of 20 people behind it, people will getless alcohol than if you had a refrigerator andpeople were throwing out beer.’

FIND A PRACTICAL RESPONSE

RE

SP

ON

D

Remove excusesSet rules

l Rental agreements. l Harassment codes.l Hotel registration.

Post instructionsl ‘No Parking’.l ‘Private Property’.l ‘Extinguish camp fires’.

Alert consciencel Roadside speed display boards.l Signatures for customs declarations.l ‘Shoplifting is stealing’.

Assist compliancel Easy library checkout.l Public lavatories.l Litter bins.

Control drugs and alcoholl Breathalysers in pubs.l Server intervention training.l Alcohol-free events.

Arriving at Disney World

l Signs tell visitors arriving by car to tuneinto Disney radio for information

l Signs direct them to the parking lot theymust use and road markings show the way

l Smiling parking attendants direct visitorsto their space and loudspeakers remindthem to lock their cars

l They are directed to rubber-wheeledtrains to take them to the monorail

l Recorded announcements direct them tostand safely behind guardrails

l They are reminded about the location oftheir parking space (e.g. Donald Duck 1)

l They are (politely) asked to sit, to keeptheir arms and legs within the confines ofthe carriage, and to make sure childrendo the same

l Before disembarking, they are told how toget to the monorail and barriers stopthem going the wrong way

l On the platform, attendants guide theminto corrals the right size to fill one com-partment of the monorail

l Safety gates at the platform edge openonly when the monorail arrives

l Any delays in service are announced andexpected times of arrival are given

l On board, passengers are asked toremain seated ‘for their own safety’

l Passengers are told how to disembark andhow to move to the first entertainment

l They are once again reminded to lookafter their children and to take their pos-sessions

l While waiting to enter each exhibit, visi-tors are marshalled in queues, whichindicate waiting times; those in line areentertained by Disney characters

l On leaving the exhibit, they are guided bysigns, barriers and attendants to the nextone

Source: Clifford Shearing and Phillip Stenning (1997). From the Panopticon to Disney World: TheDevelopment of Discipline. In Situational CrimePrevention: Successful Case Studies, edited by Ronald V.Clarke. Monsey, New York: Criminal Justice Press.

Many problems arise through the failure ofsome institution – business, governmentagency, or other organisation – to conduct itsbusiness in a way that prevents crime ratherthan causing it. In short, many problems occurbecause one or more institutions are unable orunwilling to undertake a preventive strategy,or because these institutions have intention-ally established a circumstance that stimulatescrimes or disorder. This creates risky facilities(Step 26) and other concentrations of crime.

Solving problems usually requires the activecooperation of the people and institutions thathave failed to take into account the conditionsthat lead to the problem. These people andinstitutions have shifted the ownership of theproblem from their shoulders to the shouldersof the police. Consequently, an importantobjective of any problem-solving process is toget them to assume ownership. So in develop-ing a response, you need to answer threequestions regarding ownership:

l Who should have ownership of the problem?

l Why has the owner allowed the problem todevelop?

l What is required to get the owner to under-take prevention?

Who should have ownership of the prob-lem? When a problem is located at a specificplace, it is usually easy to identify who isresponsible. The owner of the problem is theowner of the location. A problem in a park, forexample, is the body with the responsibility foroperating the park – usually a local authorityor some private agency.

It is more difficult to identify those responsiblefor problems that are spread over larger areas.If a widespread problem is focused on a specificlocation, then that location may be the sourcefor the events in the surrounding area, and theowner of the central location may be responsi-ble. A Lancashire scrap yard functioning as acriminal receiver, for example, facilitated prop-erty crime in the surrounding village.

If there are agencies charged with seeing tothe well-being of a special group of individuals– the elderly, children with special needs, orvictims of domestic violence – and these indi-viduals are targets of crime or disorder, thenthe service agency is a candidate for an ownerof the problem.

Why has the owner allowed the problemto develop? There are four generic explana-tions that alone or in combination fit mostproblems:

1. An institution may be unable to preventcrime. This might be due to ignorance asto the effect of its operations on crime orignorance as to how to prevent crime. Orthis may be due to lack of resources, evenwhen the institution knows its operationshelp create crime. It is also important torecognise the importance of institution-alised procedures. Changing procedurescan be time consuming and costly in bothmonetary resources and staffing. A newinventory control procedure to preventshoplifting and internal theft may be diffi-cult to implement because it requiresdisruptive changes in the ways employeesconduct their normal business.

2. Some institutions may be unwilling to pre-vent crime facilitated by their operations.Rather than recognise the role of opportu-nity in creating crime, some people dwellexclusively on the role of offenders. Fromthis perspective, it is the function of policeto reduce crime by stricter enforcement. Thelimitations of this approach have been notedin Step 4. Another source of unwillingness isthe belief that the police are intruding ontheir rights. A retailer might claim that hehas the right to display goods any way he wants, and that the police should notcompel, or even suggest, alternative displaysthat might reduce shoplifting.

3. Some institutions are unwilling becausethey gain more by ignoring crime thanthey lose. They may perceive that the costsof prevention outweigh any benefits to

38 Find the owner of the problem

them. Security personnel at an entertain-ment venue are costly, and quality securitypersonnel are more costly. If the costs ofthe problem are not borne by the facility,then there is little perceived need to bearthe costs of prevention. In essence, suchfacilities are exporting the costs of crimeand prevention onto others, and therebyreducing their costs.

4. Some institutions may profit from thecrimes, as can happen when a second-handgoods shop does little to verify legitimateownership of the merchandise they display.Garages can purchase stolen car partscheaper than legal car parts, thus increas-ing their profit margin.

What is required to get the owner toundertake prevention? Herman Goldsteinhas outlined a rough hierarchy of interven-tions designed to shift ownership of problemsfrom the police back to the institutions thatare responsible for them. These are shown inthe figure.

Moving from the bottom to the top of the list,interventions become less cooperative andincreasingly coercive. Because of this, as onemoves up the list, the difficulty of interventionincreases, along with the costs of failure to thepolice. Consequently, the value of informationand thorough analysis increases from thebottom to the top. As Goldstein notes, this hier-archy is a rough indicator of these trends, ratherthan an exact description. Nevertheless, it isuseful for planning a layered set of responses toa problem – beginning with the most coopera-tive and working upward only if needed andonly when supported by information.

FIND A PRACTICAL RESPONSE

RE

SP

ON

D

Herman Goldstein (1997). The Pattern of Emerging Tactics for Shifting the Ownership of Prevention Strategies in theCurrent Wave of Change in Policing: Their Implications for Both Environmental Criminology and the Police. Paperpresented at the 6th International Seminar on Environmental Criminology and Crime Analysis. Oslo, Norway.(Download from: www popcenter.org)

Read more

Goldstein’s hierarchy of ways to shift ownershipBringing of a civil action

Legislation mandating adoption of prevention

Charging a fee for police service

Withdrawing police service

Public shaming

Pressing for the creation of a new organisation to assume ownership

Engaging another existing organisation

Targeted confrontational requests

Straightforward informal requests

Educational programmesLess

coo

pera

tive

→In

crea

sing

ly d

iffic

ult

→

It can be a long and troublesome process find-ing a suitable response. You may be repeatedlydisappointed when promising interventionsare vetoed because of expense or difficulty, orbecause a crucial partner won’t cooperate. Butthere is never just one way to solve a problemand it is always possible to find an acceptableresponse among the available range of situa-tional measures. So, eventually your team willagree on a response that meets some basicrequirements:

l It is not too ambitious or costly.

l It focuses on near, direct causes rather thanon distant, more indirect ones, which givesit a good chance of making an immediateimpact.

l The mechanism through which each meas-ure should impact the problem has beenclearly articulated.

So, at last your worries are over and you canrelax, right? Wrong! Even more difficult thanagreeing on a promising response is to makesure that, once selected, it is actually imple-mented. As an analyst, you cannot ensure thison your own, but if you know the pitfalls ofimplementation, you can steer the partnershipaway from choosing responses that can fallprey to these. Tim Hope and Dan Murphyidentified these pitfalls when studying a HomeOffice vandalism prevention project in elevenschools in Manchester.

The measures to be implemented at eachschool were selected by local council officials,school staff and police. Much of the damagewas more inadvertent than malicious. Forexample, many windows had been broken byfootballs and much of the damage to the out-side of buildings had been caused by childrenclimbing about. This suggested two differentsolutions: situational measures to protect thebuildings or leisure provision to divert childreninto less harmful activities. Only one of theeleven groups recommended improved leisureprovision. The situational measures recom-mended were mostly basic target-hardening

(window grills, toughened glass and highfences), though proposals also included ascheme to encourage local residents to keepan eye on two adjacent schools and a plan tomove a playground to a less vulnerable area.

At only two schools were all the recommenda-tions implemented. In three, none was put inplace and at the remaining six schools one ormore measures failed to materialise. The fail-ure to implement measures meant there waslittle impact on vandalism. Hope and Murphyidentified five main obstacles to implementa-tion, all of which have since been encounteredin other problem-solving projects:

1. Unanticipated technical difficulties. Foreight schools, the groups recommendedthe replacement of vulnerable windowswith polycarbonate glazing or toughenedglass. However, not a single pane of eithertype was installed. Polycarbonate glazinghad been prohibited by the city architectsbecause it would prevent escape in case offire and might give off toxic fumes.Toughened glass had to be cut to sizebefore it was toughened, but the panescame in many sizes and it would have beendifficult to store a few of each size in readi-ness. The alternative of supplying a pane toorder was ruled out by the long time (sixweeks) it would take to do this.

2. Inadequate supervision of implemen-tation. At one school it was agreed to movethe playground to a less vulnerable area.The original playground was to be replacedwith flower-beds, but this had to be done bya council department that got no furtherthan providing an estimate for the work.The relocation of the playground was sub-contracted to a private builder, but due to amisunderstanding only half the proposedarea was resurfaced. After two years, there-fore, vandalism was unchanged, there wereno flower-beds, and the school hadacquired a useless, narrow strip of tarmac.

3. Failure to coordinate action among dif-ferent agencies. Every recommendation

39 Choose responses likely to beimplemented

that was the sole responsibility of the build-ings branch of the education departmentwas implemented, while none of thoseinvolving other departments or agenciesever materialised. For example, it was rec-ommended that people living near twoschools should be encouraged to keep aneye on them after hours and report any-thing suspicious to the police. The schemerequired the cooperation of the educationdepartment, the head teacher, staff andpupils of the schools and three branches ofthe police – crime prevention, communityrelations and local police. All seemed keenon the idea, but no one would take the lead.

4. Competing priorities. During the imple-mentation period many other demandswere placed on the education departmentas a result of widespread industrial actionby council employees and a substantialreorganisation made necessary by a declin-ing school-age population. The departmentwas also undergoing staff changes thatresulted in the reassignment of staff whohad been involved in the vandalism project.It is therefore not surprising that the HomeOffice vandalism project was given little pri-ority, especially when there was no obviousadvantage for officials to spend time on it.

5. Unanticipated costs. In some cases, thewider consequences of a particular courseof action outweighed its immediate bene-fits. For example, at the school with themost serious vandalism problem it wasdecided to mount a security patrol for theimpending holiday period. The school care-takers were employed to patrol the schoolfor payment during their spare time. Thismeasure was immediately successful inreducing vandalism and was extendedbeyond school holidays to provide cover-age at evenings and weekends. Otherschools began to demand the same protec-tion and more caretakers wanted theadditional overtime opportunities.Ultimately the cost became too high andthe scheme was scrapped.

You can see from the list above that someimplementation problems cannot be antici-pated and that a proportion of all measuresselected will never be implemented. However,it is also clear that certain kinds of measurescan be expected to encounter problems andthese are summarised in the box. In somecases, of course, a measure may be so promis-ing that it is worth pursuing despite the risksof implementation failure. But being fore-warned is to be forearmed.

Expect implementation problems when ameasure…..

…requires coordinated action among anumber of separate agencies

…will take a long time to introduce andinvolves a number of steps to be completed insequence

…must be implemented by staff with littleunderstanding of its purpose

…has no major supporter among thepartnership team

…lacks the support of senior administrators

Also expect problems when the measuremust be implemented by an agency…

…that is outside the partnership

…that is poorly resourced or in turmoil

…that will gain little direct benefit from thesolution

FIND A PRACTICAL RESPONSE

RE

SP

ON

D

Tim Hope and Daniel Murphy (1983). Problems ofImplementing Crime Prevention: The Experience of aDemonstration Project. The Howard Journal, XXII,38–50.

Gloria Laycock and Nick Tilley (1995). ImplementingCrime Prevention. In Building a Safer Society, editedby Michael Tonry and David Farrington. University ofChicago Press.

Read more

The first critical step in assessment is to con-duct a process evaluation. It answers thequestion, ‘Was the intervention put into placeas planned and what alterations were requiredfor implementation?’ A process evaluationdoes not tell you whether the response waseffective. For that you need an impact evalua-tion. An impact evaluation tells you whetherthe problem changed. Steps 41 to 48 discussvarious aspects of impact evaluations.

Both types of evaluations are needed in a POPproject. The table summarizes possible con-clusions based on the findings of both types ofevaluation. A) The response was implementedin accordance with the plans, and the declinein the problem was probably due to theresponse. So there is credible evidence thatthe response caused the reduction. B) Theresponse was implemented as planned, butthere was no reduction in the problem. Sothere is credible evidence that the responsewas ineffective.

But what if the response was not implementedas planned? In this case, it is hard to come to auseful conclusion. C) If the problem declined,it might mean that the response was acciden-tally effective or some other factor wasresponsible. D) If the problem did not decline,then no useful conclusion is possible. Perhapsthe implemented response is faulty and theoriginal response would have been effective,or neither is effective. So unless we can besure the response was implemented asplanned, it is hard to learn from an impactevaluation.

It is useful to think of the response as a com-plex piece of machinery with a variety ofcomponents, any of which can go wrong (Step39). A process evaluation requires examiningthe degree to which components of theresponse were carried out. The process evalu-ation checklist lists the questions that youshould ask. For each item, you should firstestablish whether the component is critical forthe success of the intervention, and finallywhether the component operated as planned.

Few responses are implemented exactly asplanned. Though unexpected developmentscan force you to modify a response, some ofthese developments can be anticipated. To seehow, we need to answer the question, whatgoes wrong with responses? Some of the pos-sible answers are as follows:

1. You may have an inadequate under-standing of the problem. You may havefocused too little on repeat victims, forexample. This can be caused by invalidassumptions about the problem or insuffi-cient analysis (you did not look for repeatvictimisation, for example). If while devel-oping the response you can identify weakspots in your analysis, then you can createcontingency plans (a plan to address repeatvictimisation should this prove to beneeded).

2. Components of the project havefailed. The process evaluation checklistshows that there are many potential pointsof failure. However, not all components are

40 Conduct a process evaluation

Interpreting results of process and impact evaluationsProcess evaluation results

Response implemented Response not implementedas planned as planned

Impact Problem declined & A. Evidence that the response C. Suggests that the responseevaluation no other likely cause. caused the decline. was accidentally effective or results that other factors may have

caused the decline.

Problem did not decline. B. Evidence that the D. Little is learned. response was ineffective.

equally important for success. Further, it issometimes possible to anticipate compo-nents with high failure rates. Citizen groupsin general are quite variable in their abilityto carry out tasks, for example. Creatingredundant capabilities or an emergencybackup plan can alleviate difficulties cre-ated by component failure.

3. Offenders may negatively adapt to yourresponse (Step 12). Some forms of nega-tive adaptation can be anticipated andplanned for. Sometimes geographical dis-placement locations can be identified beforethe response, for example, and advancedprotective actions can be taken to immunisethem.

4. There are unexpected externalchanges that impact the response. Apartner agency’s budget may be unexpect-edly cut, for example, forcing it to curtail itsefforts on the problem. As the problem willnot dissipate on its own, the only recourseis to alter the plans.

In short, it is possible to plan for some unex-pected developments, but not all. Part of theprocess evaluation should include examiningthe adequacy of planning for unexpecteddevelopments and the effect of truly unex-pected changes on the response. A completeprocess evaluation will help others learn fromyour experiences.

Process evaluation checklistComponents of response Is component critical? Did it operate as planned?

Who is supposed to act?Police unitsGovernment partnersCommunity groupsBusinessesOthers

Do they have the capability to act?Legal authorityLocal authorityResourcesExpertise

When were they supposed to act?Dates and timesCoordination with others

Who or what is supposed to receive action?PeoplePlaces

Was action delivered appropriately? IntensityDuration

Are there back up plans?Faulty plansAdaptationExternal changes

ASSESS THE IMPACT

AS

SE

SS

You should address three questions whenassessing a response:

1. ‘What was the implemented response?’ Aprocess evaluation answers this question(Step 40).

2. ‘Did the problem decline?’ Comparing thelevel of the problem after the response tothe level of the problem before answersthis question.

3. If the response declined, then the thirdquestion is, ‘Did the response cause thisdecline, or was it something else?’ Thereare usually many alternative explanationsfor the decline in the problem.

To answer the third question requires the useof controls. The purpose of controls is to ruleout alternative explanations. Different alterna-tive explanations require different types ofcontrols as described below:

l Changes in size. If there had been adecline in tenants in an apartment buildingwe would expect the number of burglariesto decline because there are fewer potentialvictims. A reduction in targets is an alterna-tive to the explanation that the drop wasdue to the intervention. To control forchanges in size you divide the number ofburglaries before and after the response bythe number of occupied flats before andafter the response. In the table, whatappears to have been a decline in burglarieswas partially caused by the reduced numberof occupied units. When this is accountedfor, we still notice a decline of 21 burglariesfor every 100 units. If the time periodsbefore and after are substantially different inlength, we control for this by dividing againby the number of months before and thenumber of months after to get burglariesper occupied unit per month.

l Cycles of activity. Human activity oscil-lates over days, weeks and months. Some ofthe most common cycles include commut-ing and attendance at school, work andleisure, and seasons and holidays. Such

cycles can cause oscillations in problems. Tocontrol for cycles compare the same part ofthe cycle before the response to the samepart of the cycle after the response. Youshould compare the number of thefts fromvehicles in July, August, September andOctober before the response, for example,to the number of thefts from vehicles inJuly, August, September and October afterthe response. With the cyclical theft patternshown in the figure, comparing the monthsmarked by dots controls for the cycles.

l Trends in the problem. Problems may begetting worse (or better) before theresponse. Without accounting for trendsone could conclude the response wasresponsible, when in fact this might haveoccurred anyway. There are two ways ofcontrolling for trend. The first method is tomeasure the problem for a long periodbefore the response so any trend can beidentified (the trend is shown by a dashedline in the figure). Substantial deviationsfrom the existing trend after the responseare evidence of a response effect (there isno evidence of this in the figure).Anticipatory effects (Step 46) are drops inthe problem due to the response, butoccurring before the response is fullyimplemented. These need to be distin-guished from longer trends prior toimplementation.

The second method is to identify a group ofpeople with the same problem (but not get-ting a response) or similar areas with theproblem (but not getting the response).These are control groups. A control groupmust be similar to the response group with

41 Know how to use controls

Using rates to control for changes in size

6 months 6 months Changebefore after

Burglaries 41 20 –21

Occupied units 83 73 –10

Burglaries/occupied unit 0.494 0.286 –0.208

regard to the problem, but cannot receivethe response. A control group tells youwhat would have happened to the responsegroup, if the response group had notreceived the intervention. If the responsegroup changes differently, this is evidence ofa response effect.

l Other unexpected events. Many thingsare changing as the response is imple-mented, one or more of which could havecaused the decline in the problem. Insteadof an accident reduction response causing adecline in accidents on a road, for example,slow traffic from the road repairs, going onabout the same time as the response mighthave been the cause. The standardapproach is to use a control group. Thecontrol area is only effective if it is influ-enced by the same influences as theresponse area. So, counting accidents onthe response and control roads, both ofwhich are influenced by the road repairs,could tell you whether the road repairscontributed to the decline in accidents.

l Change in problem measurement. Abefore–after comparison is only valid if theproblem was measured in the same waybefore and after the response. Measurementdifferences could cause a perceived change

in the problem. Use the same standardisedand systematically applied measurementprocedures before and after. Before andafter observations should occur at the samesites at the same times, watch the samethings and record them in the same way.Before and after photographs and videosshould be taken in the same light from thesame angles, at the same distance and withthe same size image area. Interviewersshould be the same before and after, askingthe same questions, in the same order.Examine official records to determine ifrecording practices are stable over time.

l Natural decline from an extreme high.Many problems are addressed becausethings have become extremely bad. So theresponse is implemented when the prob-lem is abnormally high. But even arelatively stable problem will fluctuate. Aproblem that is now abnormally bad willreturn to its normal level, even if nothing isdone (this is called ‘regression to themean’, and it applies to abnormal lows aswell). Comparison area controls may not beuseful here, if the response area is at a peakand the control area is not. Instead, exam-ine the long-term fluctuation prior to theresponse (see trends, above) to see if theproblem was abnormally bad.

ASSESS THE IMPACT

AS

SE

SS

Months1

20

15

10

5

0

Thef

ts fro

m v

ehic

les

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

25

30

A cyclical pattern with a downward trend

Geographical displacement occurs when theintervention blocks crime or disorder oppor-tunities at a facility or in an area, and offendersmove to other facilities or areas to offend.Temporal displacement also stems from suc-cessful prevention, but in this case offendersshift offending in time.

Offenders can find it difficult to move to someother location because easy crime or disorderopportunities are limited (Step 17). Targets maybe concentrated at some places and not others(Step 17). Vulnerable potential victims can befound at some locations, but not others (Step27). Some facilities have low behavioural con-trols, but others do not (Step 26). Opportunitiesthat exist are either already hot spots or arehidden from offenders – either far away or notrecognisable as fruitful places to offend.

Offenders will not usually spend time searchingfar from their hot spot when it is suppressed.So, if offenders move they are most likely tomove close to the original hot spot. The like-lihood that offenders will move to anopportunity declines the further the opportu-nity is from the original hot spot, as illustratedin the figure. Also, not all spaces are suitable to

offenders. Opportunities are not spread evenlyacross the map. In this map, the diamonds areplaces with characteristics like the original hotspot. Those closest to the original location willbe affected most by displacement. Knowingthis, displacement countermeasures can beapplied with the response.

If geographical displacement occurs, it can dis-tort your findings of prevention effectiveness.Table 1 illustrates how this can happen. Herethere are three similar areas with equal numbersof crimes before treatment: (1) a treatment area;(2) an area adjacent to the treatment; and (3) anarea distant from the treatment. The treatedarea has a decline of 25 crimes. However, theadjacent area has a 10-crime increase, indicatingthat if nothing had been done in the treatmentarea it too would have experienced 110 crimes.So the net reduction is 35 (the estimated neteffect is the treatment difference minus thecomparison difference, –25 – 10= –35).

But these extra 10 crimes could have been dueto geographical displacement. One would bebetter off using the distant control area forcomparison. As a control, the distant area sug-gests that if no treatment were implemented,crime would not have changed in either thetreatment or the adjacent area. The implicationis that the treatment caused a 25-crime declinein the treatment area, but a 10-crime increase inthe adjacent area (displacement), for a com-bined reduction of 15 crimes. Though effective,the programme is not as effective as originallyestimated. Step 45 describes formulas to calcu-late effectiveness when geographicaldisplacement or diffusion are present.

Try to select two comparison areas as part ofevaluations: one near the treatment area todetect geographical displacement (and diffu-sion–Step 45), and the other to serve as acontrol area. The control area should be pro-tected from displacement contamination bydistance or some other barrier (e.g. a motor-way or river). Valid selection of control anddisplacement areas requires you to have someidea of offenders’ normal movement patterns,

42 Consider geographical and temporaldisplacement

N

half a mile

Decreasing chances of spatial displacement

Decline of geographical displacement with distance from hot spot epicenter

as the control area needs to be outside theirroaming territory while the displacement areashould be within it.

Temporal displacement may be easier foroffenders than geographical displacementbecause it requires less effort. Temporal dis-placement can occur within a 24-hour day, if,for example the prevention is restricted to cer-tain times but leaves other times unprotected.It can also occur over a week. Or it can occurover longer periods.

If the evaluation compares times with preven-tion to times without prevention, contaminationof temporal controls can take place. In Table 2 atreatment takes place on Saturday and Sunday.The average number of crimes on these daysdropped by 25 crimes after treatment, whilecrime on Mondays and Fridays increased by 10.Was this due to temporal displacement?Midweek days may be more valid controlsbecause they have less in common with week-ends than do Mondays and Fridays.

Waiting out the prevention is a common formof temporal displacement. Enforcement crack-downs are particularly vulnerable to this formof time shifting because they are temporary bydefinition. If an intervention can be main-tained (unlike a crackdown), then offenderscannot wait it out. They then face the difficultoption of moving to less attractive places ortargets or undertaking new tactics or othercrimes. If these options are too difficult, unre-warding, risky or otherwise unattractive theymay curtail some offending.

ASSESS THE IMPACT

AS

SE

SS

Table 1: Use of adjacent and distant control area in controlling for geographical displacement

Before After Difference

Treatment 100 75 –25

Adjacent area 100 110 +10

Distant area 100 100 0

Table 2: Using days of the week to control for temporal displacement Day of week Before After Difference

Treatment days Sat & Sun 100 75 –25

Adjacent days Mon & Fri 100 110 +10

Distant days Tues-Thurs 100 100 0

John Eck (2002). Assessing Responses to Problems:An Introductory Guide for Police Problem-Solvers.Washington, DC: Office of Community OrientedPolicing Services, US Department of Justice.(Download from: www.popcenter.org)

Lawrence Sherman (1990). Police Crackdowns: Initialand Residual Deterrence. In Crime and Justice: AReview of Research, edited by Michael Tonry andNorval Morris, vol. 12. Chicago: University ofChicago Press.

Read more

In addition to geographical and temporal dis-placement, offenders can switch targets,change their tactics or change crimes. Theseforms of displacement will be more difficultfor you to detect than geographical or tempo-ral displacement.

Target displacement involves offenders shift-ing from newly protected targets to othertargets. In 1970, when steering column locksbecame required in all new cars sold in Britain,thefts of new cars dropped from 20.9% of allcars stolen in 1969, to 5.1% in 1973. However,the overall theft rate of automobiles stayedroughly constant because offenders switchedfrom the newer, protected, vehicles to older,unprotected, vehicles. This is one of the fewdocumented cases where displacement wipedout most prevention gains, at least in the shortrun. Over a longer period these devicesappeared to have curbed theft for temporaryuse. As this case illustrates, it is easy for offend-ers to switch to similar targets. Targetdisplacement is less likely when the alternativetargets are unlike old targets.

Step 42 shows how geographical or temporaldisplacement can contaminate control groups.If the evaluation of a prevention effort uses atarget control group, then a similar form ofcontamination can take place. Imagine aresponse to curb theft of handbags from olderwomen (over 60) in a shopping district. Toestimate what the trend in elderly handbagtheft would be if nothing had been done, thetheft of handbags from middle-aged women of45 to 59 is measured. If, unknown to us, thethieves displaced from the protected olderwomen to unprotected middle-aged women,we would conclude that handbag theft wouldhave gone up without a response. When wecompare this control target group change tothe treatment group change we would mistak-enly inflate the treatment effectiveness. Abetter control group might be even youngerwomen shoppers (ages 30 to 44, for example),or even better, wallet theft of male shoppers.Though neither of these alternatives is perfect,they are improvements because one would

expect far less displacement to dissimilar tar-gets. (Or select another shopping district as acontrol area. In this case you would have toguard against geographical diffusion or dis-placement contamination – Steps 42 and 45.)

Tactical displacement occurs when offenderschange their tactics or procedures. They mightuse different tools to defeat better locks, forexample. Or computer hackers might alter theirprogrammes to circumvent improved security.In medicine, some bacteria can mutate quicklyso a drug that is effective against one form ofthe bacteria becomes less effective as mutantstrains become more prevalent. One way of countering this is to use broad-spectrumtreatments that are effective against a widerange of mutations. Similarly, ‘broad spectrum’responses protect against existing methodsused by offenders and many modifications ofthese tactics. Broad-spectrum interventionsrequire offenders to make big changes in theirbehaviour which they may not be able to do.Paul Ekblom describes attempted tactical dis-placement following the installation of barriersin sub-post offices to prevent ‘over the counter’robberies; some offenders tried using sledge-hammers. This change in tactics was notparticularly successful, however, and displace-ment was limited. These barriers are anexample of a broad-spectrum intervention asthey were able to defeat new tactics.

We seldom select treatment or control groupsbased on tactics, but we sometimes use crimetypes as controls. For example, we might selecttheft from vehicles as a control in the evaluationof a theft of vehicles intervention. The sameprinciples of contamination and protectionapply here as we found with other displacementmethods. If the tactic or crime type is very simi-lar to the tactic or crime type being addressed,then displacement could contaminate thesecontrols. Dissimilar tactics or crime types areless likely to suffer contamination. But if theyare too dissimilar it is not a useful control.

There is no perfect solution to this problemand compromises must be struck. The conse-quence is that it is often difficult to know if

43 Examine displacement to othertargets, tactics and crime types

displacement is occurring and difficult tojudge the effectiveness of the intervention.Compounding these difficulties is that mul-tiple forms of displacement can occur simulta-neously. Indeed, sometimes one form ofdisplacement will necessitate another form aswell. Some target displacement may require achange in tactics, and if the new targets arenot in the same places as the old targets, geo-graphical displacement will occur too.

You cannot find displacement unless you lookfor it. This means that you should examine aproblem closely and hypothesise the most likelyforms of displacement. Are there other oppor-tunities for crime or disorder that are similar tothe opportunities your efforts are trying toblock? Will your offenders easily discover theseopportunities? Looking for displacement oppor-tunities prior to finalising a response gives youtwo advantages. First, you can develop meas-ures for detecting it should it appear. Moreimportant, you may be able to develop counter-measures that prevent displacement.

A study of target displacement:helmet laws and the reduction inmotorcycle theft

In Germany (as elsewhere) the enactment ofhelmet laws was followed by large reductions inthefts of motorcycles. After the laws were broughtinto place in 1980, offenders wanting to steal amotorbike had to go equipped with a helmet orthey would be spotted quickly. The first column ofthe table shows that by 1986 thefts of motorbikeshad dropped to about one-third of their level in1980, from about 150,000 to about 50,000. (Thegradual decline probably reflects strongerenforcement and growing knowledge about therequirement.) This fact suggests that motorbiketheft has a much larger opportunistic componentthan anyone would have thought. The existence ofexcellent theft data in Germany allowedresearchers to investigate whether the drop inmotorcycle theft had resulted in targetdisplacement to theft of cars or bikes, other formsof personal transportation.

The second and third columns of the table showthe national totals for car and bike thefts duringthe same years. These provide some limited

evidence of displacement in that thefts of carsincreased by nearly 10% between 1980 and1986, from about 64,000 to 70,000. Thefts ofbicycles also increased between 1980 and 1983,but by the end of the period had declined again toa level below that for 1980. Altogether, it is clearthat at best only a small proportion of the100,000 motorbike thefts saved by the helmetlaws were displaced to thefts of other vehicles.

A little thought shows why this may not besurprising. Motorbikes may be particularlyattractive to steal. They are much more fun to ridethan bikes for the young men who comprise mostof the thieves. Even if the intention is merely to gethome late at night, a motorbike offers significantadvantages, especially if the distance is more thana few miles. Motorbikes may also be easier tosteal than cars since the latter have to be brokeninto before they can be started. Like bikes, carsalso offer less excitement than motorcycles andthey may require more knowledge to operate.

Source: Pat Mayhew and colleagues (1989). Motorcycle Theft,Helmet Legislation and Displacement. Howard Journal ofCriminal Justice, 28: 1–8.

ASSESS THE IMPACT

AS

SE

SS

Paul Ekblom (1987). Preventing Robberies at Sub-Post Offices: An Evaluation of a Security Initiative.Crime Prevention Unit Paper 9. London: Home Office.

Barry Webb (1994). Steering Column Locks and MotorVehicle Theft: Evaluations from Three Countries. InCrime Prevention Studies, vol. 2, edited by RonaldClarke. Monsey, New York: Criminal Justice Press.(Download from: www.popcenter.org)

Read more

Thefts of motorcycles, cars and bicycles in Germany

Motorcycles Cars Bicycles

1980 153,153 64,131 358,865

1981 143,317 71,916 410,223

1982 134,735 78,543 453,850

1983 118,550 82,211 415,398

1984 90,008 72,170 376,946

1985 73,442 69,659 337,337

1986 54,208 70,245 301,890

Three principles of this manual are: (1) it takesmore than offenders to create problems; (2)people cannot offend if there is no opportu-nity structure to support this behaviour; and(3) altering the opportunity structures candramatically reduce problems. It follows thatresponses focusing on only removing offend-ers have limited effects on problems. It doesnot matter whether offender removal is basedon law enforcement or on reformation orrehabilitation through social services.

Removing offenders and letting opportunitiesto offend remain is not good practice. Aftersome offenders are removed, there may be adecline in the problem for a short time. Theneither the old offenders return to take advan-tage of the opportunities, or new offendersstart taking advantage of them. (This is some-times called perpetrator displacement.)Natural replacement of offenders can be slow,particularly if the opportunities are obscure.But if someone discovered the crime opportu-nities in the past, chances are others willrediscover them. If the old offenders wereremoved through imprisonment, some mayreturn to take advantage of the opportunitiesupon their release.

New offenders attracted by opportunities mightcontribute to long-term crime cycles. Bank rob-beries in parts of the United States may be anexample of this. For a few years there will be alarge number of these crimes and then they willdecline for several years, only to surge againlater to start the cycle over again. One hypothe-sis for these cycles is that during peak robberyyears, banks begin instituting a host of preven-tive measures and most offenders are caughtand imprisoned. These efforts drive bank rob-bery down. After several years with few bankrobberies, bank security becomes lax and theopportunities for bank robbery increase. Thennew offenders start to take advantage of thelower security beginning a new wave of rob-beries and prevention. This hypothesis drawsattention to the fact that it takes more thanenforcement to be effective and preventionmust be maintained to stay effective.

In fact, it is quite common to read descriptionsof problem-solving efforts that begin with adescription of failed enforcement efforts. Inevery situation either old offenders keepcoming back or new offenders have replacedthem. There are three ways in which newpeople are exposed to offending opportunities:

1. They are exposed to them through theirnormal daily routines. Police arrestyoung men stealing items from unlockedcars in a city centre, for example, butunlocked cars with things in them remainthere. Of the many people who use the citycentre on a daily basis, a few will noticethese cars and try their hand at theft. If suc-cessful, some of these individuals willcontinue to steal from cars.

2. They are exposed to crime opportunitiesthrough informal networks of friendsand acquaintances. People already expe-rienced with taking advantage of anopportunity to commit crime or disordermay invite others in to help them or enjoythe experience. Since we are seldom 100%successful at removing all the offenders forlong periods, there are usually manypeople around who can introduce newpeople to the opportunities.

3. They discover offending opportunitiesthrough recruitment. A criminal receivermay employ new burglars if the old onescan no longer supply him with goods. Ifprostitution is organised, then a pimp mayrecruit new prostitutes to fill the jobs leftvacant by the former prostitutes. Gangsmay bring in new members to replace oldones. In the United States, it has been sug-gested that adult drug dealers, faced withstiffer penalties for drug convictions,started hiring juveniles to carry out theriskiest tasks because the penalties for juve-niles caught with drugs were much lessthan for adults.

44 Watch for other offenders moving in

How do you find out if offenders are movingin? The most straightforward method is tocompare the names of offenders associatedwith the problem before the response to thenames of offenders associated with the prob-lem after the response. If the names aredifferent then offenders may be moving in.The difficulty with this approach is that oneseldom has a complete roster of the offendersinvolved. So it is not clear if the new names arereally new offenders, or if they have been partof the problem for some time, but have onlyrecently been discovered.

Offender interviews can also be helpful.Offenders may tell you when they becameinvolved in the problem, how they becameinvolved, and who else is involved. They canalso provide information on tactical and otherforms of displacement. However, offenderscan be uncooperative and unreliable.

Sometimes detailed examination of the meth-ods used to commit crimes can provideinsights into whether new offenders areinvolved. If the tactics are radically differentthan those used earlier, there is a possibilitythat new offenders are working. However, it isalso possible that the old offenders haveswitched tactics.

Combining crackdowns withenvironmental modifications:Controlling ‘away day’ prostitutes inFinsbury Park

Roger Matthews describes a Londonprostitution problem in Finsbury Park.Repeated crackdowns by the police over manyyears had failed to control the prostitutionmarket as the prostitutes simply returned tothe same area. When crackdowns werecombined with street barriers to make itdifficult for men to find prostitutes by drivinground the area in their cars the level ofprostitution activity dropped dramatically.Matthews suggests that it was thecombination of strategies – offender removalthrough enforcement and opportunity blockingthrough street barriers – that was responsiblefor the decline. One important reason whythese interventions were successful was thatthe prostitutes were not deeply committed tothis way of earning a living. Few were addictedor under the control of pimps. In fact, thecommonest reasons they gave for working asprostitutes was that they could earn moremoney than other forms of work, they enjoyedthe independence and enjoyed meeting avariety of men. Many of them came toFinsbury Park from outlying areas on cheap‘away day’ rail tickets. Together with otherwomen, they rented rooms in one of the manyboarding houses or residential hotels in thearea, or they conducted business in the carsof clients. When not working as prostitutes,many of them worked as barmaids, go-godancers or shop assistants. Their relativelylight commitment to prostitution and theiralternative ways of making money might helpexplain why the researchers could find littleevidence of displacement of the Finsbury Parkprostitutes to other nearby areas in London. A

SS

ES

S

Roger Matthews (1997). Developing More Effective Strategies for Curbing Prostitution. In Situational CrimePrevention: Successful Case Studies, 2nd edition, edited by Ronald Clarke. Monsey, New York: Criminal Justice Press.

Read more

ASSESS THE IMPACT

You can drastically underestimate the effectsof your intervention if you do not take accountof diffusion of benefits (Step 14). You may con-clude that the intervention is not worth theeffort or that it failed to suppress the problem.This is particularly true when diffusion con-taminates your control group.

Control groups show what would have hap-pened to problems if you did nothing (Step41). They need to be as similar to the treat-ment group as possible but they must not beinfluenced by the treatment. If prevention dif-fuses into the control area, you will get themisleading impression that things would havegotten better if no response had occurred. Thisinvalid conclusion will cause you to underesti-mate the effectiveness of your programme.(See Step 42 for related issues with displace-ment.)

Kate Bowers and Shane Johnson suggest twoways to select control areas when diffusion ofbenefits or displacement are possible. The firstis to create two concentric zones around thetreatment area. Prevention may diffuse intothe adjacent buffer zone but does not contam-inate the outer control area. This is feasible ifthe control and treatment areas are very simi-lar and diffusion or displacement does notreach the control zone. If these conditions arenot met, then option 2 should be used. Here

the control areas are widely separated fromthe treatment and diffusion/displacementbuffer. Controls are selected specifically fortheir similarity to and isolation from the treat-ment area. Multiple control areas can beselected and their crime rates averaged.

To determine the overall effect of the responseon the problem (including any diffusion or dis-placement effects) you need to answer fourquestions, in order. Each question has a for-mula that you can apply to the results from theresponse, control and diffusion/displacementareas to get answers.

1. Did the problem change from beforeto after the response? Subtract the crimein the treatment area before the responsefrom the treatment area after the response,using the formula for the Gross Effect:

GE = Rb – Ra (the subscripts indicateafter and before).

A positive number indicates a decline in theproblem.

2. Was the response a likely cause of thechange? Shane Johnson suggests compar-ing the change in the treatment area (R) tothe change in the control area (C), usingthis formula for the Net Effect (NE):

NE = (Rb/Cb) – (Ra/Ca).

If NE is close to zero, the response proba-bly was ineffective, and if NE is negative theresponse may have made things worse. Ineither case, displacement and diffusion areirrelevant so you can stop with the answerto this question. But if NE is positive thereis reason to believe the response may havecaused the improvement. This raises thepossibilities of diffusion and displacement.So you need to answer the next question.

3. What is the relative size of the displace-ment or diffusion? Bowers and Johnsonpropose the Weighted DisplacementQuotient (WDQ), to measure this:

Da/Ca – Db/CbWDQ = –––––––––––––– .Ra/Ca – Rb/Cb

45 Be alert to unexpected benefits

Treatment areas

Diffusion/displacementbuffer zones

Control areas

Option 2: Distant control areas

Option 1: Concentric areas

Figure 1: Two options forselecting control areas

Da is the crime rate in the diffusion/displace-ment area after the programme and Db is thecrime rate in this area before the programme.The bottom term (denominator) is a measureof the effectiveness of the response, relativeto the control. An effective programme willproduce a negative number in the denomina-tor. The top term (numerator) shows therelative amount of diffusion or displacement.The numerator is negative when diffusion ispresent and positive when displacement ispresent. If it is near zero, neither is presentand the WDQ is zero (so you can skip ques-tion 4). If the WDQ is positive there isdiffusion (remember, a ratio of two negativenumbers is positive), and if it is greater thanone then the diffusion effect is greater thanthe response effect. If the WDQ is negative,there is displacement. When the WDQ isbetween zero and negative one, displacementerodes some, but not all, of the responseeffects. Theoretically, the WDQ could be lessthan negative one, indicating the responsemade things worse. However, research sug-gests that this is an unlikely occurrence.

4. What is the Total Net Effect of the response(including diffusion and displacement)?Johnson suggests using the following formulato calculate the Total Net Effect:

TNE=[Rb(Ca/Cb)–Ra]+[Db(Ca/Cb)–Da].

The first term shows the effect of the treat-ment in response area (accounting for thecontrol area changes). The more effectivethe treatment, the more positive this term.The second term shows the level of diffusionor displacement (accounting for the controlarea). It will be positive in the presence ofdiffusion and negative in the presence of dis-placement. The more positive TNE, themore effective the programme.

Lets see how these formulas can be applied.Imagine a problem of assaults in a particularpub shown as R in Figure 2. The nearby pubwith a similar problem is a good diffusion/dis-placement site (D). We calculate the assaults as arate (e.g. 20 assaults per 1,000 patrons per year)so we can make comparisons across pubs withunequal numbers of patrons. This also allows usto have multiple control pubs (J, K and L) thatcan be averaged to form a single control.

The table shows the results. We see that diffusion contributed substantially to the effec-tiveness of the intervention, and the total neteffect was a reduction of 10 assaults per thou-sand patrons per year.

ASSESS THE IMPACT

AS

SE

SS

R

D

J

K

L

N

half a mile

Figure 2: Selecting control anddiffusion-displacement sites

Kate Bowers and Shane Johnson (2003). Measuring theGeographical Displacement and Diffusion of BenefitEffects of Crime Prevention Activity. Journal ofQuantitative Criminology, 19(3): 275–301.Shane Johnson, Kate Bowers, Peter Jordan, JacqueMallender, Norman Davidson, and Alex Hirschfield (inpress). Evaluating Crime Prevention Scheme Success:Estimating ‘Outcomes’ Or How Many Crimes WerePrevented. International Journal of Theory, Researchand Practice.

Read more

Calculating the response effects Before After A–B

Treatment area (R) 20 12 –8Diffusion area or buffer (D) 15 11 –4Control area (C = average of J, K & L) 18 17 –1

GE = 8 NE = 0.405 WDQ = 0.459 TNE = 10

Offenders often believe that prevention meas-ures have been brought into force before theyactually have. This leads to what has beencalled the ‘anticipatory benefits’ of prevention.Though these anticipatory effects can occur byaccident, the police can make deliberateefforts to create or accentuate them. To beeffective, police must have useful insight intohow offenders perceive the situation and havemethods for deceiving offenders as to the truenature of the intervention.

Marti Smith and her colleagues found evidenceof anticipatory benefits in 40% of situational pre-vention studies whose data could have revealedsuch benefits. They have provided six possibleexplanations for observed anticipatory benefits:

1. Preparation-anticipation effects occurwhen offenders believe the programme isoperational before it is actually working.For example, a property-marking schememay be announced to the public, but resi-dents have not yet been mobilised, orCCTV cameras may be installed but not yetmonitored.

2. Publicity/disinformation effects occurwhen offenders believe covert enforcementexists as the result of publicity or rumour.Offenders’ perception can be manipulated,at least in the short run, through disinfor-mation. Rather than disinformation,targeted communications can sometimes beeffective. In the United States, a project inBoston to reduce youth homicide useddirect communications with potentialoffenders to warn them that certain speci-fied behaviours would result in crackdowns.

3. Preparation-disruption effects occurwhen preparation for the prevention pro-gramme causes surveillance at theprevention sites. Surveys of residents mightalert offenders. Problem-solving projectscan create anticipatory responses duringtheir analysis stage if there is considerablevisible data collection in the community(note that statistical analysis of police datawill not create anticipatory effects).

4. Creeping implementation occurs whenparts of the response are put into effectbefore the official start date. The evaluatormay use 1 May as the beginning of the fullprogramme, but offenders detect stagedimplementation in the weeks leading up to 1May, and change their behaviour accordingly.

5. Preparation-training effects occur whenplanning, training, and surveys make thepublic or police better prepared to addressproblems and they use this new knowledgeprior to the programme going into effect. Acoordinated multi-business anti-shopliftingprogramme, for example, may be scheduledto begin on a particular date, but the discus-sions and training of employees makesthem more attentive prior to that date.

6. Motivation of officers or public occurs forsimilar reasons as preparation-training,except the people involved are more highlymotivated rather than better equipped. Thehigher motivation leads to improved per-formance in advance of responseimplementation.

Smith and her colleagues also identified fourdistinct circumstances that masquerade asanticipatory effects, but are really the results ofmisinterpretation or incomplete analysis:

1. Seasonal changes in crime can also createpseudo anticipatory effects. An interventionthat begins shortly after a seasonal turndownin crime will appear to have an anticipatoryeffect. Controlling for seasonality (Steps 24and 41) can eliminate this problem.

2. Regression effects refer to natural declinesin crime from extreme highs that occureven if nothing is done (Step 41). If a crimetrend for a problem has just dropped dueto a regression effect and a prevention pro-gramme is implemented, the naturaldecline will look like an anticipatory effect.Examining the long-term average crimelevel prior to the response, as suggested inStep 41, can reveal a regression effect mas-querading as an anticipatory effect.

46 Expect premature falls in crime

3. If a crime type (A) has been over-recorded by changing the classification ofanother crime (B), it is possible to get whatlooks like an anticipatory effect. This mightoccur if one type of crime was inflated inorder to gain funding to address that typeof crime, and then following the receipt offunding, the classification was changedback to normal. This bogus anticipatoryeffect can be detected by looking at oppo-site trends in the other crime. Finding twosimilar crimes that have opposite trendsprovides a clue that changes in the classifi-cation of crime are responsible.

4. Smoothing data (Step 24) to reveal atrend masked by random variation can pro-duce results that look like anticipatoryeffects. In the figure, a prevention pro-gramme was implemented betweenperiods 9 and 10. The raw data are plottedin the upper graph and the smoothed datain the lower graph. A pseudo anticipatoryeffect is visible in the smoothed data due tothe moving average used. A pseudo antici-patory effect is more likely the wider themoving average (e.g. 5 periods vs. 3) andthe bigger and more abrupt the decline incrime following the intervention.

ASSESS THE IMPACT

AS

SE

SS

David Kennedy (1997). Pulling Levers: Getting Deterrence Right. NIJ Journal, July: 2–8.

Martha Smith and colleagues (2002). Anticipatory Benefits in Crime Prevention. In Analysis for Crime Prevention. CrimePrevention Studies, Volume 13, edited by Nick Tilley. Monsey, New York: Criminal Justice Press (and Willan Publishing, UK).

Read more

Implementation starts

Even

ts

40

30

20

10

5

01 2 3 4 5 6 7 10 13

35

25

15

11 12 1498

Periods

Trends

Even

ts

35

30

20

10

5

01 2 3 4 5 6 7 10 13

25

15

11 12 1498

Periods

Smoothed trend (5-period moving average)

Pseudo anticipatory effect

Pseudo anticipatory benefits caused by smoothing data

How do you know that the problem actuallydeclined? Most problems vary in intensity,even when no one is responding to them. Forexample, there are nine vehicle thefts in a par-ticular city centre, on average, but seldom arethere weeks with exactly nine thefts. Instead,95% of the weeks have between 7 and 11thefts, and in 5% of the weeks less than 7 orgreater than 11 thefts are recorded. Suchrandom variation is normal. Clearly, a reduc-tion in vehicle thefts from an average of nineper week to an average of six per week iswithin the norm, and might be due to ran-domness alone. In other words, the responseto the problem may have had no impact. If thechange in the problem is small, random varia-tion is a plausible explanation. If the numberof thefts had dropped to about zero per week,with an occasional week with one theft, wecould be reasonably certain that this was not achance result; it is abnormal. How can weknow if the change in the problem is withinnormal variation?

The answer is use a significance test. A sig-nificance test compares the after responsevariation in the problem to the variation in theproblem before the response. It tells you theprobability of the change in the problem beingrandom. Your response may have caused thechange if there is a small probability of thechange being random.

If there is less than a 5% chance that the prob-lem’s change was due to random fluctuations,we reject the explanation of randomness as acause of the change. Here, 5% is called thesignificance level, and five per cent is theconventional choice. In short, we ‘bet’ thatsomething other than randomness caused thechange, if there is a ‘sufficiently low’ likelihoodrandomness could have been the cause. Wedefine ‘sufficiently low’ by our choice of signif-icance level. Often our choice is 5%, but wecan pick other levels.

You can pick a more stringent level, such as1%. The more stringent the significance levelyou select, the greater the likelihood you willmistakenly conclude that the response wasineffective when it actually worked. You mightpick a stringent significance level if the cost ofthe response is so high that you need to bevery certain it works.

Occasionally, analysts use a less stringent sig-nificance level, such as 10%. The less stringentthe level you pick, the greater the possibilitythat you will mistakenly endorse a responsethat has no effect. You might want to pick aless stringent level if the problem is serious,the measures of the problem are not particu-larly good, and you are very concerned aboutaccidentally rejecting a good response.

There are two ways of using significance levels.Social scientists typically use them as rejectionthresholds: below the level you reject chanceand above the level you accept it as the cause.In short, the significance level makes the deci-sion. If you follow this strategy, always pick thesignificance level before you calculate it. Infact, you should select it early in the assess-ment process to avoid ‘fiddling’ the figures toget the desired outcome.

A better approach is to use the significancelevel as a decision aid. Decision makers usethe significance test information, along withother facts (problem seriousness, programmecosts, absolute reduction in the problems andso forth) to make an informed choice. Manysciences, such as medicine, follow thisapproach. If you follow this approach, use a p-value instead of the significance test. Thep-value is an exact probability that the prob-lem’s change is due to chance. So a p-value of0.062 tells you that the chance of making amistake if you accept the response is a bitmore than 6%. This can be roughly interpretedto mean that in 100 such decisions, the deci-sion to reject randomness in favour of theresponse will be wrong about 6 times. Wouldyou or your colleagues take such a bet? Itdepends on a many things, doesn’t it?

47 Test for significance

There are many statistical software programsthat can make the required calculations. Thedifficulty you will face is how to choose amongthe various methods of determining the signifi-cance level or p-value. You can rely on theprogram’s default settings, but there is no guar-antee that these settings correspond to yoursituation. The difficulty lies with the hiddenassumptions behind the calculations. The fol-lowing are a few important considerations.

The variation in a problem is called its distribu-tion. In the example above, the distributionwould describe how many weeks (over a longperiod) had zero thefts, one theft, two thefts,and so forth. Many software packages assumea normal, bell shaped distribution. This isoften not appropriate because a normal distri-bution assumes that the numbers arecontinuous (like people’s income, or crimerates) rather than discrete events (like countsof crime and disorder). If you are using crimecount data, check the software package todetermine if there is an option to use some-thing like a binomial distribution.

Also, normal distributions are symmetrical,with the average in the centre. Most crimedata is highly distorted – a few people orplaces or times have most of the crime events,and many people, places or times have few orno such events (see Step 19). This is becausecrime and disorder is rare, particularly in shorttime intervals and small places. This alsomeans that chance can play a large role in thefluctuations. To find a significant change in aproblem when the events are rare, you willhave to look longer and harder at the prob-lem, both before and after the response. Thistypically means more time periods. Also, aPoisson distribution is a better assumptionwhen you are dealing with rare events.

The investigation of chance can become verycomplex. If there is a great deal riding on theoutcome of a significance test or a p-value andyou are not an expert in probability theory orstatistics, you should seek expert help from alocal university or other organisations that usestatistics on a regular basis.

ASSESS THE IMPACT

AS

SE

SS

You should pay attention to two types of costs.The first is the cost of crime. A response thatreduces crime reduces these costs and resultsin benefits. Offsetting these benefits are thecosts of the response – personnel, equipmentand so forth. Overall, the cost of preventingcrime should not be greater than the costs ofthe crimes averted.

How much does the response cost? Responsecosts will fall into two broad categories: fixedand variable. Fixed costs are made once orare obligations that cannot be changed for along time. Variable costs can be increased ordecreased. Changing a street pattern incurs afixed cost. It is done once; it cannot be easilyaltered and needs little maintenance. Trafficenforcement is mostly a variable cost, sincepolice deployment can be easily changed. Aprostitution reduction scheme involvingenforcement and changing street patterns willhave both.

It is important to calculate all the costsinvolved in an intervention. Police costsinclude personnel time (including pro ratafringe benefits), equipment purchases, materi-als used up (such as fuel and office supplies)and other expenditures. Arrest and adjudica-tion have costs. Non-police costs are similarand include resources supplied by citizens,businesses or other government agencies. Itdoes not matter if these were donated – theyare still costs to someone.

Many of the costs of crime are borne by victims. Property crimes have three costs asso-ciated with them:

1. The monetary loss of the items taken(sound systems stolen from vehicles, forexample). Records on monetary losses maybe part of police investigation reports,though the accuracy is sometimes ques-tionable because of ignorance, recall andexaggeration for insurance purposes.Further, replacement costs may be a betterindicator of harm to the victim than the ini-tial cost of the items taken.

2. Repair costs of things damaged during thecrime (car windows smashed, for example).Damage costs are often unknown to thepolice and may not be recorded.

3. Intangible costs to the property owner (losttime dealing with the theft and psychologi-cal costs of the stress related to the crime,for example). Intangible costs are extremelydifficult to calculate and are often left out ofbenefit calculations. Doing so short changesthe value of the prevention effort.

Personal crime costs are more difficult to cal-culate. Much of the cost of personal crime isintangible, and placing a credible monetaryvalue on it is difficult. There are some tangiblecosts, however, such as the monetary loss inrobberies, damage to property from assaults,and medical costs of injuries. Costs of consen-sual crimes, such as prostitution and drugdealing, are even harder to estimate. This isbecause most of the costs are indirect andintangible – affecting people not directlyinvolved in the events.

Sam Brand and Richard Price have made esti-mates for the Home Office for many personaland property crimes. These estimates do nottake into account the costs of fear and theimpact on families or acquaintances. Also,these are national estimates and there may belocal variation. Nevertheless, they can be usedas rough approximations of minimum benefits.

Cost effectiveness is the cost of the responsedivided by the number of crimes averted. If anassault reduction programme cost £10,000 andthe evaluation showed that 30 assaults wereaverted because of it, its cost effectiveness is£333 per assault. Cost effectiveness figures areuseful for two purposes. First, they allow com-parisons of interventions with different costsand levels of effectiveness. A cost of £333 perassault may be less expensive than alternativesthat either cost more or reduce fewer assaults.And second, you sometimes can determinethat the costs of the programme exceedknown costs of the crime. If the averageassault costs over £500 in medical expenses,

48 Calculate costs

for example, then this programme is alreadyworthwhile without even taking into accountother factors.

Benefit-cost is calculated either as a ratio ofbenefits to costs or as benefits minus costs.Valuable programmes have ratios greater thanone and positive differences. Under-perform-ing programmes have ratios less than one andnegative differences.

If all the costs and the benefits of the pro-gramme occur at the same time – within asingle year – then comparisons are relativelyeasy. However, many opportunity reductionefforts have response costs that come earlyand go down rapidly but have benefits thatcontinue coming in for a long time. Because itis a standard principle of accounting that abenefit today is worth more than the samebenefit some time in the future, adjustmentsneed to be made that put the future benefits(and costs) in present day pounds. Theseadjustments also account for economic infla-tion that raises costs of future responses.

Sanjay Dhiri and Sam Brand provide guidancefor making cost-effectiveness and benefit-costcalculations. Benefit-cost analysis can be ahighly technical undertaking. If it is criticallyimportant to obtain highly reliable and preciseestimates then you should ask the advice of anexpert in this field.

ASSESS THE IMPACT

AS

SE

SS

Sam Brand and Richard Price (2000). The Economicand Social Costs of Crime. Home Office ResearchStudy 217. London: Home Office.http://www.homeoffice.gov.uk/rds/pdfs/hors217.pdf

Sanjay Dhiri and Sam Brand (1999). Analysis of Costsand Benefits: Guidance for Evaluators. CrimeReduction Programme. London: Home Office.http://www.homeoffice.gov.uk/rds/pdfs/cdp1costeff.pdf

Read more

The purpose of your work is to help peoplemake better decisions. To assist decisionmakers you must tell a clear story that leadsfrom an important question to possibleanswers and to effective actions. To do thisyou need to know who your audience is andthe questions they want answered. Your storyhas to address their particular needs. Thisstory can be told in a written report or in anoral presentation (see Step 53).

You should not simply recount what you didto analyse the problem. This is tedious anddoes not help people make actionable deci-sions from your work. So once your analyticalwork has been completed you will have totranslate it into a story that addresses theneeds of your audience.

Your work can help answer four basic ques-tions. These questions correspond to thestages of the SARA process:

● Scanning – What is the nature of theproblem?

● Analysis – What causes the problem?

● Response – What should be done aboutthe problem?

● Assessment – Has the response broughtabout a reduction in the problem?

Clearly, these are general questions that can bemade more specific depending on the facts ofthe matter being examined. Local residents,for example, might complain about late nightnoise and finding litter along their street.Instead of the general scanning question, a setof specific questions could be developed: Arethere recurring instances of late night noiseand litter that disturb residents? When andwhere do these incidents occur? Who is dis-turbed? Are these complaints symptomatic of adeeper set of issues? So in this example,answering the general question requiresanswering a set of more specific questions.

Your first task in telling a coherent story is todecide which kind of question you are seekingto answer. Next, you should try to structureyour account around the basic theories andapproaches described in this manual (e.g. thecrime triangle or the 80–20 rule). These areframeworks. A framework is a general ‘storyshell’ that links the multiple interacting factorsthat cause problems. Your choice of frame-works depends on the problem, your findingsand the needs of decision makers. Be surethere is a logical flow from the basic question,through the framework and findings, to theanswers. Check for gaps in logic. Now outlinea story. There are four basic story outlines thatcan guide your work. The details of the storywill depend on the specifics of your problem.

Do not religiously stick to these outlines, buttailor them to the amount of time you haveand, above all, to the concerns of the peoplewhom you are addressing. Try to anticipatetheir questions, and modify the appropriateoutline accordingly. Though we have usedtechnical terms from this manual in these out-lines, you may need to use a commonvocabulary in your presentation. If your audi-ence is not already familiar with theterminology of problem analysis, you probablyshould use it sparingly, or not at all.

49 Tell a clear story

COMMUNICATE EFFECTIVELY

CO

M.

1. WHAT IS THE NATURE OF THE PROBLEM?

A. Organising framework – e.g. elements ofproblem definition.

B. Systematic description of evidence aboutproblem type and existence:● Time trend of crimes or disorder.● Geographic analysis and hot spots.● Types of harm.● Types of people most harmed.● Evidence from other forces and areas.

C. Implications for analysis and collaborativeproblem solving:● Questions that need answering.● Partners who need to become involved.

D. Summary.

2. WHAT CAUSES THE PROBLEM?

A. Organising framework for problem – e.g.problem analysis triangle.

B. Systematic description of problem answeringthe following questions:● Is this an enduring problem or a new

problem?● What brings the offenders and targets

together at the same places? ● What behaviours are each engaged in? ● Why don’t others step in to prevent these

encounters?C. Implications for general form of responses that

fit the information:● Offender access or control.● Victim/target behaviours or protection.● Facility access or management.

D. Summary.

3. WHAT SHOULD BE DONE ABOUT THISPROBLEM?

A. Organising framework for response – based onanalysis conclusions:● Offenders.● Targets/victims.● Places.

B. Systematic description of response strategy:● Increasing risk or effort.● Reducing rewards, provocations or excuses.● Who will carry out actions, when and

where?● Additional resources required.

C. Implications and anticipated outcomes: ● Direct results.● Displacement.● Diffusion.● Other side effects.● How evaluation should be conducted.

D. Summary.

4. HAS THE RESPONSE BROUGHT ABOUT AREDUCTION IN THE PROBLEM?

A. Organising framework – Why the response wasexpected to be effective.

B. Systematic description of evaluation:● Was the response implemented as

planned?● Did the problem change?● Why it is likely response was a direct cause

of change.● The magnitude of displacement, diffusion

and other side effects.C. Implications for further action:

● Is this problem-solving effort complete?● What further actions are necessary?● Should further analysis be conducted and

the response changed?D. Summary.

Four story outlines

Problem-solving maps serve four importantpurposes:

1. They show where problems are located(important at scanning and early in analysis).

2. They help test hypotheses about problems(important during analysis).

3. They show how problems are handled(important during the response and earlyassessment stages).

4. They show changes in problems (importantfor assessment).

We will first illustrate purposes 1, 3 and 4 witha pair of maps, and then illustrate purpose 2with a third map.

Map 1 shows the location of a commercial bur-glary problem (first purpose). This straddles ahalf-mile stretch of Battersby Road, beginningat the Sophie River, but is most intense in thetwo middle street segments. The elongatedellipse indicates the problem’s axis is BattersbyRoad and the problem does not extend muchinto the surrounding areas. A colour rangeshows the intensity of the problem. The dis-tance scale at the bottom of the map helpspeople judge size, and the arrow shows North.The map shows the relationship of the prob-lem to roads and the river. The only roadslabelled are those needed to understand theproblem’s position. The amount of labellingdepends on the local knowledge of the mapusers and the nature of the problem. Later wewill see why two additional streets should havebeen labelled.

The grey rectangle in Map 2 along BattersbyRoad shows the bounds of a merchants’ associ-ation created to address the commercialburglary problem. A more detailed map couldshow every shop colour-coded by participationin the association. This might indicate wherefurther efforts are needed, if the locations ofthe least involved shops correspond to theremains of the problem (third purpose).

The hot spots in Map 2 depict the remains ofthe problem after the response. ComparingMap 2 to Map 1 gives an indication of how theproblem changed (fourth purpose). It shrank

to about a quarter of a mile and the intensityof the problem declined because no dark blueremains in the original hot spot. The problemmay also have partially displaced to the inter-section of Young’s Road and Crowley Street.Note that these two streets were not labelledin Map 1 because they had no apparent rela-tion to the problem.

Before-and-after maps are insufficient to estab-lish that the response caused changes, butalong with other information can help makepersuasive arguments. In this example, datashowing that offenders associated with theBattersby Road hot spot are now found at thenew hot spots across the river (and they werenot there before) could lend support to thedisplacement hypothesis. In the absence ofother information, however, the displacementhypothesis has no more credibility than thehypothesis that the new hot spot grew inde-pendently of the response.

It is important that the before-and-after mapsare created in the same way (size, distancescale, directional orientation, problem inten-sity measurement, and so forth) so that anychanges can be attributed to the problem. IfMaps 1 and 2 are to be used together in thesame presentation, then they should have con-sistent street labelling.

Map 3 tests the association of problem locationswith other features (second purpose). Here, amap of the pub assault problem, introduced inStep 19, shows the eight pubs in one area of apolice district. The numbers within the circlesand squares indicate the number of assaultsassociated with each pub, and the shapes andshading of the icons indicate the pubs relative‘riskiness’. Three ‘risky’ assault pubs’ are con-centrated on Dorcus Hill and another is lessthan a mile away. All four pubs are located tothe east of the river and railway line.

This map calls into question two possiblehypotheses. The association of ‘risky’ pubswith main roads is suspect because all the lowassault pubs are also on main roads. The asso-ciation of ‘risky’ pubs with Dorcus Hillneighbourhood is not clear. One of the ‘non-

50 Make clear maps

risky’ pubs is located amongst the ‘risky’ pubs.Further, two other ‘non-risky’ pubs are locatedwithin a mile of the cluster of ‘risky’ pubs. Thissuggests that the neighbourhood effect is notabsolute and that site-specific circumstancesshould be investigated. The problem solvershould investigate what ‘protects’ cold pubsand what facilitates assaults at the sites of the‘risky’ pubs.

As useful as maps are, they do not tell a com-plete story. Aspects of a problem that are notgeographic cannot be shown on maps. For thisreason, maps are often a part of presentation,but they are seldom the only part.

Creating useful maps

1. Keep maps simple. Eliminate all featuresthat do not contribute to understanding theproblem.

2. Always include a scale and compassorientation (usually North is to the top).

3. Use meaningful gradations to showintensity of hot spots. For example, showcolours becoming increasingly hot (yellowto red) as the problem worsens.

4. Apply the correct dimension of crimeconcentration: dots for places (andsometimes victims); lines forconcentrations along streets and highways;and areas for neighbourhoods.

5. Avoid graphics that draw more attention tothemselves than the data. The readershould not have to ask about thetechniques used, but only about theproblem.

6. Make use of tables and figures along withmaps.

COMMUNICATE EFFECTIVELY

CO

M.

Old Snorthrum Rd

Battersby Rd

Mill Street

Sophie River

N

1 Mile

Dorc

us R

d

1.1 - 2.5

0.7 - 1.0

0.0 - 0.5

Burglary rate(/1,000 addresses)

Map 1: Where is the commercial burglary problem?

Old Snorthrum Rd

Battersby Rd

Mill Street

Sophie River

N

1 Mile

Dorc

us R

d

1.1 - 2.5

0.7 - 1.0

0.0 - 0.5

Burglary rate(/1,000 addresses)

Youn

gs R

d Crowley Street

MerchantAssociationMembershipArea

Map 2: Assessing the response tothe commercial burglary problem

Bacon Rd

Dew

ey R

d

Dorcus Hill

1713

31

11Mill St

Sherw

ood St

Pier

ce R

d

Sophie

River Ru

ssel

l St

Hum

e St

Pier

ce R

d

Berkely R

d

Campbell Rd

3

0

0

StreetRiverRail

Half a Mile

Lowest3 Badger1 George0 Swan0 Black Bear

Highest31 Whitehart17 Union13 Feathers11 King’s Head

Map 3: Association of pub assault problem with geographical features

Mark Monmonier (1993). Mapping it Out: ExpositoryCartography for the Humanities and the SocialSciences. Chicago: University of Chicago Press.

Read more

Tables are effective tools for telling a com-pelling story if they are made simple, but this isoften difficult. The software used to createtables adds unnecessary and distracting packag-ing – the lines and labels used to interpret thedata. In addition, analysts do not always organ-ise tables in a way that makes intuitive sense.

Assume that you are trying to show that beerthefts from off-licences may be facilitated bythe way beer is displayed. Some off-licencesdisplay the beer near the front entrance andsome at the rear. You are trying to show thatrear display off-licences have fewer incidentsof beer thefts than those where the display isin the front. Table 1 gets in the way of thismessage. The data are poorly organised andthe packaging is distracting.

Table 2 properly organises the data. The per-centages are made central to the story.Because the raw numbers do not tell the mainstory, but may be useful to a reader who wantsto look more closely, they are made sub-servient by enclosing them in brackets. Finally,instead of row percentages (as in the firsttable), column percentages are used.

Whenever we examine a relationship in whichone factor may be the cause of another factor,it is best to put the causal factor in the columnand use column percentages. Comparisons arethen made across the row. Here we see imme-diately that 29% of the front-displayoff-licences had no thefts compared to almost83% of the rear-display off-licences. At theopposite extreme, almost 46% of the front-dis-play off-licences had three or more thefts, butnone of the rear-display off-licences did.

Table 2 has less packaging. The bold bordershave been removed and replaced by thin lines.Inside, the only remaining line separates thetitle from the content. Instead of lines, space isused to guide the reader’s eye across rows anddown columns. By informing the reader in thetitle that the important numbers are percent-ages (and the raw numbers are in brackets),

there was no need to include a per centsymbol in each cell. Only the column totalremains. This tells the reader that the impor-tant sum is vertical. Finally, all percentages arerounded to one decimal place, thus allowingthe column figures to line up, making inter-pretation easier. With all of these changes,most of the content of the table is data ratherthan packaging.

A problem often has multiple causes. Thoughtables can be constructed to show large num-bers of causes, a single table communicatespoorly if the number of causes is greater thantwo. The basic principles of table constructionremain the same:

● All the causes go in the same direction(usually columns).

● Summation goes in the direction of thecausal factors.

● Comparison of causes goes in the oppositedirection (rows if causes are in columns).

Table 3 is a three-dimensional table (the earliertables were two dimensional). Table 3 answersthe question: is the relationship between dis-play location and theft different for twodifferent off-licence chains (Roberts andBolgers). The answer is readily seen; it doesnot. There is the same basic pattern for bothchains that we saw in Table 2. In both cases wesum the column and compare front-displayoff-licences to rear-display off-licences. Thisimplies that any off-licences that display beerat the rear will experience less theft.

In effect, Table 3 holds constant type of off-licence. Other factors can be held constant ifwe think they are important. For example, wecould group off-licences by size – small,medium and large – and separately analyse therelationship between display location andtheft for each size category. This would requirethree panels, but otherwise the same princi-ples apply.

51 Use simple tables

Take note of several other features of Table 3:

● If you add the raw numbers (in brackets) inthe Roberts cells to their correspondingcells under Bolgers, you get the raw num-bers in Table 2. In other words, Table 2 is asummary of Table 3. This also means thatyou cannot go the other way: derive Table 3from Table 2.

● Because Table 3 contains two possiblecauses of the problem, we have added avertical line to draw attention to the twotypes of off-licences.

● The row labels apply to both store types, sothere was no need to duplicate them.Because of rounding in the percentages,they sometimes add to over 100. In someinstances these sums can be just under 100,usually 99.9. Such small deviations areseldom of much concern.

If you routinely produce the same tables forthe same decision makers, show them severaldifferent table formats with the same data.Determine which format most effectively com-municates to them, and then use this standardformat.

COMMUNICATE EFFECTIVELY

CO

M.

Table 1: Location and beer theft (June)Location of display

Number of theft reports Front Rear Total

0 7 (17.5%) 33 (82.5%) 40

1–2 6 (46.15%) 7 (53.85%) 13

3 or more 11 (100%) 0 (0%) 11

Total 24 (37.5%) 40 (62.5%) 64

Table 2: Percentage of off-licences with reported beer thefts (numbers in brackets)

Location of display

Thefts in June Front Rear

0 29.2 (7) 82.5 (33)

1–2 25.0 (6) 29.2 (7)

3 or more 45.8 (11) 0.0 (0)

Total 100.0 (24) 100.1 (40)

Table 3: Percentage of off-licences with reported beer thefts by retail chain (numbers in brackets)

Roberts Off-licences Bolgers Off-licences

Thefts in June Front display Rear display Front display Rear display

0 30.8 (4) 84.2 (16) 27.3 (3) 81.0 (17)

1–2 23.1 (3) 15.8 (3) 27.3 (3) 19.0 (4)

3 or more 46.2 (6) 0.0 (0) 45.5 (5) 0.0 (0)

Total 100.1 (13) 100.0 (19) 100.1 (11) 100.0 (21)

Like tables and maps, figures and charts areeffective tools for conveying information, butonly if they are kept simple. All figures consistof two parts – packaging and content. Contentis the information you are interested in con-veying to others. The purpose of thepackaging is to ensure that the content can bequickly, easily and accurately interpreted.Simplicity means keeping the packaging to aminimum. The most common mistake isadding elements that get in the way of thestory. To illustrate this, we begin with an exam-ple of a poorly designed figure. Then we willshow how figures become more powerful bymaking them simpler.

Figure 1 shows a pie chart that is supposed toshow how burglars entered homes. Thethree-dimensional image distorts the message.As we will see later, doors are the biggest prob-lem and the rear window ranks fourth, behindother as the entry of choice for these burglars.The 3-D effect inflates the importance of theslices in the front while deflating the impor-tance of the slices in the back. The singlevaluable feature of a pie chart is that it showshow the parts contribute to the whole. This islost when a 3-D effect is used. Note that a vari-ety of shades and patterns need to be used todisplay the six categories. This adds clutter.

Figure 2 shows the distortion that 3-D effectscan produce in bar charts. Comparing barheights is difficult because one has to choosebetween the front top edge and the back topedge of the bar. Three-dimensional effectsshould never be used.

This chart has a number of other features thatmake it hard to use: surface shading thatmasks contrasts between the bars and back-ground, redundant bar labels and vertical axislabels, and distracting horizontal lines. Theframe around the figure is superfluous.

The simple bar chart in Figure 3 communicatesinformation very effectively because all the con-fusing features of Figure 2 have been removed.We even removed the horizontal axis. If wewanted each bar to show the exact percentage,we could label the tops of the bars. But then weshould remove the vertical axis, as this featurecommunicates the same information.

Additionally, the data in Figure 3 have beenreorganised. Instead of raw numbers of burgla-ries, the chart shows the percentage of thetotal. This communicates two points: whichmethods are more frequent, and what part ofthe whole each method represents. If youneed to show the relative contribution to awhole, use percentages in a bar chart ratherthan a pie chart.

Another feature of Figure 3 is that the cate-gories are arranged in a meaningful order:from most to least. This points to where atten-tion should be focused. Meaningful order ishard to communicate in a pie chart because ithas no obvious beginning or end. There reallyis no need to use a pie chart as bar charts cancommunicate better. When you have data incategories, bar charts are simple and effective.

52 Use simple figures

Front Door Rear Door

OtherUnknownFront WindowRear Window

Figure 1: Methods of entry

FrontWindow

60

50

40

0Rear

WindowFrontDoor

RearDoor

Other Unknown

30

20

10

6

16

54

31

20

7

Figure 2: Methods of entry

Do not forget the figure title. In Figure 3 thetitle boldly tells a story. Not only is this farmore interesting than ‘Methods of Entry’, itmakes the story unambiguous. In short, Figure3 stands on its own. Without reading anyaccompanying text, the reader gets the point.

The final figure depicts a line graph. These aretypically used when tracking data over time. InFigure 4, the data cover six months. The dotssymbolise the burglary count, and the linesindicate a continuous connection over time.The vertical axis shows the number of burgla-ries. If this chart were in a report, we wouldexpect the text to indicate that these numbersare counts, rather than rates. Generally, youshould label the vertical axis so the figure com-municates the story on its own.

If you prefer to show the number of events ateach time period, label the dots, but removethe vertical axis; it’s now redundant. Be care-ful, however. Numerical labels at every timepoint can make a chart difficult to read. If mul-tiple graphs are shown in the same figure (forexample, the trend in burglaries for severalpolice districts), make sure the different linesare clearly marked and easily differentiatedover the chart.

Designing effective figures

● Keep them simple. Don’t over package.

● Do not use superficial effects, like 3-D.

● Avoid pie charts.

● Use bar charts for data that comes incategories.

● Use line graphs for trends over time.

● Use labels effectively.

● Choose titles carefully.

● Make them stand on their own, withouthelp from the text.

COMMUNICATE EFFECTIVELY

CO

M.

FrontDoor

45

40

35

0RearDoor

Other RearWindow

Unknown FrontWindow

25

15

10

30

20

5

%

Figure 3: Doors are the problem

April

60

50

40

0May June July August Sept

10

30

20

Figure 4: The decline in burglary

A presentation should begin with a basic ques-tion, move through a description of findingswithin a framework, and end with a set of spe-cific conclusions (see Step 49). Graphicalmaterial should be prepared following theguidance in Steps 50 through 52. The mainfocus of your presentation should be toanswer specific questions that will aid deci-sion-making and it should consist of:

● A set of slides organised around your story.

● A graphical motif or outline slide to keepyour audience focused on the story.

Figure 1 illustrates a presentation of analysisfindings. The presenter, Sergeant Smith, hastwo goals. The first is to answer the question,‘What causes this problem?’ The second is toopen up a discussion of possible responses.The title slide asks the question (and intro-duces the presenter). This and Slides 2 to 4constitute the introduction. Slide 2 reinforcesa set of already agreed points that serve as afoundation for what follows. Slide 3 outlinesthe presentation. And Slide 4 summarises datacollection.

The framework is presented in the fifth slide.Sgt. Smith uses the crime triangle. All of thefindings that follow will be keyed to this trian-gle (Note that this only works if Smith’saudience is already familiar with the triangle. Ifthey are not, then Smith should use a differentframework.) To reinforce this message, and tokeep the audience from getting lost, Sgt.Smith uses the triangle motif throughout thepresentation of findings, with slight but impor-tant modifications: the shaded side and thecolour change as slides move from targets andguardians, to places and managers, and tooffenders and handlers. The circular arrow inSlide 5 indicates the anti-clockwise order Sgt.Smith will present the findings. So in this slide,Sgt. Smith has simultaneously described hisframework and provided an outline of themain findings.

Slides 6 to 13 present tables, figures and mapsthat tell the audience about the elementsdescribed in the framework. A bar chart mightshow the actions used to protect targets. Alocation map might show the places where theproblem is particularly prevalent in contrast towhere it is absent. Photos might show particu-larly important features of these sites. A tablemight show the frequency with which offend-ers are arrested.

Slide 14 summarises these findings. Here thetriangle shows all sides shaded, reinforcing thepoint that the separate findings are part of alarger whole. The final slides list responseoptions that are consistent with the findingsand options that are inconsistent. Though Sgt.Smith gives his expert opinion, it is the deci-sion makers who have the final say in thismatter. So these final slides are meant to openup a discussion that is informed by the earlierfindings.

It is important to keep your audience focusedon the larger story and from becoming lost inthe details. Two methods for accomplishingthis are to use an ongoing motif (like the tri-angle in the figure) or a highlighted outlineslide. When using an outline slide, the outlineis shown before each topic. The topic to bepresented is highlighted on the outline andthe other topics are dimmed. In Sgt. Smith’spresentation, the outline slide would beshown four times, before each of the maintopics.

Handouts of your slides are very useful, butthere are some limitations. You can make last-minute changes in the slides more easily thanthe handouts. If you expect major last-minutechanges, handouts may not correspond to theimages. Colour slides are often not legiblewhen photocopied in black and white. If youare using PowerPoint, then the ‘pure black andwhite’ option in ‘Print’ menu will temporarilyconvert your colour slides to black and whitefor printing.

53 Design powerful presentations

Most decision makers are not as interested asyou in the methods you used to analyse yourproblem. Therefore, do not spend a great dealof time describing your methods, unless this isthe objective of the presentation. Rather, sum-marise the main elements (see Slide 4 inFigure 1). You can prepare separate methodsslides, held in reserve, should audience mem-bers have methods questions. These might bea separate slide show, additional handouts, orother media.

COMMUNICATE EFFECTIVELY

CO

M.

Rand ‘Guidelines for Preparing Briefings’. Downloadfrom: www.rand.org/publications/CP/CP269/CP269.pdf

Read more

What Causes the XYZ Problem?

Sgt. Rodney SmithProblem Analysis Section

1

What is the XYZ Problem?

- High number of reported X events- Concentrated in sector Y- First documented in 1986- Does not respond well to

enforcement- Common to other police forces

2

This Presentation Shows

- Data sources used- How the data was organised- Why we have this problem- Possible responses

3

Analysis of XYZ Problem

- Undertaken by Problem Analysis Section- Data from a Variety of Sources

Reported X eventsInterviews with Merchants& ShoppersOffender InterviewsRevue of CCTV Recordings

- Experts from other Police Forces

4

Elements of the XYZ Problem

5 6 - 13

Targets

Guardian

Place

Offe

nder

ManagerHa

ndle

r

Slides describing targetsand guardians

Slides describing placesand managers

Slides describing offendersand handlers

Causes of theXYZ Problem

A bulleted summaryof the previous slides

14

InconsistantResponses

Target and Handlersa.b.

Places and Managersa.b.

Offenders and Handlersa.b.15 16

ConsistantResponses

Target and Handlersa.b.

Places and Managersa.b.

Offenders and Handlersa.b.

Slides for a hypothetical presentation

Nowadays, all professionals are required tomake presentations and presentational skillsare becoming almost as important as goodwriting. The key to a good presentation isthorough preparation.

Preparation

Never try to ‘wing it’. Thorough preparationhelps keep nerves under control.

Be sure to:

1. Know your topic.

2. Know your audience (including who else ispresenting).

3. Establish presentation length.

4. Prepare to finish a few minutes early (buttake full time if you need it).

5. Rehearse presentation (and time it).

6. Rehearse again, if necessary.

Check out the room on the presentation day

1. Is the equipment you need in place?

● Flip charts

● Black/white boards

● Chalk/marker

● Projectors

● Microphone

● Laser pointer

● Other

2. Do you know how to use the equipment?

3. Have you tried it out?

4. Do you know how to summon the technician?

5. Do you know how to dim lights?

Projector

1. Locate this in the best position for the audi-ence and yourself.

2. Make sure it does not block the view of thescreen .

3. If necessary, get help with presenting slides.

4. Make sure slides can be read from the backof the room.

5. Draw curtains or blinds if necessary.

Presentation style

1. Avoid reading your paper (even if youhave supplied a written version).

2. Speak from notes (using cards preventsyou losing your place).

3. Begin politely (thank chair, introduceyourself, greet audience, etc.).

4. If possible, stand up and speak (this helpskeep control of the audience).

5. For lengthy presentations, you can varywhere you stand (but don’t walk aboutrestlessly).

6. Do not block the audience’s view of theslide images.

7. Make sure you can be heard.

8. Don’t speak too fast (about 120 words perminute is good).

9. Maintain eye contact with the audience(but not just one person!).

10. Deal with questions quickly (or put themoff until the end).

11. Make sure handouts are clear (and thatyou have enough).

12. End on time.

13. Try to enjoy yourself!

54 Become a good presenter

PowerPoint

PowerPoint and other similar presentationsoftware allow the audience to receive theinformation simultaneously in two modes:visually from the slides and aurally from yourtalk. They are therefore more likely to under-stand and remember key points. The maindanger is that PowerPoint can result in stan-dardised presentations that quickly becomeboring for more sophisticated audiences.

1. Don’t read your slides – your talk shouldnot just be a repetition of the slides.

2. Look at the audience – not at your slides!

3. Begin with the title of the presentation,your name and affiliation (but not yourqualifications).

4. Use only one form of slide transitionthroughout, and only use a simple transi-tion that does not distract the audiencefrom your main points.

Individual PowerPoint slides should:

1. Make only one point.

2. Present just enough detail to address thepoint and no more.

3. Avoid distracting sound effects, animation,type fonts and transitions.

4. Use dark background (e.g. deep blue) andlight colours for text (e.g. yellow).

5. Use large fonts and contrasting (not clash-ing) colours.

6. Avoid thin lines and letters that cannot beeasily distinguished from the background.This can be a particular problem with linecharts and maps.

7. Avoid too much red – you should use redselectively to emphasise important points.

8. Use visual graphics rather than wordswhen possible.

9. Use clear and simple maps, figures andtables.

10. Use short bulleted phrases, not narratives,on word charts.

11. Ensure that each bullet is related to themain point of the slide.

Finally, be safe

1. Avoid technologies that break down fre-quently or ones that you are not familiarwith.

2. Have a backup plan for equipment and soft-ware failures.

3. Provide handouts of slides as supplements.

COMMUNICATE EFFECTIVELY

CO

M.

Much of what we know about problems todaywas unknown 20 years ago. This accumulationof knowledge is largely due to the sharing of knowledge by police practitioners andresearchers in the UK, United States, and othercountries. Steps 49 to 52 described how to com-municate to decision makers in your policeforce and in your community. You also have aduty to improve your profession by sharing yourwork outside your local force and community.

There are two approaches to communicatingwith your peers. The first is through writtenmaterials. These may be published in reports,professional periodicals, or popular press art-icles. The second is through presentations atprofessional conferences and meetings. Themost effective strategy for communicating infor-mation is to use a combination of theseapproaches.

Written reports can present a wealth ofdetailed useful information that others can useas reference material. There are a number ofways of disseminating written information. Itcan be made available in a downloadableformat from websites. It can be published inprofessional periodicals. Shorter piecesdesigned to capture people’s attention can bepublished in professional newsletters andother periodicals. Finally, encouraging profes-sional journalists to write about your effortscan reach even a wider audience. Shorter andmore easily accessible pieces reach a wideaudience, but contain less information.

Conferences allow face-to-face communica-tions, questions and answers, and discussionsof the latest developments. Informal discus-sions are useful for exchanging viewpoints onideas that have not developed enough to bepublished. And they allow you to seek advicefrom peer experts on difficult problems.

The United Kingdom and the United Stateshold annual conferences on problem-orientedpolicing. There are also a host of other policeconferences around the globe where you canpresent new information on problem solving.

Finally, you should also consider conferencesof other professions, particularly if you havebeen working with partners from other fields.The principal drawbacks to conferences arethe limited time available to present material,the lack of detailed permanent records of con-ference proceedings and the relatively smallnumbers of people who attend. But attendeescan spread information to those not present.

A comprehensive communications strategyshould include the following:

1. For people interested in the details, a tech-nical report downloadable through aneasily used website.

2. For a large audience of general interest,one or more short articles in professionalor popular periodicals, with references tothe website.

3. For professional colleagues and academics,a longer article in a professional journal.

4. For a small but influential group of profes-sional colleagues, at least one presentationat a professional conference.

Additionally, it is helpful to send copies of arti-cles to people who are interested in the topicyou are investigating. This not only communi-cates your ideas, but also allows you to solicitadvice as to how to communicate your ideasto others. Professionals are particularly inter-ested in:

1. Discoveries of new or changing problems.

2. Advances in analytical techniques that cananswer new questions, or answer old ques-tions more precisely and with less error.

3. New responses to problems or new applica-tions of old responses.

4. Evidence about the effectiveness, lack ofeffectiveness or side effects of responses.

Each of these topics can be written as a casestudy of your particular problem. The basic out-line for a useful case study covers four points:

55 Contribute to the store of knowledge

1. Dissatisfaction with the old situation – whythe standard understanding or practice isinsufficient in particular circumstances.

2. Search for alternatives – how a new under-standing or practice was discovered.

3. Evidence supporting alternatives – compar-ison of old and new approaches.

4. Conclusions and implications – summary ofwhat people should consider, given thisnew information.

Notice that this outline follows the SARAprocess. Scanning reveals dissatisfaction with aparticular circumstance. Analysis is a search fora new understanding of the problem. Response

requires a systematic comparison of alternativeapproaches and the selection of a particularnew approach. And assessment summariseswhat one has learned from the experience.

The table shows how this outline can be appliedto each of the four case study topics. Thesetypes of case studies can be combined, as cir-cumstances require. A new technique forproblem analysis might reveal a new type ofproblem, for example. In such a circumstance,the first two types of case study can be com-bined. Similarly, a description of a new responseto a problem might include evaluation informa-tion, thus combining the last two types of casestudy. Other combinations are possible.

COMMUNICATE EFFECTIVELY

CO

M.

Four types of case studyOutline 1. New problem 2. New analytical 3. New response 4. New evidence on

technique effectiveness

I. Dissatisfaction Discovery of an Why old technique Why old response Uncertainty of anomalous situation. is limited. is limited. effectiveness of

response underparticular circumstances.

II. Search Exploration of what How the new How the new Difficulties in is different. technique was response was evaluating response

discovered. discovered. in these circumstances.

III. Evidence Comparison of old Systematic Systematic Evaluation methodsproblem to new comparison of old comparison of old used and theirproblem. technique and new response to new results.

based on objective based on objectivecriteria. criteria.

IV. Conclusions What this implies Circumstances Circumstances Circumstances wherefor problem solving. where new where new response should be

technique is response is used and expected particularly helpful. particularly helpful. results.

Term and definition Step3-D mapping High-definition mapping that portrays locations within buildings. 22

80–20 rule The principle that a few people or places are involved in a large proportion of events. 19, 20, 26, 27, 28, 49

Activity space The nodes and routes routinely used by an offender during his/her daily activity (see Crime pattern theory). 17

Acute problems or hot spots Problems or hot spots that suddenly appear (i.e. have not been present for a long time, not chronic) (see Chronic problems or hotspots). 15, 21

Acute temporal clustering A very high concentration of crime in a small part of the 24-hour cycle. 23

Adaptation Long-term changes in offender population behaviours in response to crime prevention. 12, 40

Analysis The second stage in the SARA process involving systematic 1, 5, 6, 7, 8, 9, 17, 19,examination of the problem to identify possible causes that might be 20, 21, 31, 32, 38, 40,susceptible to responses. 46, 49, 55

Anticipatory benefits Benefits from crime prevention that begin prior to initiation of crime prevention treatment. 12, 46

Anticipatory benefits, pseudo The appearance of anticipatory benefits caused by smoothing data (i.e. the use of a moving average). 46

Aoristic analysis A statistical method for determining the 24-hour rhythm of crimes when the exact time of crime commission is unknown. 23

Assessment The fourth stage in the SARA process involving evaluating the 1, 2, 5, 8, 22, 25, 32, 40,effectiveness of the response. 47, 49, 50, 55

Attractors, Crime Areas of criminal opportunities well known to offenders. 18, 26

Behaviours One of two criteria for classifying problems describing aspects of harm, intent and offender–target relationships (see Environments). 16, 49

Benefit-cost The ratio of benefits to costs or benefits minus costs. 8, 48

Benefits The value of crime prevented and of positive side effects of 8, 28, 35, 38, 39, 45,prevention. 46, 48

Boost accounts An explanation for repeat victimisation that suggests that the rewards to the offender for the first crime encourage the offender to repeat the offence against the same victim or to tell other offenders who then attack the same victim (see Flag accounts). 27

Broad-spectrum treatments Crime prevention measures that are effective 43against a wide variety of methods for committing a type of crime.

Broken windows policing 6, 36

Buffer zone Area around a facility, hot spot or treatment area used to account for diffusion or displacement (see Diffusion/Displacement area). 17, 45

CHEERS Acronym for elements defining a problem: Community, Harm, Expectation, Events, Recurring and Similarity. 15, 16

Glossary and index

Term and definition StepChronic problems or hot spots Problems or hot spots that persist for a long 15, 21

time (see Acute problems or hotspots).

Community policing 5, 6

Content The substantive information in a table or figure. 51, 52

Control area A geographical area not receiving treatment but compared to a treatment area (see Control group). 41, 42, 43, 45

Control group A group of people or an area that is similar to the treatment group or area, but does not receive treatment. Used in evaluations to control for the impact of other, non-treatment, influences on crime (see Controls (for analysis)). 41, 42, 43, 45

Controls (for analysis) Statistical and evaluation design procedures to isolate the effect of one factor on some outcome from that of others. A group of people or areas not getting a response that are compared to those receiving the response to show what would have happened to the response group, if the response group had not received the intervention (see Control Group). 41, 42, 43, 45

Controls (on offenders) People and situations that reduce potential 7, 10, 16, 18, 20, 26,offenders’ willingness or capabilities to commit crimes. 33, 35, 36, 37, 42

Conventional policing Policing that relies primarily on the use of patrolling, rapid response and follow-up investigations to prevent crime. 4

Cost effectiveness The ratio of response costs to crimes prevented. It is a measure of the cost of preventing a single average crime. 15, 48

Costs Expenses or hardships associated with criminal events or 8, 13, 15, 32, 38, 39,prevention measures. 47, 48

Costs, fixed Costs that cannot be changed for long periods. 48

Costs, response Costs of the intervention to prevent crime. 48

Costs, variable Costs that can be changed in a short time period. 48

CPTED See Crime Prevention Through Environmental Design.

CRAVED An acronym describing the characteristics of items most likely to be stolen and standing for Concealable, Removable, Available, Valuable, Enjoyable and Disposable. 26, 29

Crime pattern theory A theory of how offenders’ normal routines create crime patterns. 17

Crime Prevention Through Environmental Design A set of principles for designing and laying-out secure buildings and public spaces. 22

Crime reduction partnerships 6

Crime triangle 9, 17, 49, 53

Crime-neutral areas Areas attracting neither offenders nor targets, with adequate controls on behaviours. 18

Glossary and indexGlossary and index

Term and definition StepCycles Regular fluctuations in crime that correspond to daily, weekly, monthly,

annual or longer changes in human activity. 23, 24, 41, 44

Decomposing a time series Breaking down a problem into parts and plotting the time series for each part. 23

Defiance Offenders challenge the legitimacy of prevention efforts and commit more offences rather than fewer. 12

Den (of iniquity) problems Problem characterised by substantial involvement of repeat places (see Crime triangle, Place). Occurs when new potential offenders and new potential targets encounter each other in a place where management is weak. 9, 16, 19

Deterrence, General Communicating a public perception that the risks and penalties of offending are high, so anyone considering such behaviour will refrain. 4

Deterrence, Specific Communicating a perception of high risk and penalty to specific individuals so they will refrain from committing crimes. 4

Diffused temporal clustering A relatively even, or random, spread of crime throughout 24-hour cycles. 23

Diffusion contamination Occurs when diffusion of benefits influences the control group or area during an evaluation. Leads to undervaluing the treatment (see Displacement contamination). 43, 45

Diffusion of benefits Reducing crime beyond the focus of the prevention scheme; a multiplier of effectiveness. 12, 14, 32, 42, 43, 45, 49

Diffusion of benefits, crime type Additional crime types blocked. 12

Diffusion of benefits, geographical Additional prevention over space. 12, 42, 43, 45, 49

Diffusion of benefits, tactical Additional methods thwarted. 12

Diffusion of benefits, target Additional targets protected. 12

Diffusion of benefits, temporal Additional prevention over time. 12

Diffusion/Displacement area Areas used to detect diffusion of benefits and displacement that are separate from control group and treatment group. 45

Displacement Offenders changing their behaviour to thwart preventive actions. 1, 5, 12, 13, 14, 27, 32, 34, 40, 42, 43, 44, 45,

49, 50

Displacement contamination Occurs when crime is displaced into the control group or area during an evaluation. Leads to inflation of effectiveness (see Diffusion contamination). 42, 43

Displacement countermeasures Prevention implemented to prevent expected displacement. 42, 43

Displacement, crime type Offenders change type of crime. 12, 13, 43

Displacement, geographical Offenders move spatially. 12, 13, 27, 40, 42, 43, 45

Displacement, tactical Offenders switch method for committing crime. 12, 13, 43, 44

Displacement, target Offenders switch type of target or victim. 12, 13, 43

Glossary and index

Term and definition StepDisplacement, temporal Offenders switch time or day. 12, 13, 42, 43

Distribution A description of the numerical variation in the magnitude of a problem. 23, 47

Distribution, binomial A distribution used for crime count data. 47

Distribution, normal A bell-shaped distribution used for continuous data. 47

Distribution, Poisson A distribution used for discrete data where one of the outcomes is rare. 47

Duck (sitting) problems Problems characterised by substantial involvement of repeat victims (see Crime triangle). Occurs when victims continually interact with potential offenders at different places, but the victims do not increase their precautionary measures and their guardians are either absent or ineffective. 9, 16, 19

Edges Boundaries between areas where people live, work, shop or seek entertainment. 17

Enablers, Crime Places with little regulation of behaviour. 18, 25, 26

Environments A criterion for classifying problems describing where the problem takes place (see Behaviours). 16, 23, 26, 30

Facilitators Physical items, social situations or chemical substances that help offenders commit crimes or acts of disorder. 30

Facilitators, chemical Substances that increase offenders’ abilities to ignore risk, reward or excuses. 30

Facilitators, physical Things that augment offenders’ capabilities, help overcome prevention measures or incite deviancy. 30

Facilitators, social Situations that provide support that stimulates crime or disorder by enhancing rewards from crime, by legitimating excuses to offend or by encouraging offending. 30

Facilities Places that have special functions, like schools, businesses and 6, 21, 23, 26, 27, 28, 30,restaurants. 32, 33, 36, 37, 38, 50

Facilities, risky Facilities that are frequent sites for crime and disorder. 26, 27, 29, 30, 38

Flag accounts An explanation for repeat victimisation that suggests that some people are particularly vulnerable because of their occupation or their ownership of hot products (see Boost accounts). 27

Focused temporal clustering Clustering of crime in distinct time ranges during 24-hour periods. 23

Generators, crime Areas to which large numbers of people are attracted for reasons unrelated to criminal motivation. 18

Gross Effect Crime in the response area before the treatment, minus crime in the response area after the treatment. A positive result indicates a crime reduction (see also Net Effect, Weighted Displacement Quotient, and Total Net Effect). 45

Handler Someone who knows an offender well and who is in a position to exert some control over his or her actions. 9, 23, 53

Glossary and index

Term and definition StepHot areas Hot spots showing neighbourhoods where crime is concentrated. 21

Hot dots Hot spots showing locations with high crime levels. 21

Hot lines Hot spots showing street segments where crime is concentrated. 21

Hot products Things that are particularly attractive as targets (see CRAVED). 19, 26, 29

Hot spots Geographic concentrations of crime. 1, 4, 9, 17, 18, 21, 23, 29, 42, 49, 50

Hypothesis An answer to a question about a problem that can be true or false, and may or may not be supported by evidence. 20, 25, 44

Impact evaluation A research study to determine if the response changed the problem. 40

Incapacitation Removing active offenders from society to prevent crimes that they would have committed if they were not locked up. 4

Intervention The response being applied to a problem (also called a 5, 8, 11, 12, 19, 25, 30,treatment or response – see Response). 37, 38, 39, 40, 41, 42,

43, 44, 45, 46, 48

Manager A person who has some responsibility for controlling behaviour in a specific location. 6, 9, 22, 23, 28, 32, 34, 53

Mechanism The process by which a response works on a problem to have an effect, or a process that causes a problem. 18, 32, 39

Moving average A method for reducing random fluctuation in a time series by recomputing the value for every data point based on the average of preceding time periods (see Smoothing). 23, 46

National Intelligence Model 1

Net Effect The change in crime in the response area relative to change in crime in the control area (see also Gross Effect, Weighted Displacement Quotient, and Total Net Effect). 45

Nodes Destination places such as home, work, shopping, entertainment and school (see Paths). 17

Offender A person who commits a crime or act of disorder. 1, 2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 23, 26, 27, 28, 29, 30, 31, 32,

33, 34, 35, 36, 37, 38, 40, 42, 43, 44, 46, 49, 53

Offenders, repeat People who commit many crimes or acts of disorder (see Wolf). 26, 28, 29

Opportunity/opportunity structure Physical and social arrangements that 10, 12, 13, 14, 38,make crime possible. 42, 44, 48

Owner of a problem People or institutions that are unwilling or unable to undertake prevention measures, thereby helping to create a problem. 38

Packaging The lines and labels used in tables and figures (see Content). Small amounts are needed to help interpret content, but large amounts obscure content. 51, 52

Paths Routes connecting nodes. 17, 21

Glossary and index

Term and definition StepPerceptions, offenders’ How offenders view situations and prevention

measures (see also Deterrence, specific). 4, 12, 28, 37, 46

Place A very small area, such as an address, street corner, or block face 5, 9, 12, 16, 17, 18, 19, 21,(see Crime triangle, Den). 23, 26, 28, 32, 33, 34, 37,

38, 42, 47, 49, 50, 53

POP See Problem-oriented Policing.

Problem analysis triangle See Crime triangle.

Problem-oriented Policing Policing that changes the conditions that gives rise to recurring crime problems and does not simply rely on responding to incidents as they occur or forestalling them through preventive patrols. 1, 2, 4, 5, 6, 8, 13

Process evaluation Assessing how a response was implemented. 8, 40, 41

Provocations Physical designs or the way places are managed that provoke misconduct. 26, 30, 32, 36

p-value The probability that the difference between two sets of statistics is due to randomness (see Significance test). 47

Random fluctuations Short-term changes in problems caused by a large number of very small effects. 23, 46, 47

Rates, crime The ratio of crimes to targets for an area. Used to control for 18, 20, 21, 25, 26, 41, differences in the number of targets (see Risk, crime). 45, 52

Regression to the mean The tendency for abnormal high or low levels of crime to move back to their normal levels. 41, 46

Response The third stage in the SARA process involving the development 1, 5, 6, 8, 9, 11, 15, 18, 20, and implementation of an intervention designed to reduce a problem. 21, 24, 31, 32, 38, 39, 40,Also a term for the preventive treatment or intervention being applied 41, 42, 43, 44, 45, 46, 47, (see Intervention or Treatment). 48, 49, 50, 53, 55

Response area See Response group. 41, 45

Response group People or places receiving prevention in contrast to control group. 41

Risk, crime The chance a target will be involved in a crime. 5, 7, 17, 18, 21, 23, 24, 25, 26, 27, 29, 35, 50

SARA An acronym for the problem-solving process (see Scanning, Analysis, Response and Assessment). 1, 5, 6, 8, 49, 55

Scanning The first stage in the SARA process involving problem identification, 1, 2, 5, 8, 15, 17,verification and classification. 19, 32, 49, 50, 55

Self-containment index Proportion of crimes in an area committed by offenders living in an area. 26

Significance level A threshold below which one rejects the possibility that the difference between two sets of statistics is due to randomness. Often 0.05 (or 5%) is the rejection threshold (see Significance test). 47

Significance test A statistical procedure used to determine whether the difference between two groups of numbers is due to randomness. 47

Glossary and index

Term and definition StepSituational Crime Prevention The science of reducing opportunities for 1, 2, 14, 30, 32, 33,

crime. A set of 25 techniques divided among five categories: increase 34, 35, 36, 37offenders’ efforts, increase offenders’ risks, reduce offenders’ rewards, reduce provocations to offend and reduce excuses for offending.

Smoothing Removing random fluctuations from a time series by using a moving average (see Moving average). 23, 46

Target The person or thing an offender attacks, takes 3, 7, 8, 9, 11, 12, 13, 14, 15, 16, 17, 18, 19, or harms (see Victim). 20, 23, 24, 25, 26, 27, 28, 29, 32, 33, 37, 38,

41, 42, 43, 49, 53

Target area An area with many potential targets (e.g. a car park for a car thief). 17

Temporal clustering Concentration of crime over 24 hours (see Acute, Diffused and Focused temporal clustering). 23

Time-window effect The underestimation of repeat victimisation due to using a set time period. 27

TNE See Total Net Effect.

Total Net Effect A formula for measuring the full impact of a response on a treatment area and diffusion/displacement area, while accounting for changes in a control area (see also Gross Effect, Net Effect and Weighted Displacement Quotient). 45

Treatment See Response or Intervention. 42, 43, 45

Treatment area An area receiving a response (see Response group). 42, 45

Treatment group See Response group. 43, 45

Trend A steady increase, decrease or stable level of crime over some 10, 20, 23, 41, 43, 46,period of time. 49, 52

Victim A human target or the owner of stolen goods or damaged 2, 3, 4, 5, 7, 9, 11, 15, 16, 17, 19, 21, property (see Target). 23, 25, 26, 27, 28, 29, 30, 31, 32, 38,

41, 42, 48, 49, 50

Victim, repeat A person or place with multiple crimes or acts of disorder 9, 21, 23, 24, 26, 27,(see Duck). 29, 40

Victimisation, repeat The process leading to repeat victims. 9, 21, 24, 26, 27, 28, 32, 35, 40

Virtual repeats Victimisation of targets that are very similar, though not identical (as in the case of repeat victims or places). 27

WDQ See Weighted Displacement Quotient.

Weighted Displacement Quotient A formula for measuring the effects of diffusion of benefits and displacement (see also Gross Effect, Net Effect and Total Net Effect). 45

Wolf (ravenous) problems Problems characterised by substantial involvement of repeat offenders (see Crime triangle). Occurs when offenders are able to locate temporarily vulnerable targets and places. 9, 16, 19

Glossary and index