new perspectives in policing: one week in heron city (a

Sparrow—One Week in Heron City (Case B) | 1

New Perspectives in Policing

IntroductionThe Heron City case study is divided into three parts — Case A, Case B and Teaching Notes. The case study is

designed to serve as a basis for discussions regarding: (a) the relationships among a range of current policing

strategies, and (b) the nature of analytic support that modern operational policing requires.

The broad strategic or organizational approaches discussed in the case study include:

• Communitypolicing.

• Compstat(asanorganizationalapproachtocrime-reductiontasks).

• Problem-orientedpolicing.

• Evidence-basedpolicing.

• Intelligence-ledpolicing.

One Week in Heron City (Case B)A Case Study

Malcolm K. Sparrow, Ph.D.

S e P t e m B e r 2 0 0 9

National Institute of Justice

Executive Session on Policing and Public Safety This is one in a series of papers that will be published as a result of the Executive Session on Policing and Public Safety.

Harvard’s Executive Sessions are a convening of individuals of independent standing who take joint responsibility for rethinking and improving society’s responses to an issue. Members are selected based on their experiences, their reputation for thoughtfulness and their potential for helping to disseminate the work of the Session.

In the early 1980s, an Executive Session on Policing helped resolve many law enforcement issues of the day. It produced a number of papers and concepts that revolutionized policing. Thirty years later, law enforcement has changed and NIJ and Harvard’s Kennedy School of Government are again collaborating to help resolve law enforce-ment issues of the day.

Learn more about the Executive Session on Policing and Public Safety at:

NIJ’s Web site: http://www.ojp.usdoj.gov/nij/topics/law-enforcement/executive-sessions/welcome.htm

Harvard’s Web site: http://www.hks.harvard.edu/criminaljustice/executive_sessions/policing.htm


One Week in Heron City: Case B

Heron City is fictional. So are all the characters in the following narrative.

Wednesday, Late Morning: Nigel Jewett, Junior Analyst, IT1 ServicesAsChiefHarrisonwasleavingtheITdepartmentafterhermeetingwithITDirectorPhilGoring,shestoppedto

say hello to Nigel. Nigel showed her some license plate images on his computer screen, which the Chief agreed

were pretty easy for the human brain to interpret, and which the system ought to have read correctly. Nigel

explained to the Chief that the biggest problem was the system’s failure to distinguish between six and eight,

particularly if a license plate was dirty. He believed that the misreading of sixes and eights, where the software

interpreted one of these digits as the other, accounted for about 80 percent of the errors he had examined. The

nextbiggestsourceoferrorswasfailuretodistinguishtheletter“I”fromthenumber“1.”

AsChiefHarrisonwasabouttoleave,Nigelaskedher,somewhatshyly,ifhecouldtalktohersometime.“About

anythinginparticular?”sheasked.“Aboutstolencars,”hereplied.Shetoldhimshewasabouttogoandfinda

cupofcoffeeinthecafeteriaandaskedhimifhe’dliketowalkdownwithher.

FollowingareexcerptsfromthefirstmeetingbetweenChiefLauraHarrisonandNigelJewett,junioranalyst,IT

department.Themeetingtakesplaceintheheadquarterscafeteria.

Chief Harrison: IunderstandfromMr.Goringthatyourjobisprettyrepetitive.Weappreciateyoudoing

thisanalysis.Iunderstanditisimportantforthesakeofimprovingthesystem.SoIhope

you can stand it.

Nigel Jewett: Isupposeitisimportant.ButitiscertainlynotwhatIliketodo.

Chief Harrison: Whatdoyouliketodo?

Nigel Jewett: Dataanalysis.Imajoredinoceanographyandmarinebiology,andIjustlovetofindout

thewaytheworldisandthewayitworks.

Chief Harrison: Whatyou’redoingisn’tdataanalysis?

1 Information Technology

4 | New Perspectives in Policing

Nigel Jewett: No.Nobodyheredoesmuchdataanalysis.WhatI’mdoingisdataqualityanalysis.As

iftheendgoalistohaveavastlibraryofperfectdata!What’sthepointofhavingavast

libraryofperfectbooksifnobody’sreadingthem?

Chief Harrison: Whatsortofanalysisdoyouliketodo?

Nigel Jewett: Oh,allsorts.Whateverhelpsfindouthowthingswork.WhenIwaseightyearsold,my

parentsgavemeabookcalledHow Things Work,andeversincethenI’vebeentaking

thingsapartandsometimes[hesmiles]puttingthembacktogetheragain.Then,incollege

Idiscoveredhowtotaketheworldapartusingarangeofdatasources,andI’mwaiting

forajobatNOAA,2whichthey’vepromisedme,soIcanstartdoingthatkindofworkall

over again.

Chief Harrison: Whatisityouwanttotakeapart,onceNOAAletsyouin?

Nigel Jewett: Idoneedtotellyousomethingaboutthestolencars.ButwhatIreallywanttoworkon,

at NOAA, is hypoxia.

Chief Harrison: Hypoxia?Forgivemyignorance.Whatishypoxia?

Nigel Jewett: Deadzonesintheocean.ParticularlyintheGulfofMexico.Theyareseasonal,butthey’re

getting bigger. They can be hundreds of miles across now. One year, the hypoxic zone in

theGulfwasasbigasthestateofMassachusetts.

Chief Harrison: Deadzones?

Nigel Jewett: Yes. No oxygen, or extremely low levels of oxygen, from the bottom of the ocean all the

way up to a depth of about 9 meters. Hypoxia is related to excessive nutrient loadings and

red tides, but it is a distinct phenomenon and not the same as red tide and algae blooms,

whicharemuchbetterknownandbetterunderstoodonthewhole.

Chief Harrison: So,howdoyoustudysuchathing?Howdoesthatrelatetodataanalysis?

2 The National Oceanographic and Atmospheric Administration


Nigel Jewett: OneofmyspecialtieswasSONAR.3IusedtolovesiftingthroughdatabasesofSONAR

andsoundrecordingsandfiguringouthowwemightormightnotbeabletomonitor

thesedeadzones.Ortheiredges.Shoalsoffishgatheraroundtheedgesofthezone.They

can’t go into it because they can’t breathe.

Chief Harrison: Isthattypicalofthetypeofworkthatmarinebiologistsdo?

Nigel Jewett: Well,Iliketodoanythingthatinvolvesfindingmeaning,orpatternsorstructures,from

hugedumpsofotherwiseuselessdata.That’swhatImeanbydataanalysis.

Chief Harrison: Whatsortofpatternsdoyousearchforinmarinebiology?

Nigel Jewett: Allkindsofthings.There’soneclassofanalysiswhereyoulookforthingsyoucandescribe

butthatshouldn’thappen.Likeahypoxicareabecausetherearenocrabsonthebottom

whenthereshouldbe,andnoshoalsoffishformiles.Whereabottom-trawlwouldcome

up with no living organism. And for very different types of research, other oddities, for

example:afemaledolphintravelingalone.Dolphinsdon’tgenerallytravelalone,except

for the older males who peel off when it’s time to go off and die. First you describe what

you’relookingfor.Thenyouworkoutwhatthedatarepresentationofthatthingwould

looklike,ifitexisted,withinthedatathatyouhaveordatayoucanget.Thenyouwork

outhowtosearchforitwithinthedatawithoutevenknowingwhetheritmightbethere

or not.

Chief Harrison: Are there other classes of analysis you use?

Nigel Jewett: Yes.Anotherbigcategoryisdeviationsfromknownpatterns.Whenmigrationpatterns

change.Whenoceantemperaturesdeviatefromseasonalnorms.That’sparticularly

important for many species in spawning seasons. A couple of degrees off normal is

enoughtosubstantiallyreduceaspawningyield.AndIdidsomeworkonceonearly

signsofsicknessandstressincoralreefecosystems.

Chief Harrison: Canyoutellifpeoplearesick?Ifthepopulationhadfluforinstance?Fromdataanalysis

of various types?

3 Sound navigation and ranging (SONAR) is a method of locating objects in the water and determining their echo characteristics.


Nigel Jewett: I’msureyoucan.Googlecan.Theycanseehowmanypeoplearelookingup“flusymptoms”

forinstance.Youcouldwatchsalesofhealth-relateditemsorprescriptions,assuming

youcancollectallthatdatafromthepharmacies.Epidemiologistshaveallthosekinds

ofmonitoringsystems,I’mprettysure.

Chief Harrison: Couldyoutellafluoutbreakfromtrafficpatterns?Iguessyoucouldseeiftherewasany

dipinoveralltrafficvolume,andthatmightgiveyouacitywideindication.

Nigel Jewett: Youcouldprobablydomuchbetterthanthat.Maybeifyoufocusedoncommuters.

Chief Harrison: Couldyoupickoutthecommuters?

Nigel Jewett: Sure.Atleast,youcouldifyoucouldjustgetthedataontosomeusableplatformwhereyou

couldplaywithit.Wedon’thavesuchathinghereatthemoment.Butyoucouldidentify

thecommutersbythefactthattheywouldexhibitadaily-repeatpatternintheirmove-

ments.Andthenifyouwereinterested,youcouldcountby10a.m.—forinstance—how

manyofthenormalThursdaymorningcommutershadnotgonetoworkthisparticular

Thursday. That might be more accurate than overall volume, because you could ignore

allthetrafficpassingthroughandthetourists,andallofthatothernoise.

Chief Harrison: Couldyoutellifsomeonewasbeingfollowed,fromtheALPRdata?

Nigel Jewett: Someoneinparticularorsomeoneingeneral?

Chief Harrison: Someoneinparticular.LikeHayleyScott.

Nigel Jewett: Shouldbeableto.It’saneasyenoughpatterntodescribeandthensearchfor.Onecar

following close behind another, and doing it more often than you’d expect by chance.

Chief Harrison: AreyouawarethattheScottmurderinquiryhadanoutsidecontractorlookatthisalready?

Andtheydidn’tfindanythingofinterestexceptthatherhusbandoftenescortedherhome,

which is hardly surprising in the circumstances.

Nigel Jewett: Ididn’tknowthat’swhatthedumpwasfor.Mybossaskedmetodothedatadumpofone

month’sALPRdata,whichIdid,wartsandall.Heonlytoldmehehadaspecialprojectto

doforthedetectivebranch,andwouldIpleaseruntheextractionroutine.Itwasallabit


hush-hushforsomereason,andweweren’tsupposedtotalkaboutit.Thisseemsreally

oddtome.It’snotjustthatdataanalysisisn’tdonehere.Ialmostgettheimpressionit’s

prohibited.Likewe’dgetintolegaltroubleifweactuallyusedthedataforanything!

Anyway,Ididthedumpforhim.ButIbetthecontractorsdidn’tdothejobright.

Chief Harrison: Whatdoyoumean,“Didn’tdothejobright”?

Nigel Jewett: Theydon’tknowthedata.Dotheyknowaboutthetranspositionerrors?Didtheycorrect

for them or allow for them in the analysis?

Chief Harrison: Idon’tknow.Doyouthinkan8-percent-errorratemattersmuchwhenyou’relookingfor

apatternofstalking?

Nigel Jewett: That’sactuallyratheraninterestingsubjectforscientistswhocareaboutpatternrecog-

nition.Somestructuresthatyoulookforfallapartprettyquicklywithadataerrorrate

ofonly1percentorso.Otherstructuresremainprettyvisibleevenwitherrorratesupto

50percent.Somepatternsyoucanseequiteclearlywithonly10percentoftherelevant

information.ButIhavenoideahowthisoutsidegroupapproachedtheproblemand

whatallowancestheymade.AndsoIdon’tknowhowrobustthesearchmethodthey

used would be in compensating for missing or bad data.

Chief Harrison: Yousaid“structures.”Whatdoyoumeanbythat?Likeadolphintravelingalone?

Nigel Jewett: Well,structuresarehigher-levelobjectsthanindividualitems,likeadolphinorasick

whale. You can search for structures of coincidence that ought not to be there. Or you can

searchforstructuresthatshouldbetherebutaren’t.Likeawholegroupofcommuters

notgoingtoworktoday.Thesearethefourmaingeneralclassesofthings-to-search-for

withinmassivedatabases:(1)known“bads,”(2)deviationsfromnormality,(3)presence

ofstructuresthatshouldn’tbethereand(4)absenceofstructuresthatshould.Whatever

thethingisyouareinterestedin,youfigureoutthedatarepresentation,thenyoucreate

an algorithm to search for it and to display the results in a form you can use, and then

you test the output to see if it means what you thought it meant. Then you go around

throughtherefinementloopseveraltimes,untilyou’vegotarobustalgorithm,showing

yousomethingyouwantedtoknow.


Chief Harrison: I’veheardthatword“algorithm,”butneverreallyunderstoodwhatitmeans.Infact,one

ofmyhusband’sbrothersdescribeshimselfasanalgorithmist.IknowheworksatLos

Alamoslabsand,lastIheard,hewasworkingoncredit-cardfrauddetectionforcredit

cardcompanies.Ithoughthewasacomputerprogrammer.Isitthesamething?

Nigel Jewett: Not really. You can represent an algorithm in a program, but you shouldn’t need to do any

programmingyourselfifyou’vegottherightkindofhigh-levelquerytools.Analgorithm

is simply a series of logical steps, in order, that get you the analytic product you want from

thedatayoustartwith.IuseExcelalot—simplespreadsheets—andwhateveryoudo

once,youcanrecordasamacro.Youdon’tneedtoknowanyprogramminglanguage.

Then,ifyouwanttorepeatthatanalysis,youjustcallthemacro.Onekey.Canbepretty

efficient.Theonlyproblemisthatyoucanonlystickamillionorsorecordsintoanyone

spreadsheet,whichisabitofalimitationforsomethings.So,forbiggerdatasets,Ihave

to use some other tools.

Chief Harrison: Whydoyousayyoucan’tdoanalysisoftheALPRdatahereintheITdepartment?

Nigel Jewett: That’swhatIwantedtotellyouabout,ma’am.Here,thesystemisrestrictedtotheuses

alreadydefinedinthecontractaheadoftime.Thesystemgivesalerts.Andapartfrom

thatitjuststoresthedata.There’snofacilitytosliceanddiceorplayaroundwiththe

data. The contract was set up on the assumption that they could predict, ahead of time,

everythingthey’deverwanttodowiththedata.That’scrazy.Nooneeverknowsthatup

front.

Chief Harrison: Soyoucan’thuntfordolphinstravelingalone?Orforanythingelseofinterest?

Nigel Jewett: Well,Ican.That’sthepoint.Butnothere.Andmybossisonlyinterested,unfortunately,

indataquality.IfItrytopokearoundanddosomethingusefulwiththedata,likefigure

outwhat’shappeningtothestolencars,notonlydoIhavetomesswiththeoperating

system(andpotentiallyupsetthevendor)butIalsodon’tkeepupwithmyworkload

schedule,whichalreadytakesmesevenhoursadayattheworkstation.Andapparently

it might be frowned upon, legally, as well.

Chief Harrison: Sowhatdoyoudo?


Nigel Jewett: WhatInormallydoisgetbored,andmyeyesgetsore.WhatIdidthistimeis,Itooka

copyofonemonth’sdata—theextractIwasaskedtorun—andtookithomewithme.

OnmyUSBstick,25gigabytes.Nottheimages,ofcourse;thatwouldbewaytoomuch

data.Justtheextracteddata.MaybeIshouldn’tbetellingyouthis!

Chief Harrison: Whyonearthdidyoutakeithome?

Nigel Jewett: SoIcouldloaditontomycomputerandusemyowntoolsandmacros.I’vebuiltupa

prettyusefulcollectionofthemovertime.Thesystemdumpcomesoutinasortofflat-

file,comma-delineatedformatandyoucanreadthoseintoExcelifyouknowwhatyou’re

doing.Youhavetostripoutawholebunchofspuriouscharactersandsoon.Butyoucan

getthedatainaformyoucanshuffle,andfilterandcarveup.So,thenIcoulddivedown

into the data and swim around a bit.

Chief Harrison: And what did you see?

Nigel Jewett: Ithinkyou’vegotthreedifferentstolencarproblems.Notjustone.

Chief Harrison: Whatdoyouthinktheyare?

Nigel Jewett: Firstofall,there’stheordinaryjoyriding.Seemstobeoldercarsonthewhole.They

mostly stay within the city. They’re found — according to the stolen car report data in

the system — within a day or two. And damage seems to be minor and miscellaneous.

Notdeliberatevandalismoranythinglikethat.

Chief Harrison: OK.That’snotthepiecethateveryone’sbaffledbyatthemoment.Whataboutluxury

cars?

Nigel Jewett: That’sthenextpiece.Ifyoufilterouttherecoveredcars,andrankordertheremainderby

reportedvalue,thenthetop100casesaredominatedbythreemakes:Lexus,BMWand

Mercedes.

Chief Harrison: Yes,weknewthat,Ithink.Butwhyaren’twegettingalertsonthese?Don’ttheygopast

theALPRlocations?


Nigel Jewett: Yes, they do. The reason you don’t get alerts is because they haven’t been reported stolen

atthetimetheygopasttheALPRs.Theyaretypicallyreportedstolentwotothreehours

afterthelastALPRsightingwithinourarea.

Chief Harrison: Soyouthinktheownersaredrivingthemawaysomewhere,andreportingthemstolen

later?

Nigel Jewett: No.Itisnottheowners.Ilookedatthestolen-fromlocations,andtheyseemtobedif-

ferentparkinglots,spreadallovertown.Buttheyareallconnectedtosimilartypesof

establishments.

Chief Harrison: Likewhat?

Nigel Jewett: Cinemas.Restaurants.Theaters.Parkinglotsinthetheaterdistrict.Theseareallplaces

whereyouknow,whenacararrives,thattheownersaregoingtobebusyforatleasttwo

hours.Watchingamovie.Havingtheirmeal.Whateverthey’retherefor.Which,Ithink,

gives the thieves time — assuming they watch the car arrive and steal it right away — to

driveitoutofourarea,andevenpastourALPRs,beforeitgetsreportedmissing.

Chief Harrison: Smartenough.Soyouthinktheyaretargetingnotonlyparticulartypesofvehiclesbut

patronsofparticulartypesofestablishmentjusttocutdowntheriskofalerts?Canyou

see where they are going?

Nigel Jewett: Ithinkthisisatheft-to-orderoperation.Theheaviestconcentration,intermsoftime,is

WednesdayandThursdayevenings,between7and10p.m.Thosearethetimesofthe

lastALPRsightings,withinourarea,ofluxuryvehiclessubsequentlyreportedstolen

butreportedthesamenight.Ihaveagraphofthesetimesforthatdefinedsubset.The

spikesareobvious.Andwherearetheygoing?ThelastsightingsarenearlyallonI-572,

southboundforMaynard.

Chief Harrison: Howcomewe’renotgettingrecoveriesfromMaynard?Doyouthinktheydon’tstayin

Maynard?

Nigel Jewett: Idon’t.There’sacontainershipthatleavesfromMaynardeveryFridaymorningforRio

deJaneiro.Ithinkthey’rebeingshippedoutonthat.


Chief Harrison: Howonearthdoyoumakethatconnection?

Nigel Jewett: Ifigureditmustbeorganized.Thecoincidenceswerejusttoomanytobeisolatedcriminal

actions.Thisisoneofthosestructuresthatshouldn’tbethere.Itshowsthatthisisnot

arashofisolatedcrimes.AndwhenIsawtheI-572connection,Iwonderedaboutthe

portandwonderedwherethemarketwasforallthesecars.SoIGoogled“Lexus,BMW,

Mercedes,”justtoseewhatwouldcomeup.

Chief Harrison: And what came up?

Nigel Jewett: Thatwashopeless.Ahugelistofcardealerships.So,thenItriedthesamethingwith

“secondhand.”Samething.Hopeless.ThenItypedinthethreemakeswith“wholesale,”

andIstartedgettinghitsonalotofpartsdistributorsinLatinAmerica.Apparentlythere

isarealclamorforparts,forthesecarsinparticular,alloverLatinAmericaandBrazilin

particular.Ithinkthesecarsareworthafortunechoppedupintoparts.Notwhole.So

it’sawholesaleexportoperation,inparts,andweshouldprobablytalktoCustoms.

Chief Harrison: All this analysis … is this the type of analysis that criminologists do? Or do you marine

biologists have your own approach?

Nigel Jewett: Ithinkthecriminologistsandsocialscientistsworryalotmoreaboutstatistics,andregres-

sions,andthingslikethat.I’mactuallynotverygoodatthatstuff.Ifailedstatsincollege

andwasthereforenotallowedtotaketheregressioncourse.Iwasneverreallycomfort-

ablewiththeideaofrandomnessanyway.OrmaybeIjustwasn’ttaughtprobabilitytheory

particularlywell.Butthat’snotthetypeofanalysisscientistsdoverymuchanyway.When

you’restudyingnaturalsystemsandtryingtofigureouthowtheywork,yougoatitwitha

much more open mind. You don’t assume one particular analytic method is going to show

youanything.Youkeepanopenmindaboutthenatureofthething,andthestructure

underneath. You have to get under the covers, down in the weeds, and see what you see.

Thenyoucanworryaboutmodelsandtheories,andpickandchoosemethodsthatmight

actuallymakesenseandarebasedonwhatyou’veobserved,andsoon.Butatamuch

lowerlevel.Allthestuffthatreallycountsisbeneaththesurface.Andyouhavetofindit

first,andlookatit,andunderstandwhattypeofthingeachoneis.

Chief Harrison: Isthatlike“findingdistinctsubcomponentsofacrimeproblem,andstudyingtheirunique

characteristics,”likeHermanGoldsteinsaid?


Nigel Jewett: Ratherthanassumingacrimerateisjustacrimerate,andtreatingitasifitwerejustone

genericthing?Yes.Ihaven’theardofGoldstein.Butthatdoessoundexactlyright.

Chief Harrison: Whydon’tyouuseregressionanalysissincecriminologistsseemtouseitallthetime?

Nigel Jewett: Yes,theydo.Andeconomists,andsocialscientists.Butthat’sbecausetheyfocusona

veryparticularsetofmodelsortheoriesabouthowtheworldworks.And,onthewhole,

Idon’tthinktheworldworksthatway.Mostproblemsaren’tjustonethingoroneobject,

dependentonabunchofhigh-levelfactors.Thatseemstobethebasisformostregres-

sionanalyses.Mostproblemshaveatexturebeneath,andalotofpartsandpieces,and

eachofthoseseparatecomponentsallbehavequitedifferently.Andtofigureouthow

the natural world behaves, you have to dive down among the parts.

Chief Harrison: And swim around?

Nigel Jewett: Yes,ma’am.Swimaround.Seewhatthereistosee.Enjoythecolorsandthespectacle.

Chief Harrison: Andwatchoutforthesharks.Andstingrays.What’sthethirdpiece,bytheway?

Nigel Jewett: Oh,newercarsbutnotluxuryones.They’rebeingdriventosomespecificlocationsand

dumped.And,theyhavethebuilt-inGPS4 systems stripped out of them.

Chief Harrison: Nigel,thisisveryinterestingstuff.HaveyoulookedattheHayleyScottcaseatall?

Nigel Jewett: No, ma’am.

Chief Harrison: Doyouhaveherlicenseplatenumber?

Nigel Jewett: Icangetitfromtheregistryfiles,providedithasn’tbeenblocked.

Chief Harrison: Nomoretakingthedatahome,youunderstand?That’snotallowed.Encryptedornot.

Doesn’tmatter.Bringyourcomputerandyourmacrosinhere.Doyoustillhavethedata

for that one month leading up to the murder?

4 Global Positioning Satellite


Nigel Jewett: Yes,ma’am.Onmydatastick.

Chief Harrison: Ihavetwojobsforyou.Canyouwriteupwhatyou’vetoldmeaboutthethreecomponents

ofthestolencarproblem?Lessthanfivepages.Justnotesisfine.I’llwantyoutocome

with me to the senior staff meeting Friday morning, if you can.

Nigel Jewett: Yes,ma’am.WhatIshouldsaytoMr.Goring?

Chief Harrison: Youshouldleavethattome.Theotherthingis,takeafewhours,ifyoucan,andseeifyou

canfigureoutifHayleyScottwasbeingfollowed.OK?

Nigel Jewett: Whichonefirst?

Chief Harrison: HayleyScott.

Nigel Jewett: OK.Bytheway,there’soneotherthingI’dbeinterestedtolookatsometime.

Chief Harrison: What’sthat?

Nigel Jewett: The accident rate in town.

Chief Harrison: Why?Itjustwentdownthislastmonthbyabout10percent,didn’tit?

Nigel Jewett: Yes,ma’am.Itdid.Butlastyearatthistime,itwentdownquitealotmore.Iremember

last year, seeing the press stories about how the police department was so proud of the

reductions,andIrememberthinkingitwastheendofthetouristseason,andtherate

oughttohavedroppedoffalot,justbecauseofthat.Thisyear,thedropisonly10percent.

Lastyear,itwasmorelike35percentmonthtomonth,ifIremembercorrectly.

Chief Harrison: Sure.Takealookifyouhavetime.ButHayleyScottfirst.

Wednesday Afternoon: Dr. Tom Boden, Compliance Monitor for Evidence-Based PolicingFollowingareexcerptsfromameetingbetweenChiefLauraHarrisonandDr.TomBoden.Dr.Bodenisa

criminologistwhohasbeenworkingwithintheHeronCityPoliceDepartmentfortwoyearstoadvancethe

useofevidence-basedpolicing.HeronCityisoneof18cities,spreadthroughouttheUnitedStates,Canada


andEurope,participatinginamajorinternationaleffortfundedbyaconsortiumofphilanthropicfounda-

tionstomoveevidence-basedpolicingfromtheacademicliteraturetotheworldofpolicepractice.Underthe

termsoftheprogram,eachcitygetsaqualifiedcriminologistworkingfulltime,onsite,foraninitialthree-

yearperiod.Dr.BodenwaspostedtoHeronCity,andhissalaryispaidfromgrantfunds.Thedepartmentalso

receivesagrantforaccommodatingandsupportingtheworkoftheircompliancemonitor.

Chief Harrison: Veryhappytomeetyou,Dr.Boden.Ireadoneofyourearlypapers,sometimeago,when

Iwasstudyingforoneofmypromotionexams.

Dr. Boden: Likewise,chief.Happytomeetyou.AndwelcometoHeronCity.I’mnotsureyourpre-

decessorandIalwayssaweye-to-eye.I’llcertainlydoeverythingIcantomakesurewe

makethisdepartmentjustaseffectiveasitcanpossiblybe.

Chief Harrison: Iappreciatethat.ButI’mcurious.Whatdoesyourtitle“compliancemonitor”actually

mean?

Dr. Boden: I’mnotsureIlikethetitleallthatmuch.Buttheideais“evidencecop,”orsomething

likethat.ItcomesfrompapersbyProfessorLarrySherman,andhesaystheroleofthe

evidencecopis“toredirectpracticethroughcomplianceratherthanpunishment.”Soit

isactuallysupposedtosignifyacloseworkingrelationshipasopposedtoanadversarial

one.Isupposeitcouldsoundasif“compliancemonitor”meantsomethingadversarial,

butthat’snotthewayIinterpretit.

Chief Harrison: I’mhappytohearthat.Because,Imustsay,Ifindtheideaofanacademicresearcher

“redirecting”apolicechief—eitherthroughpunishmentorcompliance—ratheramus-

ing.Butwhosecompliancearewetalkingabout,justsoI’mclear?Mine,Isuppose?

Dr. Boden: No,ma’am,notyours.Itisaboutthedepartment’suseofvariousmethods.Thewhole

ideaistomovethedepartmentfromchoicesbasedonassumptionsandguesswork

towardchoicesandmethodsbasedonfactsandproventoworkthroughtheuseofproper

scientificmethods.

Chief Harrison: Whatkindofscientificmethods?

Dr. Boden: Empiricalresearch.Theideaisthatpolicingshouldbeassolidlygroundedinscientific

researchasthepracticeofmedicine.Sowe—andwhenIsay“we,”Imeanthecommunity


ofprogram-evaluationexpertswhoareparticipatinginthisprogram—useonlythevery

best studies, and even then we normally require that there be at least two or more studies,

in different locations, that show a program to be effective before we accept that it really

is.Notjustanystudieseither,butstudiesofthehighestquality.

Chief Harrison: OK.Butwhatmakesonestudyofhigherqualitythananother?Andwho’stojudge?

Dr. Boden: Well,alotofworkhasbeendoneinthatarea.Themostreliableevidencecomesfrom

randomized controlled experiments, where you can really isolate the effect of an inter-

vention.Someformsofquasi-experimentaldesignrankprettyhighly,too,providedthe

researchersinterpretingthemknowaboutanddealwithalloftherelevantinference

problems.Intermsofjudging,wearemovingtowardafairlybroadinternationalagree-

ment about how many studies you need, and of what quality, before the effectiveness of

a particular program should really be accepted.

Chief Harrison: Soundslikeaprettyhighstandard!Andalotofwork.Howlonghasthisbeengoingon?

Bytheway,howmuchpoliceexperiencedoyouhave?

Dr. Boden: I’veneverworkedasapoliceofficer.ButIhaveworkedonpolicedataandcrimedataever

sincemyPh.D.Mythesiswasontheimpactoflead-paintexposureinearlychildhood

anditssubsequentimpactsondelinquency.Weknowthistypeofresearchtakesalong

time, and one of the ways we are trying to accelerate the process is by reviewing stud-

ies already completed by others and, if they’re done well enough, we can combine their

findingsthroughmeta-analyses.Thatgivesusevenmorerobustfindingsandsavesthe

need to set up brand new experiments.

Chief Harrison: Whatkindsofprogramshaveactuallypassedthistest?Andhowmanyofthemare

there now?

Dr. Boden: TheNationalCrimePreventionCentre,inCanada,publishedasummaryin2007.They

listed14typesofprogramsthatcouldnowbeaccepted,thatis,proveneffectiveinpre-

venting crime.

Chief Harrison: Just14?Canyougetmethatlist?I’dlovetoknowwhattheyare.


Dr. Boden: Ihaveithereinmycase.[Handsoverdocument5 and points to listing on page three.]

Chief Harrison: Isee.Sofamily-basedpreventionisonecategory,andthatcategorycontainsfiveapproved

programs.Isthatright?Includinghomevisitation,daycare,parenttrainingwithyounger

childrenandsoon.Thesedon’treallysoundlikepolicebusinesssomuch.

Dr. Boden: Yes,ma’am.Buttheyarecrime-preventionprograms.Fiveofthemapprovedunderthe

family-basedpreventioncategory,threeundercommunity-basedprevention,three

underschool-basedpreventionandthreeunderplace-focusedprevention.

Chief Harrison: Whataretheplace-focusedones?Maybetheysoundmorelikepolicework.

Dr. Boden: Thesearetheplace-focused,crime-preventionprograms.[Hepointstodocument.]They

arenuisanceabatement,closed-circuittelevisionsurveillancecamerasandimproved

street lighting.

Chief Harrison: Soyouhavejust14programsintotalthatyou’veestablishedactuallywork.Isthat

right?

Dr. Boden: Ithinkacouplemorehavebeenaddedsince2007,butroughlythatnumber.Wecould

acceptmore,butweneedtokeeptheresearchtothehighestpossiblestandardssowe

don’tmakemistakes.Thewholeideaistodirectpublicfundstothoseprogramsthatwe

knowforsuremakearealdifference.TheCanadiansusedaveryhighstandardforthis

listing.Theysayhere,onpagefive,thataswellasusingproperexperimentalorquasi-

experimentaldesigns,thestudiesshouldalsoinclude“largesamples,longfollow-up

periods,follow-upinterviews,andprovisionforeconomicanalysis.”6

Chief Harrison: IfIhadtoguess,I’dsaythatthisdepartmentprobablydoesatleastathousanddiffer-

ent things each day, all supposed to contribute to crime prevention. Are you saying we

shouldn’t be doing the other — however many it is — 986?

Dr. Boden: Well,theselistedherearemajorcrime-preventionprogramsthatarerunacrossbroad

segmentsofthepopulation.Notjustordinarydailydecisions,likewhatthebeatofficer

shoulddonext.Thisisdesignedtohelppoliceandsocietymakethebigcrime-prevention

investmentsonthebasisofscienceratherthanguesswork.

5 See appendix: Evidence-Based Crime Prevention: Scientific Basis, Trends, Results and Implications for Canada, by Brandon C. Welsh. 6 Ibid., p. 5


Chief Harrison: Istherereallynothinginbetween—betweenyourkindofscienceandguesswork?It

sounds as if it has to be high science, complete with enormously expensive randomized

experiments,armiesofPh.Dsand,Iassume,ahugeamountofmoney.Andifit’snot

that,thenit’switchcraft,ormaybejuststupidity?Don’tpolicemanagersdoanawfullot

of perfectly sensible stuff that you haven’t even had a chance to consider yet? And don’t

theyconfrontproblemsthatnobody’sseenbefore,wheretheyjusthavetotrywhatseems

bestatthetime?Howlongdoesittake,fromthetimeanewproblemappearstothetime

thatyourcommunityofscholarsactuallydecideswhatworks?Mustbeagoodfiveyears

orso?Whathelpisthatforoperations?

Dr. Boden: It’ssurelybettertoknowwhatworks.Otherwise,we’djustkeeppouringpublicmoney

down the drain.

Chief Harrison: Don’tyouthinkpoliceknowanythingwithoutyourhelp?Hey,waitaminute.I’mlooking

atthislist.DoyouknowwhatIdon’tseehere?Catchingmurderers.Doyouthinkcatching

murderersisimportant?Ifwecontinuethisveryhigh-profilemurderinvestigationthat

wehavenow,whichistearingupourcityatthemoment,willIgettobe“noncompliant”

inyoureyes,justbecausearrestingmurderersdoesnotappearonthislist?Haveyou

evaluatedcatchingmurderersyet?Whataboutmeetingwiththemayor?Canyoutellme

theoptimalnumberoftimesIoughttomeetwiththemayorpermonthtoreducecrime

tothelowestpossiblelevels?Doyouhaveequationsforthat?

Dr. Boden: Wehaven’tlookedatthosethings.No.Amurderonlycountsforoneinthereportedcrime

figures.Itprobablywouldn’tfigureverymuchinthetypesofanalysiswenormallydo.

Chief Harrison: So,whatmakesthings“figure”inyourtypeofanalysis?Publicoutrage?Politicalpres-

sure?Dreadfulfearwreckingpeople’slives?What?

Dr. Boden: The things that matter in our analyses generally start off as factors or programs that end

upwithastatisticallysignificantcoefficientinaregressionanalysis,orasignificanteffect

under hypothesis testing of some sort. The dependent variable is usually the crime rate or

somepieceofit.Econometricmethodspointustothefactorsthatreallymakeadifference.

Andthen,whenwearereadytoevaluateaspecificprogram,wetrytosetupaproperly

constructed experimental design so we can measure the impact on crime rates of a spe-

cificinterventionandbereallysureitistheintervention,andnotsomethingelse,thatis

makingthedifference.That’sthestagewhenwetrytoestablishacausalconnection.


Chief Harrison: Alright.Justsoyouknow,Ineverdidaregressionanalysisorranarandomizedexperi-

mentinmylife.AndI’msorryyoudon’thaveastudy,orhaven’tdoneone,thatshowsthat

investigatingmurdersisactuallyimportant.ButI’mgoingtoseethatmurderinvestiga-

tionthroughanyway,ifthat’sokaywithyou.

Dr. Boden: You don’t need my permission, ma’am. You’re the chief.

Chief Harrison: Thismakesmewonder.Maybeyourrole,withallthissophisticatedresearchatyour

disposal,shouldbetoinformmeaboutthingsthatmightworkandthatIwouldn’thave

thoughtofbymyself.Howaboutthat,asabasisforourrelationship?Iguessthatmight

seem more reasonable to my operational managers than the idea that you’d prohibit

themfromdoingthingsjustbecausescholarshiphadn’tgotaroundtoevaluatingthem

yet.Butletmeaskyouonemorething:HaveyoureadGoldstein?

Dr. Boden: Yes, ma’am. Of course.

Chief Harrison: Doesn’tGoldstein’sapproachsuggestthatweshouldtackleproblemsonce,makingsure

thatwedeviseasolutionthatnotonlysuppressestheminaneffectiveandresource-

efficientmannerbutalsokeepsthemsuppressedforagoodlongtimeandwithminimal

subsequent effort?

Dr. Boden: Yes.Iguessthat’safairdescriptionoftheproblem-solvingapproach.

Chief Harrison: Itstrikesmethatthisveryshortlistofprogramsthatyouandyourcolleagueshaveactually

approvedforusebypolice—suchasdaycare,orafter-schoolrecreation—arepro-

grams that once you’d started, you’d probably continue forever and at enormous expense.

Presumablyyouhavetocontinuetheseprogramsallthetime,onceyou’vestarted,and

foreachsuccessivecohortandeachgenerationofmight-bedelinquents.Isthatright?

AreyouandGoldsteintalkingaboutthesamekindofapproach?

Dr. Boden: Well,theclosestthingtoGoldstein’sideathatmostpolicedepartmentsactuallydois

place-basedpolicing.Policetendtodohot-spotanalysesandplace-basedinterventions.

Sowehaveevaluatedsomeofthosemethods,andthat’swhysomeoftheseplace-based

ideas,likecamerasandstreetlighting,appearonthislist.Thenuisanceabatementidea

islinkedcloselytothebrokenwindowstheory,andthat’sbeenprettywellvalidatedover

theyears.SowebelievewehavevalidatedGoldstein’sapproach.


Chief Harrison: SoyoutestedGoldstein’sapproachbyfindingoutthatplace-basedproblems,likehot

spots,tendtohaveplace-basedsolutions.Isthatright?

Dr. Boden: Webelievethathasbeendemonstratedinenoughplacesnowsowecanbefairlysureit

works.Yes.

Chief Harrison: Isn’tthatratherobvious?Couldn’twehaveguessedthatplace-basedproblemswould

tendtohaveplace-basedsolutions?Doesn’tthelikelyshapeofthesolutiongowiththe

shapeoftheproblem?Imean,ifitwasafamily-shapedproblemratherthanageographic

hotspot,wouldn’tyouexpectafamily-centeredsolutionofsomesort?

Dr. Boden: Thewholepointisthatweshouldn’tberelyingonguesswork.

Chief Harrison: Bytheway,doesn’thot-spotanalysisbypolicepredateGoldstein?IthoughtGoldstein

startedoffbycomplainingaboutjusthownarrowthepoliceprofessionhadalwaysbeen

in the types of analysis they used. The police always did hot spot analysis, even when it

waswithpin-mapsontheofficewall.IrememberGoldsteinstartsoffoneofhisbooksby

listingabout13otherdimensionsinwhichweoughttobeabletoidentifyproblems:like

repeat victims, repeat offenders, modus operandi, criminogenic commodities, patterns

of organized crime and so on.

Dr. Boden: Ithinktheintelligence-ledpolicingsystemfollowsupontherepeatoffendersidea.And

wehaven’thadachancetoevaluatethatapproachyet.Thatoneisdifficulttoevalu-

atebecausedifferentdepartmentsactuallyimplementintelligence-ledpolicinginsuch

different ways.

Chief Harrison: ButwhathaveyouandyourfriendsdoneaboutGoldstein’sideas?Itsoundstomeasif

you’vetestedGoldsteinbytestingtheonethinghecomplainedaboutasbeingtheonly

thingthatpoliceknewhowtodo?IfIwasHermanGoldstein,Idon’tthinkI’dbethrilled

about that!

Dr. Boden: Ithinktheworkdoneonplace-basedapproacheshasbeenprettythorough.Prettyhigh

quality.Wecansay,forsurenow,thatsomeofthesemethodsactuallywork.

Chief Harrison: Fromwhatyousay,itseemstomeyou’vevalidatedsomeofthemostlimitedandsimple-

mindedformsofGoldstein’sapproach.Isupposeyoudohavetofocusonthemethods


thatenoughpolicedepartmentshaveused—likeimprovingstreetlights—andina

sufficientlystandardizedmannerthatyoucanactuallyperformyourmeta-analyses.

Soyouenduplookingatstandardizedtools,generallyunderstoodeverywhere.Didn’t

Goldsteintalkabouttailor-madeinterventionsforcarefullyidentifiedproblems?He

talkedalotaboutthefactthatrealproblemshaveuniquefeaturesand,fromwhatIread,

I’dsayhestressedthatyouhadtounderstandtheuniquenessofaproblemtohaveany

realchanceofsuccess.Idon’tthinkheassumedthatatleastthreeotherdepartments

musthaveusedthesamesolution,exactlythesameway,beforewecanknowanything

about effectiveness. He was all in favor of brilliant, surgical, locally invented interventions,

wasn’the?Isn’tthatwhattheannual“Goldsteinproblem-solvingawards”aresupposed

to celebrate?

Dr. Boden: Theevidence-basedtraditiondoesrecognizethatcontextsvary,andofcourseyouhave

totakethatintoaccount.So,onceaprogramisacceptedaseffectiveandapprovedforuse

in crime prevention, then we do understand that the program may need to be tailored,

tosomedegree,ifyoudecidetouseitinanotherjurisdiction.

Chief Harrison: Butyouhaven’tansweredmyquestion.Goldsteinstartswithproblems.Youstartwithpro-

grams.Goldsteinsaystheproblemsareoftenuniqueandrequirecarefulstudy.Healso

saysyoufirstneedtodisaggregatetheoverallcrimerateintoindividualandactionable

pieces.Isn’tthatwhattheydoinmedicine?Inevidence-basedmedicine,don’ttheystart

with a particular disease or condition? And don’t they also start with a careful description

orprofileofthepatient,withallofthepatient’sriskfactorsandhistory,andsoon?Isn’t

thatlikeGoldstein’sratherpreciseproblemdefinition,whichhesaysyoumustdofirst,

beforeyouevenbegintothinkaboutsolutionsormethodsorprograms?

Bycontrast,youseemtostartwithmethods.This[holdsupEvidence-Based Crime

Prevention]isalistofprograms.Shouldn’tIbelookingatalistofspecificproblemsfirst?

What’sthediseasethattheseprogramsaresupposedtocure?Idon’tseealistofproblems

here,exceptforcrimeingeneral.Maybedelinquencyingeneral.

Dr. Boden: I’llhavetothinkaboutthat,ma’am.Ithinkitismostlyadifferenceofsemantics.Ithink

we all want the same thing.

Chief Harrison: What’sthat?


Dr. Boden: Tobescience-based.Forpolicingtobebasedonfacts.Notonsuperstition,orguesswork

oruselesstraditions.AndIdon’tmindifyou’dpreferadifferenttitle,somethingother

thancompliancemonitor.I’dbehappytothinkofmyselfasyourchiefscientistinstead,

ifthatmakessensetoyou.

Chief Harrison: Dr.Boden,I’mnotsureIneedachiefscientist.IthinkIjustneedsomesensibleanalysts,

whocanactuallyhelpmegetthejobdone.Let’ssee.CanyouhelpmesolvetheHayley

Scottmurder?

Dr. Boden: No.Idon’tthinkso.

Chief Harrison: Can you tell me what might prevent or reduce our rash of car thefts — brand new, luxury

class vehicles — and explain why they seem to be disappearing into thin air at unprec-

edented rates and only in this city? Can you offer any insight, from your studies, as to

where they might be going?

Dr. Boden: I’mnotawareofanyworkspecificallyontheftsofluxurycars.Iguessthethingthatmight

helpistousesomeoftheseplace-basedapproachesthathavealreadybeenproven.We

doknowthatCCTVactsasafairlygoodpreventivemeasure,althoughitdoesseemto

workbetterinEnglandandWales,forsomereason,thanitdoesintheUnitedStates.

Chief Harrison: [Laughs].MaybetheUnitedStatesisbigger,sothingsaremore spread out. Oh, excuse

me.I’monlyguessing.That’snotallowedanymore,isit?

Dr. Boden: Are these car thefts concentrated in hot spots at all?

Chief Harrison: Absolutely not, according to my chief of operations. They’ve been mapping them for a

good long time, and they seem to fall all over the city. There are many, many tiny clusters,

butnobigclusters.Therearealmostasmanyclusters,now,asthereareparkinglotsin

the city. And that’s several hundred. No way could CCTV coverage be affordable.

Dr. Boden: Wecouldcutthecostinhalf,ifyoulike.Wecouldputthecamerasupinhalfofthepark-

ing lots, rather than in all of them. Then we could really see what difference they made.

Mightbeareallygoodexperiment.I’dbeveryhappytohelpyouwiththeexperimental

design.


Chief Harrison: Any other ideas, apart from running experiments so you can get your name in print?

Dr. Boden: Whataboutimprovingthestreetlighting?That’sbeenshowntoworkinmanydifferent

cities.

[ChiefHarrisonlooksupbutdoesnotreply.]

Dr. Boden: Iguess,ifthethievesturnedouttobehighschoolers,thentheafter-schoolrecreation

programs might help?

Chief Harrison: ThankyouDr.Boden.Letmeshowyouout.Iappreciateyoucomingup.

ChiefHarrisonstartstoleadDr.Bodentowardtheofficedoor,buthalfwaythereshestopssuddenly,turns,and

(beingquiteaccomplishedinmartialarts)deftlystrikesDr.Boden’snosewiththebaseofherflat,openpalm.

Nothard,butaccurately,andjusthardenoughtomakehisnosebleed.Dr.Bodenlooksstartledandhorrified,

andfumblesforhishandkerchiefasbloodbeginstotrickledownhisface.ChiefHarrisonsmilesathimkindly

andhandshimaboxoftissuesthatwasontheedgeofherdesk.

Dr. Boden: Excuseme,ma’am.Whatwasthatabout?

Chief Harrison: Oh,I’msorry.Justamodestlittleinterventionofmyown.

Dr. Boden: You made my nose bleed!

Chief Harrison: No,no,Idon’tthinkso.JustbecauseItouchedyourfaceIdon’tthinkyoucanjumpto

theconclusionthatitwasIwhomadeyournosebleed.Onthecontrary,there’sastorm

coming,andIassumeitwasthedropinbarometricpressure.Maybeyoualwaysget

nosebleedsthistimeofdayorinthistypeofweather.Ireallydon’tthinkithadanything

to do with me.

Dr. Boden: Butyouhitme.Andnowit’sbleeding.It’sobvious.

Chief Harrison: Surelyyoucan’tpossiblydeduce,basedonjustasingletrial,andwithnocontrolsample

ofanykind,thatmystrikingyouactuallycausedyournosetobleed.Wouldthatkindof

claimmakeitthroughyourfiltersforhigh-qualityscientificevidence?OrtheCanadians’


evenmorerigorousstandards?Wherearethelargesampleandtheextensivefollow-up

interviews?

[Dr.BodenlooksatChiefHarrison,bemused.Hemovestowardthedoor.]

Chief Harrison: IfwereallywantedtoknowforcertainthatImadeyournosebleed,wouldn’twehave

toconductasignificantnumberoftrials?Wecoulddothisagain.Isupposeitwouldbe

good if we could get some of the other compliance monitors also, to stand right next to

you.Andwe’dhitsomeofthem,butnotothers.Mayberandomselectionwouldhelp?

Whatdoyouthink?Shouldwedoarealexperiment?

Thursday Morning: Nigel Jewett, Junior Analyst, IT DepartmentWhenChiefHarrisonarrivedatherofficeat8a.m.onThursday,NigelJewettwassittinginthewaitingarea

outside.

Chief Harrison: Nigel, good morning. Come right in.

Nigel Jewett: Thankyou,ma’am.Thiswasfun.

Chief Harrison: Howlongdidittakeyou?

Nigel Jewett: Mostofthenight.ButIthinkitwasworthit.

Chief Harrison: Sotellme.Whatdidyoudo?

Nigel Jewett: Anumberofthings.Wedohaveanaccidentproblem.AndIdidsomeworkontheScott

case, too.

Chief Harrison: Whatkindofaccidentproblem?

Nigel Jewett: Thanks,first,forgettingmetheaccidentdata.OnceIhadthatinExcel,itwasprettyeasy

tofigureoutwhatwashappening.Ijustlookedtosee,monthbymonth,whatsegment

of the accident data was not following the aggregate trend downward.

Chief Harrison: And was there one?


Nigel Jewett: Yes.Itseemstobewomendrivingalone.Andnotalwaysalone,butwithnootheradult

inthecar.Sometimestheyhavetheirkidsinthecar.

Chief Harrison: And what’s happening to them?

Nigel Jewett: Theyjustrunintothings.Thecarinfront.Afewintotreesorotherobjects.Justseems

likethey’renotpayingattentiontotheirdriving.

Chief Harrison: So?Whatdoyoumakeofthis?Arethemenbeingequallycarelessallofasudden?

Nigel Jewett: No.Themen-drivingnumbersarebehavinglikeeverythingelse—droppingoffbytotal

volume.IthinkitmightberelatedtotheScottcase.Maybewehavejusttoomanywomen

worryingaboutbeingfollowed.Maybethey’reallspendingtoomuchtimestaringinto

theirrear-viewmirrors.That’stheonlyreasonIcouldthinkof.

Chief Harrison: Areyousureaboutthetrend?Iknowyoudon’tmuchcareforprobabilitytheory,butcould

thisjustbeafluke?Isthisstatisticallysignificantatthisstage?

Nigel Jewett: Probablynot.ButIthoughtyou’dwanttoknowaboutitbeforeitgottobestatistically

significant.Itisquiteaparticulargroupingofaccidents,onceyouseeit,anditincludes

two fatalities already this month. That’s twice as many as the murder itself.

Chief Harrison: WhatabouttheHayleyScottcase?Whathaveyoudoneonthat?

Nigel Jewett: Withherlicenseplatenumber...first,IranthesimpletestontheALPRdata.Itook30

secondsasthefollowingintervalworthchecking.IstartedwithHayleyScott’scarnum-

ber,atleasttheoneregisteredtoher,andsheshowsup403timesinthemonth.Imerged

thelistsforallthecarsthatpassedwithin30secondsafterwardsandlookedformultiple

hits.

Chief Harrison: And what did that show you?

Nigel Jewett: Notmuchthat’sinteresting.Herhusband’scarjumpsout,followingherseventimes.And

therearelotsofcarsthathitthreetimes.Butnofours.


Chief Harrison: Thatmarriesupprettymuchwithwhatthecontractorfound,Ibelieve.

Nigel Jewett: Iftherereallywasastalker,andonlyone,thenI’dexpecttobeabletofindthemmuch

morecleanly.I’dexpecttofindawayoflookingatthedatatomakethemreallystickout,

if they were there at all.

Chief Harrison: Soyoudidn’tstopthere,right?

Nigel Jewett: No,ma’am.Ididacoupleofotherthings.Thefirstwastoallowfortheerrors.HayleyScott’s

numberisDBR862.SoIalsowantedtotryallofthemisreadingpossibilities.Thatcould

bereadas882,or682or662.Soonethingwastobroadenthesearchandcoverallfour

possibilities.ThatbumpedupthenumberofHayleyScottsightings,potentially,to432.

Chief Harrison: Whatelse?

Nigel Jewett: ThenIwonderedaboutthehusband.Notasasuspectsomuch.Butwhatwasittheywere

doing?Seventimeshewasbehind,andfourtimeshewasinfront.Butstillcloseinfront.

SoIwonderedwhatthatmeant.Ifthiswasmywife—andIdon’thaveoneyet—butifI

didandshethoughtshewasbeingfollowed,I’mprettysureI’dfollowrightbehind,notin

front.Soiftheyweredoingthat,itmeanstheyswapcarssometimes.Whichmakessense,

ifyou’vegottwodifferentsizesofcar.Dependswho’scarryingthetrashtothedump,or

who’sdrivingthekidsaround,right?SoIwonderedwhathappensifweassumeshealso

drives her husband’s car around town on a regular basis. And his car shows up 320 times

in the month.

Chief Harrison: Sonowyoucombineallthesethings?

Nigel Jewett: Exactly.Husband’snumberdoesn’thaveanysixesoreights.SoIdecidedtotreathis

number, and all four of her possibilities, as a single entity and see who might be follow-

ingthatentity,whicheveroneofthefiveformsittook.Thatgaveme432plus320,which

is752possibleHayleyScottsightings.Imergedthoselistsandstuckto30secondsfor

theinterval.TheScotts’owncarscomeupamongstthefollowers,asyou’dexpect,with

scoresofnineand13occurrences,followingbehindtheother.Ijustignoredthose.There

was still a big pool of threes. And a single car with a score of nine.


Chief Harrison: Nine?Sowhatexactlydoesthatmeanforthisonecar?

Nigel Jewett: ThatoneisloggedasbeingclosebehindHayleyScott’scar,asDBR862,twiceinthemonth.

Itisbehindhercar,whenherplateisreadasDBR882,threetimes.Iguessshehadadirty

licenseplate.Anditisloggedasbehindherhusband’scarfourtimes.It’senough.Itsticks

outlikeasorethumb.Thenexthighestscoreisthree.Thisisnine.

Chief Harrison: Doyouhavethenumber?

Nigel Jewett: Yes,ma’am.[Hepassesoveraspreadsheetprintoutwith18rowsshowingtheninesight-

ingsandninerelated“followings.”]Icouldhavetakenthisonelevelfurtheranddone

thepotentialcorrectionjobforallothercarshavingsixesoreights,or“I”sor“1”sintheir

plates.

Chief Harrison: How would you have done that?

Nigel Jewett: Iwouldhavecreatednewrecordsforallofthepossiblemisreadingsforallcarsinthe

database.Actually,Ididcreatethemacrosforthat,inExcel.Butitwasn’tnecessaryand,

in this case, it probably wouldn’t change the result.

Chief Harrison: Whynot?

Nigel Jewett: Becausethevehiclenowofinterestdoesn’thaveanyofthosecharactersinitslicense

plate.PYN447.Soyouwouldn’tincreasethescore.Andthattypeofcorrection—putting

backinpotentiallymissingdata—canonlyincreasethescoresifyou’reusinganykind

ofadditivescoringfunction,likeIwashere.

Chief Harrison: Thisisprettyimpressivestuff,Nigel.Soundslikeacomplicatedprocess.Asalgorithms

go, was this a tough one?

Nigel Jewett: No,notreally.Usuallyittakesawholesequenceofstagestogettoanythinguseful.We

callthese“multipassalgorithms”or“multistage”ones.Whatcomesoutofeachstageacts

asthestartingpointforthenextstage.Youfilterthisway,thensortthatway,thenscore

whatyousee.Andthen,usuallyyoudon’tknowwhattodonextuntilyoustareatwhat

youhave,atleastforawhile.Soyoucan’tspecifythewholeprocessupfront.That’swhy


youneedthekindsoftoolswhereyoucanseewhatyou’redoingandwhatyou’regetting,

andwhyyoutryalotofdifferentideasprettyquicklyuntilsomethingjumpsoutatyou.

Chief Harrison: Whatmakessomeonegoodatthiskindofresearch?

Nigel Jewett: Mymarinebiologyprofessorusedtosay“abandoninguselessalleywaysquickly.”You

know,getoutofthemandintoanotherone.Nottoofast,ofcourse.Butnottosticktoo

long to any one analytic idea when it is not showing you anything useful.

Chief Harrison: Nigel,listentome.Thankyouforthis.Thismayturnouttobeveryimportant.ButIhave

61otherwomennowwhoareworriedtheyarebeingfollowed.Maybebythissamecar,

PYN447.Maybebyanotherguy.Maybebynobodyatall.Canyoudothisoveragain,61

times,foreachofthem?Didyousaveyourmacros?

Nigel Jewett: Iwasthinkingaboutthat.Yes,ofcoursewecandothesamethingfor61othernumbers.

Butthat’snotthebestthingtodo.

Chief Harrison: Whynot?Whatwouldyoudo?

Nigel Jewett: Whynotaskthebiggerquestion?Isanyonefollowinganyone?Maybepoliceoughtto

knowifanyoneisfollowinganyone.Maybesomeone’sbeingfollowedandisthenext

murdervictimandhasn’tevenbecomesuspiciousyet.Wouldyouwanttoknowthat?

Chief Harrison: You could do that?

Nigel Jewett: Ithinkso.Thereisoneextradifficulty,though,whenyoudon’thaveanyparticularplace

tostart,likeasuspectoravictim,andyouhaveverylargedatabases.Youhavetothink

carefullyaboutallthepermutationsandcombinationsyouneedtotry.AndIcheckedthe

volumesoverall.Wehaveroughly300,000uniquecarseachday,withmultiplesightings

ofmostofthem.Inamonth,closeto2millioncarsseenatleastfivetimeseach.Ifyou

startedwithjustthose2million—leavingoutthedrive-through-onceset—thenyou

wouldwanttoknowtherelationshipbetweenanyoneofthose2millioncarsandany

other one of those 2 million cars. You’d want to see which pairs have the most frequent

linksbetweenthem.You’ddefinetwocarsas“linked”iftheypassedanyALPR,anytime,

lessthansomanysecondsapart.Theintervalisasearchparameteryoucouldadjust,if

you needed to control the volumes of the output or change the sensitivity of the search.


Chief Harrison: Logically,thatsoundsalotsimplerthanwhatyou’vealreadydoneinthiscase.

Nigel Jewett: Logically,itis.Butcomputationally,itisn’t.Youmightstartoffwithamatrixorarray,2

millionby2million,andstartloggingthelinksinthearray.JustworkthrougheachALPR

historyonebyone,creatingallthelinks.Thencountthemupandlookatthebiggesttotals.

Butonmostmachines,unlessyouhavemassesofinternalmemory,youcan’tdeclare

arrays2millionby2million.Andifyoucan,youcan’tdoanythingwiththem.Itispretty

easy to blow up your machine, not literally of course, but grind it to a miserable halt by

writing a thoughtless algorithm that will produce billions and billions of possibilities.

Chief Harrison: Sowhatcanyoudo?

Nigel Jewett: For those types of problems, you have to get crafty and creative about the actual search

methodevenafteryou’vealreadydecidedwhatyou’relookingfor.Soyoufindwaysto

chopouthalfofthedata,hopefullywithoutloss.Ortolookfirstamongsubsetsathigher

risk.Andifthestructureyouwanttofindisoneofthosethatisnotatallrobusttomiss-

ingpiecesofdata,thenyousimplycannotthrowanythingawayor,forallyouknow,the

patternmightevaporatebeforeyoucanfindit.

Chief Harrison: So,whatthen?

Nigel Jewett: Forthose,youhavetouseothercomputationaltechniques,likesparsematricesand

linked-listprogramming,thatcutdowntheamountofmemoryspaceyouneedbywork-

ing only with the nonzero entries. Or use massively powerful computing technologies.

LiketheyhaveatLosAlamos,Iguess!Iwonderhowmanycredit-cardtransactionsthere

areeverydayontheglobalsystemsandwhatstructuresyourbrother-in-lawendsup

searching for. That must be fun, too!

Chief Harrison: Nigel,I’mintriguedbyyourideaoffun.I’dlikeyoutodomoreofthistypeofwork.Ijust

havetofigureoutwheretoputyouexactly,andhowthistypeofworkfitsinwithallthe

other things we have to do here.

Thursday Afternoon: Telephone Call With Dr. Tom Boden, Compliance Monitor for Evidence-Based PolicingFollowingareexcerptsofatelephonecallplacedbyChiefHarrisontoDr.TomBodenat2p.m.onThursday.


Chief Harrison: Dr.Boden.I’mhopingwecanworkcloselytogether.AndIpromisenottoslapyourface

anymore.

Dr. Boden: Ma’am.I’mdelightedtohearthat.Ihopeso,too.

Chief Harrison: I’minterestedinsettingupanewanalyticalunittosupportoperations.I’mnotsureif

it’sthetypeofthingyou’dliketodo,orwhetheritservesyouragenda.ButIwouldneed

sometechnicallyverywell-qualifiedpeopletorunitandworkinit.

Dr. Boden: I’mhonored.Itsoundslikeagreatidea.ItisexactlythetypeofengagementIhadhoped

forinthisdepartment.Whatsortofstaffdoyouhaveinmind?

Chief Harrison: PeoplelikeNigel.DoyouknowNigelJewett?

Dr. Boden: Yes,IknowNigel.

Chief Harrison: Couldyouworkwithhim?Andpeoplelikehim?Maybewecouldgetsomeofhisfriends

or friends from college in here, too?

Dr. Boden: I’mnotsurethat’swhatyouneed,ma’am.Iactuallydon’tthinkNigelisqualifiedtobe

an analyst.

Chief Harrison: No,whynot?Whatmakesyousaythat?Heseemsquitebrighttome.

Dr. Boden: Maybe.Buthedoesn’tknowthefirstthingaboutprogramevaluation.

Chief Harrison: OK,Iunderstand.Thanks.Don’tworry,I’llfindsomeoneelse.

Dr. Boden: Ma’am,I’dbehappytohelpyoufindsomeoneelse.Icouldbringinsomereallygood

graduate students who would be really good for the department. And they’d be well

qualified.Verywellqualified,Icanpromiseyou.

Chief Harrison: Idon’tthinkyouunderstoodwhatImeant,Dr.Boden.I’mkeepingNigel.


Friday Morning: Nigel Jewett, Junior Analyst, IT ServicesAt10a.m.onFriday,NigelJewettcomestothechief’sexecutivesuiteandasksthechief’sexecutiveassistant

if he can see Chief Harrison. Nigel says it is urgent. Following are excerpts from his subsequent conversation

with the chief.

Nigel Jewett: I’verunit.

Chief Harrison: You’ve run what?

Nigel Jewett: Who’sfollowingwhom.Ifoundawaytodoitonmycomputer.Andmycomputerisnow

inmyoffice,ma’am.Ithoughtyou’dbepleasedtoknowthat.Iactuallyhadtoprogram

itinC++usinglinkedlists.Butitworked.

Chief Harrison: So,who’sfollowingwhom?

Nigel Jewett: There are three women. They are all being followed by the same car, but on different days

oftheweek.Onlyoneofthemisinthegroupof61thathavereportedanythingodd.The

other 60, by the way, have nothing to worry about.

Chief Harrison: Thesamecarasyoufoundformetheotherday?Wearrestedthatguyasasuspect!

Nigel Jewett: Notthatone.Anotherone.Butthesamecarforallthreewomen.Andthat’snotall.

Chief Harrison: Heavens.Whatelse?

Nigel Jewett: You’re being followed.

Chief Harrison: I’mbeingfollowed!Bywhom,forheaven’ssake?

Nigel Jewett: Twocars.Theyworktogether.Normallythere’soneinfront,close,within15seconds,and

onebehind,thatcanbeupto45secondsbehind.Theyswitchplaces,too.It’snotalways

the same way around.


Chief Harrison: Whichcararetheyfollowing?

Nigel Jewett: Yourofficialcar,sometimes.Butmoreoftenyourprivatecar.Eveningsandweekends.

Thisgoesbackthreeweeks.Itstartedtwoweeksbeforeyouarrivedhere,ma’am.

Chief Harrison: Butaftermyappointmentwasannouncedinthepress,right?

Nigel Jewett: Isupposeso.

Chief Harrison: Whoisfollowingme?

Nigel Jewett: One car is private but registered to a lieutenant in the department. And the other is an

unmarkedcarregisteredtotheDrugSquad.

Chief Harrison: Who’dhavethoughtit?

Nigel Jewett: [Pause.]It’sfunnysometimes,whatyoudiscoverbyaccident.LikeMarieCurie.When

shediscoveredradium,shewasn’tlookingforit,oranythinglikeit,atall.

Chief Harrison: Butshewaslooking,andshewasascientist.Sosheknewhowtolook.Whoknows,maybe

she was having fun, too!

Nigel Jewett: Whywouldtheybefollowingyou,ma’am?

Chief Harrison: That’sanyone’sguess,atthispoint.Maybewefoundsomesharks,andweweren’teven

lookingforthem.TheyprobablywanttoknowifandhowImightbecompromised.


EVIDENCE-BASED CRIME PREVENTION: SCIENTIFIC BASIS, TRENDS, RESULTS AND IMPLICATIONS FOR CANADA

The State of Science on What Works to Prevent Crime

Research on “what works” in preventing crime has long been of interest to practitioners, policy-makers, and academics alike. Only in recent years, however, has there been an increased effort to improve the trustworthiness of claims about what works in preventing crime. This has come about through the use of the highest quality scientific evidence and the most rigorous and transparent review methods to assess what works. It has come to form the state of science on evidence-based crime prevention.

Family-Based Prevention

Five types of family-based programs have been found to be effective in preventing crime:

Home visitation; Day care/preschool; Parent training (with younger children); Home/community parent training (with older children); Multisystemic therapy.

Community-Based Prevention

Three types of community-based programs are considered to be promising in preventing crime:

Gang member intervention programs that are focused on reducing cohesion among youth gangs and individual gang members;

Community-based mentoring; After-school recreation.

School-Based Prevention

Three types of school-based programs have been found to be effective in preventing crime:

School and discipline management; Interventions to establish norms or expectations for behavior; Self-control of social competency instruction using cognitive-behavioral instruction

methods.

Placed-Focused Prevention

Three types of place-focused programs have been found to be effective in preventing crime:

Nuisance abatement; Closed-circuit television surveillance cameras; Improved street lighting.

3

Appendix: Excerpt from Evidence-Based Crime Prevention*

* Welsh, Brandon C., Evidence-Based Crime Prevention: Scientific Basis, Trends, Results and Implications for Canada, Research Report 2007-1, Ottawa, Ontario, Canada: National Crime Prevention Centre, Public Safety Canada, June 2007, available at http://www.publicsafety.gc.ca/prg/cp/_fl/evidenced-based-cp-e.pdf.


Author Note: MalcolmK.SparrowisprofessorofthePracticeofPublicManagementattheJohnF.Kennedy

SchoolofGovernmentatHarvardUniversity.ThiscasestudywaswritteninsupportoftheExecutiveSession

onPolicingandPublicSafetyattheHarvardKennedySchoolofGovernment.Itisdesignedtoserveasabasis

fordiscussionsregardingthenatureofanalyticsupportformodernpolicing.Theauthoracknowledgesvalu-

ableresearchassistanceprovidedbyBaillieAaron,withrespecttopolicingstrategies,andbyDr.LibbyJewett,

HypoxiaResearchProgramManagerattheNationalOceanographicandAtmosphericAdministration,with

respect to marine biology.

U.S. Department of JusticeOffice of Justice ProgramsNational Institute of Justice

Washington, DC 20531Official BusinessPenalty for Private Use $300

presorted standardpostage & fees paid

doJ/niJpermit no. g –91

*NCJ~227665*

NCJ 227665

Members of the Executive Session on Policing and Public Safety

Colonel Rick Fuentes, Superintendent, New Jersey State Police

Chief George Gascón, San Francisco Police Department

Mr. Gil Kerlikowske, Director, Office of National Drug Control Policy

Chief Cathy Lanier, Washington D.C. Metropolitan Police Department

Ms. Adrian Nicole LeBlanc, Visiting Scholar, New York University

Professor Tracey Meares, Walton Hale Hamilton Professor of Law, Yale Law School

Chief Constable Peter Neyroud, Chief Executive, National Policing Improvement Agency (UK)

Christine Nixon, Chair, Victorian Bushfire Reconstruction & Recovery Authority (Australia)

Chief Richard Pennington, Atlanta Police Department

Mayor Jerry Sanders, City of San Diego

Professor David Sklansky, Professor of Law, Faculty Co-Chair of the Berkeley Center for Criminal Justice, University of California, Berkeley, School of Law

Mr. Sean Smoot, Director and Chief Legal Counsel, Police Benevolent & Protective Association of Illinois

Professor Malcolm Sparrow, Professor of Practice of Public Management, Kennedy School of Government, Harvard University

Chief Darrel Stephens, Charlotte-Mecklenburg Police Department (retired)

Professor Christopher Stone, Guggenheim Professor of the Practice of Criminal Justice, Kennedy School of Government, Harvard University

Mr. Jeremy Travis, President, John Jay College of Criminal Justice

Mr. Rick VanHouten, President, Fort Worth Police Association

Professor David Weisburd, Walter E. Meyer Professor of Law and Criminal Justice, Director, Institute of Criminology, Faculty of Law, The Hebrew University and Distinguished Professor, Department of Administration of Justice, George Mason University

Dr. Chuck Wexler, Executive Director, Police Executive Research Forum

Chief Anthony Batts, Oakland Police Department

Professor David Bayley, Distinguished Professor, School of Criminal Justice, State University of New York at Albany

Dr. Anthony Braga, Senior Research Associate, Lecturer in Public Policy, Program in Criminal Justice Policy and Management, Kennedy School of Government, Harvard University

Chief William J. Bratton, Los Angeles Police Department

Chief Ella Bully-Cummings, Detroit Police Department (retired)

Ms. Christine Cole (Facilitator), Executive Director, Program in Criminal Justice Policy and Management, Kennedy School of Government, Harvard University

Commissioner Edward Davis, Boston Police Department

Chief Ronald Davis, East Palo Alto Police Department

Chief Edward Flynn, Milwaukee Police Department

Learn more about the Executive Session at: NIJ’s Web site: http://www.ojp.usdoj.gov/nij/topics/law-enforcement/executive-sessions/welcome.htm

Harvard’s Web site: http://www.hks.harvard.edu/criminaljustice/executive_sessions/policing.htm

new perspectives in policing: one week in heron city (a

Documents