Improving Police: What's Craft Got to Do with It?

7/28/2019 Improving Police: What's Craft Got to Do with It?

1/14

Ideas in

American

PolicingNumber 16

June 2013

Ideas in American Policingpresents commentary and insight rom leading crimi-

nologists on issues o interest to scholars, practitioners, and policy makers. The

papers published in this series are rom the Police Foundation lecture series o the

same name. Points o view in this document are those o the author and do not

necessarily represent the ocial position o the Police Foundation. The ull series

is available online at http://www.policeoundation.org/docs/library.html.

2013 Police Foundation. All rights reserved.

James J. Willis is an associate proessor in the Department o

Criminology, Law and Society at George Mason University and

a ormer member o the Police Foundation Research Advisory

Committee. His research interests include police organizations,

police reorm, police technology, and police decision making.

POLICE

The ly so short, the crat so long to learn.Georey Chaucer

To put it bluntly, it is not likely that police work generally, and

the work o individual ocers, will be appreciated at its actualvaluethat is as a service o being complex, important and

serious, until we begin to give a damn whether it is done well.Egon Bittner

Improving Police: Whats

Craft Got to Do with It?By James J. Willis

Over the last century

or so, the police have

been the object o

almost continuous and intensive

attempts at reorm. Currently,

one o the most powerul orces

or transorming what thepolice do and how they do it

is the evidence-based policing

movement, an approach that

challenges the police to base their

actions on scientic evidence

about what works. This puts

scientic research squarely in the

drivers seat o police decision

making, unlike past reorms that

have aspired to proessionalism

by ocusing on the legal and

administrative eatures o the

police environment (Klockars

1988). For example, the origins

o the proessional policing


2/14

2

model that dominated much

o the twentieth century lay in

the implementation o laws,

organizational policies, and

departmental rules. As eatures

o bureaucratic organization,

their purpose was to strengthen

accountability and to infuencethe choices made by patrol

ocers and the organizations that

employed them (Reiss 1992).1

Community policing, the most

popular reorm o the last ew

decades, then sought to reverse

this trend by assigning a greater

role to the needs and concerns

o local communities or guiding

police actions (Mastroski andGreene 1993, 80).

Whatever the reorm

approach, it is a common lament

among those seeking to improve

policing that the policing

crat, or the culmination o

knowledge based on hands-on

experience, is a eature o police

culture that poses a ormidable

obstacle to implementing new

policies and practices. Supporterso community- and problem-

oriented policing, two recent

and highly-touted reorms

(Goldstein 1990; Skogan 2006),

have expressed this concern,

as have advocates o evidence-

based policing. For example,

Cynthia Lum (2009) notes in

a previous Ideas in American

Policinglecture that despiteresearch demonstrating the crime

control benets o concentrating

police ocers in high-crime

areas or hot spots, there is little

indication that police agencies

have actually tried to reallocate

their patrol resources accordingly.

Here she echoes a lament made

over twenty-ve years ago by

Lawrence Sherman, a leader

in the evidence-based policing

movement, about police ocers

general resistance to scientic

discovery. Despite evidence

showing that arresting batterersin misdemeanor domestic

violence incidents in Minneapolis

was the most eective response

or reducing uture oending,

patrol ocers in the department

said they would continue to

use their standard responses,

including talking to both parties

or asking one to leave (Sherman

1984, 75).2 While proponentso evidence-based reorm are

careul to avoid attributing a

reluctance to embrace research to

a single cause, it is clear that they

consider the lower status that

police ocers assign to science

than to crat as a signicant

impediment to reorm (Sherman

1984, 1998; Weisburd 2008;

Lum 2009).Today it appears that

scholars are being attracted

in increasing numbers toward

the evidence-based movement

and policymakers, such as the

United States Department o

Justice, are encouraging police

to do the same.3 Consequently,

1 My comments in this essayconcentrate on patrol ocers. Thus,policing and police work reer to theactivities o these ront-line practitionersunless otherwise noted.

2 Subsequent replications o thisresearch at additional sites revealedmore complex ndings, including thecontribution o arrest to uture domestic

violence under some circumstances(Sherman 1992).

3 For example, the Oce o JusticePrograms Bureau o Justice Assistance

Web site provides resources on evidence-based approaches and practices (http://

www.ojp.usdoj.gov/BJA/evaluation/evidence-based.htm).

Improvements in policing

rest heavily on the shoulders

o those who do policing atthe coalace, and patrol

ocers have long thought

o the way they perorm

their work as a crat.


3/14

3

it would seem to be a good time

to reconsider the value o the

police crat in relationship to

police science (Bayley and Bittner

1984). Improvements in policing

rest heavily on the shoulders o

those who do policing at thecoalace, and patrol ocers have

long thought o the way they

perorm their work as a crat

(Wilson 1978, 283). Thus, unless

more attention is given to the

crat aspects o policing within

the context o the evidence-based

movement, it is unlikely that

eorts to integrate science with

policing will deliver the resultsthat reormers desire (Weisburd

and Neyroud 2011). Ultimately,

any attempt to improve police

perormance must take into

account the views o those who

constitute any departments largest

resource and their understanding

o what constitutes superior police

work (Skogan 2008).

Some might argue thatcrat and science already work

well together, noting that the

relationship between police

researchers and practitioners has

improved substantially over the

past ew decades (Bayley 2008).

This might be, but, aside rom

a ew scholars (see Bayley and

Bittner 1984; Mastroski 1996),

not much attention has ocusedon examining howscientic and

proessional knowledge might

contribute to one another in

mutually supportive ways in the

context o street-level decision

making. In this essay, I consider

what a true marriage o crat and

science might look like or guiding

the decisions o rank-and-le

ocers in two domains relevant

to police practice: (1) advancing

knowledge about what works,

and (2) making decisions about

the right thing to do. In doing

so, I hope to illuminate somepossibilities or reorm that

policymakers, practitioners, and

researchers might wish to consider

in their eorts to improve the

police o the uture.

Let Me Introduce

Our Couple, Science

and CraftFrom the perspective o evidence-

based policing, it is social science

that promises to revolutionize

the use o police discretion

(Sherman 1984, 61). A glance

through any research methods

textbook reveals that social

science encompasses a range

o methodologies. However,

the scientic gold standard othe evidence-based policing

movement is the experimental

study, as it is the most rigorous

methodological tool or

determining whether a causal

relationship exists between a

particular treatment and a desired

outcome (Sampson 2010). I you

want to learn whether problem-

oriented policing is more eectivethan directed patrol or reducing

crime at hot spots (Taylor,

Koper, and Woods 2011), or

whether arrest is the best option

or reducing recidivism in

domestic violence cases (Sherman

1992), then randomized trials

are your best hope. This view

o police science conjures an

image o the police proessional

as a technical expert, someone

whose eorts to solve crime and

disorder problems are infuenced

powerully by scientic research.

Indeed, Lawrence Sherman(1984, 7677) has envisioned

police ocers making street-level

decisions by accessing research

results on laptop computers in

their patrol cars. Having entered

data about a particular suspect, a

preprogrammed algorithm would

advise the patrol ocer on the

best course o action based on

the likely eects o the ocersactions on the suspects behavior.

From the perspective o

crat, proessionalism is dened

quite dierently. Experience,

not scientic knowledge, is the

oundation o eective police

work. By encountering a variety

o situations and people over

time, patrol ocers learn valuable

practical knowledge and developspecic skills. Some o these

situations might seem clear cut,

like dealing with a bank robbery

or other violent crime in progress

but others, such as domestic

disturbances or trac stops, are

more complex and uncertain.

This makes the decision about

how best to respond much more

challenging.Under these conditions,

what is embraced is situated

knowledge that oers an ocer

immediate guidance about the

constraints and possibilities or

responding to a specic incident

(Thacher 2008, 51). Crat

places a high value on fexibility


4/14

4

to t the right response to the

particulars o the situation and

does not necessarily demand

orthodoxy in response. It also

recognizes that what works

well or one ocer might not

work so well or another due todierences in skills and personal

traits. Scientic knowledge

predicting the likely outcomes o

a specic action is certainly useul,

but patrol ocers generally

assign much greater importance

to knowledge o laws and rules

and in-depth understanding o

people, places, and events. The

ormer helps dene the nature othe problem and the outer limits

o the police ocers authority

and responsibilities (Mastroski,

Willis, and Revier 2011, 16), and

the latter increases an ocers

ability to predict intention and

behavior when conronting an

unamiliar setting or the rst

time (Bittner 1990, 252).

Crat combines thisknowledge with a specic set o

skills. For patrol ocers, these

include the ability to remain calm

under pressure, to talk and listen

to people, to use orce sparingly,

and to exercise good judgment

by weighing up a complex set

o actors beore coming to a

reasoned decision (Bayley and

Bittner 1984; Fielding 1984;Kritzer 2007, 335; Muir 1977).

Those who have mastered these

tools o the trade are regarded by

their peers as master cratsmen or

women, that is, or being cool,

poised, inventive, careul, active,

and nonviolentocers who

can cope without jeopardizing

themselves or others (Bayley and

Bittner 1984, 5152).

In contrast to a computer-

driven robocop, the crat image

o the proessional police ocer

is o someone who thinks quickly

on her eet to behave in waysthat are wise, compassionate, and

air. Under some circumstances,

especially when an ocer senses

immediate danger, choices are

made automatically and intuitively

(Sherman 2012), but decision

making also involves considerable

observational and analytic rigor,

and even admits the possibility

that creativity in devising acourse o action may be valuable

(Muir 1977, 189259). Thus,

when police ocers make street-

level choices, they combine

experience-based intuition and

science-like analytic strategies or

thinking both ast and slow

(Kahneman 2011).

What Might a GoodMarriage Between

Science and Craft

Look Like?

This image o science and crat

vying or supremacy, in the

hope that springs eternal rom

American police reormers desire

to improve policing, generates

at least three distinct ways thatscience and crat may be coupled

together. In the rst coupling,

science is dominant and crat is

suppressed. While this model

is not unduly dismissive o

experience as a guide to police

action (Sherman 1984, 62), it

emphasizes the limitations o

intuitive knowledge and expresses

alarm about its potential or

negative consequences (Lum

2009). From an evidence-

based policing standpoint,

policing as a crat projects an

overly romantic conception o adiligent and skillul practitioner

making judicious decisions. In

doing so, it ails to consider

ully that even the most well-

respected patrol ocers have

a limited range o experiences

to draw upon and oten lack

the ability to learn much about

the long-term consequences o

their actions (that is, beyondwhat they are able to observe

during an encounter) (Sherman

1984). Where science is absent,

ocers are ree to dispense their

own version o justice based

on any number o problematic

or unethical actors, including

guesswork, personal biases, and

oensive stereotypes (Lum 2009,

3; Maynard-Moody and Musheno2003). Moreover, because

ocers oten do not have reliable

knowledge about the results o

their actions, what they do, no

matter how well intentioned, may

be ineectual or even harmul to

suspects, victims, and oenders

(Sherman 1984, 64).

These are important

criticisms, but there are alsolimitations to sciences capacity

to dictate decision making.

Leaving aside the act that

police may simply not trust

scientic ndings, supporters

o good scientic evaluations

oten attribute the ailure o

practitioners to embrace science


5/14

5

to their not being aware o

relevant ndings published in

academic journals, or to not

understanding the orm in which

they are delivered (Birkeland,

Murphy-Graham, and Weiss

2005). The obvious solution isto disseminate knowledge about

what works more widely and

in more digestible ormats

(Lum, Telep, Koper, and Grieco

2012). Clear examples o such

an approach are the recent

creation o the Oce o Justice

Programs CrimeSolutions.gov

Web site to highlight and rate

the eectiveness o dierentcriminal justice approaches on a

straightorward color-coded scale,

and the Evidence-Based Policing

Matrix summarizing the eects o

dierent crime control strategies

in a simple three-dimensional

cube.4

No doubt distrust and the

unavailability and inaccessibility

o scientic ndings help explainsome o the gap between

research, policy, and practice,

but there are other good

reasons why crat may be dea

to sciences exhortations. One

is the crucial recognition that

police organizations and the

ocers who work or them are

expected to accomplish multiple

goals at once. The evidence-based movement has tended to

identiy crime reduction as the

primary end o policing, but

police work is characterized by

a swath o values or ends, such

as equity, legitimacy, liberty, and

eciency, which oten confict

(Thacher 2001, 392). Under

these conditions, scientic

knowledge that identies the

best means to a given end, suchas crime control, oers useul

guidance, but it cannot resolve

these value conficts and thus

serve as a rm guide to action

(Thacher 2001, 389). This

is particularly true or patrol

ocers working at the street

level who must try to reconcile

the numerous requirements that

justice demands, say betweenbeing responsive to the needs and

wishes o a victim while reducing

the risk o uture oending. For

example, in the case o a minor

assault, should the ocer make

an arrest when a husbandthe

amilys only breadwinnerswears

that he will never slap his wie

again and she supports his claim

while pleading or leniency?

There are also times when

an ocers judgment should

override the dictates o even

well-established scientic ndings.For example, we probably do not

want police ocers to single out

domestic violence oenders or

arrest based on their employment

status or where they live, even i

evidence suggests that arresting

those with jobs or who live

in afuent areas is the most

eective means or reducing the

risk o uture violence (Sherman1998, 8). Finally, social science at

best improves the odds o success

o a given sort but, given the

complexities o human behavior,

it rarely guarantees it and it is

ar rom matching the predictive

power o the natural sciences

(Gutting 2012). Sherman (1984,4 http://gemini.gmu.edu/cebcp/

The evidence-based movement

has tended to identiy crime

reduction as the primary endo policing, but police work

is characterized by a swath

o values or ends, such as

equity, legitimacy, liberty, and

eciency, which oten confict.


6/14

6

1998) requently acknowledges

this act, but not in the context

o subjecting to empirical testing

the assumption that social science

is superior to crat in producing

desirable outcomes.

In the second coupling

between crat and science, crat is

dominant and science is missingor, i it is present, it is merely

a presentational strategy or

justiying traditional police

policies and practices (Manning

1992, 365). That is, science

is not really driving decision

making, at least in the ways it

is supposed to. Compstat could

be considered an example o

this version o the science-crat relationship. Implemented

in the New York City Police

Department in 1994, Compstat is

an inormation and management

tool that tallies and maps crime

statistics to hold command sta

accountable or crime levels

in their beats. According to

Compstat doctrine, police are

supposed to go beyond their

own experiences and to apply

innovations in crime prevention

theory and research in order to

solve crime problems (Bratton

1998). In practice, however,

in-depth research conducted

at multiple sites has shownthat despite the availability o

electronic maps, timely crime

data, and crime analysis, police

continued to rely heavily upon

what they have learned in the

course o their careers about

where crime occurs and how best

to respond. Rather than careully

assessing a range o promising

alternatives beore selecting themost eective crime strategy,

district commanders requently

used tactics they have tried in the

past and that they believed work,

such as saturating an area with

patrol or increasing arrests (Willis,

Mastroski, and Weisburd 2007).

In this case, impressive electronic

maps and crime statistics

the harbingers o science

help coner legitimacy on police

actions while the experience-

based aspects o police work

continue to hold sway.

A third model regards cratand science as a true marriage,

where each partner bestows

equal worth on the other and

has a rightul place in guiding

the decisions o the rank-and-

le ocer. In the next section, I

try to envision what a successul

marriage o this sort would look

like. Which partnercrat or

sciencedoes what, and howmight some dierences between

the two be resolved?

Advancing Knowledge

About What Works

In the rst instance, police

science could probably do more

to pay attention to police crat

in order to validate what worksand under what conditions.

In ocusing so much o its

energy on identiying the best

means to preventing crime, the

evidence-based movement has

improved understanding about

how police can contribute to

public saety. In comparison,

it has largely overlooked the

many concerns ocers musttake into account when making

a decision, including the many

dierent tactics that are available

to them (Bayley and Bittner

1984). Egon Bittner noted long

ago that the work o ront-line

police ocers could not be

adequately captured and assessed

In the rst instance, police

science could probably do

more to pay attention to

police crat in order to

validate what works and

under what conditions.


7/14

7

in such simple terms as ghting

crime or enorcing laws (Bittner

1970). In their encounters

with the public, ocers must

consider other important goals,

including preventing disputes

rom escalating, ensuring saetyat the scene, and responding

adequately to legitimate citizen

needs (Bayley and Bittner 1984;

Mastroski 1996). Little is still

known about the best treatments

available or accomplishing

these kinds o goals. Albert

Reiss (1995, 103) noted with

surprise almost twenty years

ago that interpersonal confictsand disputes were the cause o

much violence but that little is

known about the eectiveness

o police in preventing the

occurrence o dierent kinds

o disputes or their escalation

into crimes when they occur.

Little seems to have changed

in the interim, suggesting that

researchers need to do moreto learn rom practitioners

themselves at close range, that

is, through direct observations

and interviews, about the goals

they identiy as most important

in any given situation and the

specic tactics they use or their

accomplishment. Ethnographers,

such as Egon Bittner and William

Muir, have provided valuableinsights into how patrol ocers

use their discretion but not as

part o a sel-conscious attempt to

propose standardized treatments

and then subject them to

empirical evaluation.

In the absence o scientic

attempts to identiy, analyze,

and validate much o what

experience has taught patrol

ocers, practitioners currently

must rely on what they have

learned by listening to old hands

in their agency or through their

acquisition o personal experience.As Bayley and Bittner (1984, 47)

observed almost thirty years ago,

those interested in improving

the quality o police work

stand to gain much by using

systematic ways to identiy and

test scientically the operational

imperatives that ocers consider

important in their daily work

dealing with the public.Like a good marriage, the

willingness o both partners to

consider alternative perspectives

on any given issue beore making

judgments is an important

criterion or success. In this

regard, proponents o science

or crat should not just assume

that one is more eective than

the other in leading to desirableoutcomes. This leads crat or

science to merely proselytize to

one another rather than using

evidence and reason to try to

resolve where their dierences

lie. Thus, it would be ruitul or

science and crat to consider the

assumption that one is superior

to the other to be a hypothesis

worth testing, but to date thisidea has received little attention.

To some degree, this squabble

could be overcome by conducting

rigorous evaluations that compare

police discretion that is exercised

using science to police discretion

exercised using crat (and some

combination o the two).

One could envision an

experiment that uses a variety

o treatment conditions. Some

ocers could be provided with

access to scientic knowledge that

they are prepared to use in concert

with the kind o computer-aideddiscretion model espoused by

Sherman and discussed earlier.

Ocers assigned to this group

would be required to have their

discretion governed by scientic

ndings. In responding to a

particular situation, such as a

domestic dispute, a computer

would advise ocers about the

preerred strategy given theiranswers to a set o relevant

questions that might include the

oenders prior record and the

seriousness o the oense.

A second treatment condition

might be a special training and

supervision program that brings

out the best o what crat has to

inorm ocers about what to do.

Bayley and Bittner (1984, 54)propose such a model when they

recommend that those ocers

identied as master cratsmen or

women are used to train their

peers on those skills they consider

most important to producing

high quality police work.

Supervision by these experienced

and respected cratspeople might

include regular debriengs,especially ollowing situations

that an ocer ound particularly

challenging. Their purpose would

be to allow ocers to discuss and

seek counsel on their decision-

making process, including the

accuracy o the initial diagnosis o

the situation, the appropriateness


8/14

8

o the goals being pursued, and

the strengths and limitations

o the tactics used or their

accomplishment.

A third treatment condition

could be a hybrid approach,

where ocers are exposedto scientic evidence and are

prepared to use it, but they are

also exposed to the best that

crat has to oer. These ocers

are then given the reedom to

ashion their actions to best t

particular circumstances. The

control condition would be that

the department provides no more

guidance than is usually availableto ocers through its existing

rules, training, supervisory

practices, and perormance review

process.

The experiment could be

designed to measure a range o

outcomes, including reductions

in crime and disorder and also

citizen satisaction or alienation.

Thus, inormation could beprovided not just on which o

these approaches is more likely

to produce desirable results, but

which o these produces more

desirable results.

The experiment should also

incorporate a process evaluation

or patrol ocers to provide

eedback to researchers about

how well each method workedand what were the most helpul

or dicult aspects o doing each.

In addition to helping illuminate

why a particular innovation did

or, as importantly, did not work

(since inormation about ailures

is equally valuable to advancing

reorm), this eedback can open

up potentially ruitul avenues or

uture scholarly inquiry (Willis

and Mastroski 2011). Moreover,

here crat might be particularly

useul in identiying the specic

circumstances under which one

response is likely to be moreor less eective. Replications o

the Minneapolis experiment on

domestic violence showed that

the eects o arrest depended on

the status o oenders and the

degree to which they experienced

procedural justice (Sherman

1992), but similar insights can

be revealed by tapping requently

and systematically into the richvein o practitioners experiences

and then testing them.

But it is not just the oender

who matters in predicting

outcomes. Ethnographies and

surveys reveal that ocers

generally believe that what works

best or one ocer might not

work well or another based on

the skills and characteristics othe individual ocer doing the

job (Muir 1977; Mastroski,

Willis, and Revier 2011). These

characteristics include an ocers

gender, amount o experience,

physical size, and verbal acility.

Identiying and then using science

to measure these interactive

eects might reveal that dierent

ocer styles are equally eectiveand thereore provide a range

o alternative responses rom

which to choose. In turn, such

customization o responses

could increase the prospects that

any research would actually be

incorporated into ocers daily

decision making by eschewing a

one-size-ts-all approach (Bayley

and Bittner 1984, 51). In short,

validating what works, under

what conditions, and by whom

would help strengthen the bond

between science and crat by

serving the needs and perspectiveso ront-line workers directly.

In doing so, it would redress

a current imbalance in police

research that tends to ocus on

the kinds o program evaluations

that are more relevant to

policymakers and police managers

than street-level decision makers

(Thacher 2008).

Deciding on the

Right Thing to Do

The marriage between science

and crat could be urther

improved by partisans rom

both sides working more closely

together to help police ocers

deliver more justice than is

currently on oer. To date, socialscientists have commonly ignored

the undamental normative

component to assessing the

quality o work that street-

level patrol ocers perorm,

preerring to ocus on explaining

and predicting variations in

what police do and how they

do it (e.g., use o orce, arrests).

This trend toward separatingact rom value, or how much

rom how well, extends back

at least as ar as Max Weber,

who asserted that while science

could oer actual statements

and causal explanations, it was

incapable o resolving questions

about important public values


9/14

9

or about what ought to be

(Thacher 2006). According

to Weber, answers to these

questions ultimately depended

on ones particular moral or

political outlook and could not

be validated empirically. Theproblem here, o course, is that

measuring the quantity o police

work an ocer perorms tells us

very little about its quality, or

whether or not an ocer uses his

or her discretion to do the right

thing. A consequence o this

traditional divide is that science

is virtually silent on those aspects

o police work that matter mostto police leaders, their ocers,

and the communities they serve.

This is a major oversight. Over

the last ew decades, many police

leaders have demonstrated a

clear commitment to promoting

better policing (Bayley 2008),

patrol ocers have expressed

interest in more sophisticated

approaches to assessing theirperormance than simple tallies

o work outputs (Mastroski,

Willis, and Revier 2011), and

research has shown that citizens

care mightily about the quality

o treatment they receive in their

personal encounters with police

ocers (Tyler 2004; Bottoms

and Tankebe 2012). Surely as

researchers we can do more tolearn rom patrol ocers and

help them make better choices.

Take, or example, two

patrol ocers responding to a

dispute between two neighbors

in an apartment building. A

complainant is upset that the

woman living in the adjacent

apartment, who may or may

not be suering rom mental

illness, has been pounding on

her door with a fat iron and

physically threatening her. When

questioned, the woman with the

fat iron says she is rustrated

by the complainants tendencyto slam her door when entering

and leaving. The ocers express

puzzlement that this should be

the cause o so much hostility,

but they do not explore this in

detail with the second neighbor.

They advise the complainant that

she should get a summons, and

advise the second woman that

she could also get a summonsbut that she is not allowed to

take the law into her own hands

and retaliate by damaging the

complainants door. When she

remains deant, they tell her

that i they have to return that

night, she will be taken down

to the precinct. Having given

this warning, the ocers leave,

disposing o the fat iron in a

nearby trash can as they wait or

an elevator. The entire encounter

lasts less than ten minutes.

This is the kind o run-o-

the-mill dispute that characterizes

everyday police work, but whatis the best response in this

particular case? Is it delivering

a sense o justice to the citizens

involved? Resolving the

underlying cause o the dispute?

Giving these citizens the capacity

to solve this problem without

summoning the law? Making

costly police resources available

to those who are in greater needo police services? This dispute

is taken rom a video clip5 o

an actual incident, which my

colleagues and I showed to

The marriage between science

and crat could be urther

improved by partisans rom

both sides working more

closely together to help police

ocers deliver more justice

than is currently on oer.

5 This clip is rom The Police Tapes(1976), which was produced, edited, anddirected by Alan and Susan Raymond.


10/14

10

patrol ocers as part o a studyexamining how they judge the

quality o the work that they

perorm (Willis et al. 2011).

What we discovered was that

there was little consensus about

what constituted good police

work, suggesting that at present

work quality at the street level is

let largely to the will and skill o

the individual patrol ocer.While it is impossible to

generalize rom a study o a

single department, ocers

responses to this clip appear

to indicate that existing

mechanisms may not be doing

a very good job o promoting

and advancing a common

vision o what constitutesgood policing. There are good

reasons or this. Available

options or oering practical

guidance, namely, bureaucratic

rules and laws and, as has been

identied here, increasing

scientic knowledge, can oer

only partial solutions as they

are challenged by their general

qualities (Marx 2006, 280).The art o street-level decision

making is in guring how, i at

all, these can be applied to the

contextual richness o individual

cases where inormation is

oten limited, inchoate, and

conficting, and where there is

pressure to act quickly (Schon

1983). The advantage o crat

is that it provides a stock o

knowledge acquired through

years o handling many dierent

situations and contingencies, and

yet this source is seldom tapped

systematically and made availableas a source o guidance. What

is more, science can play an

important role in this process.

Take the disgruntled neighbor

scenario described above. The

video clip could be shown to a

group o ocers who have been

recruited or their skillul work.

They could be asked to judge the

quality o the ocers responsein the clip and to identiy the

strengths and weaknesses o the

approach on display. A police

researcher could play a useul role

in this process by helping identiy

and clariy the major concerns

that arise. Based on what we

heard during our interviews,

these might include a range o

dimensions as shown in the tableon this page.

Working in this collaborative

environment, the group could

then be asked to identiy more

or less desirable responses, with

the researcher oering insights

based on scientic theory and

evidence, particularly regarding

potentially harmul consequences.

It would also be necessary to tryto establish priorities among these

criteria in order to make their

application useul to others, a

tough challenge but one that it is

still possible. Studies on the crat-

based culture o policing have

oten examined its undesirable

or negative eatures, such as

Possible Dimensions for MeasuringQuality of Police Performance

Procedural

justice

To what degree were the victim and

oender treated airly and respectully?

Problem

diagnosis

Did the ocers demonstrate good skills

in identiying the nature o the problem?

Safety and orderat the scene

Was the potential or immediate violence

adequately controlled?

Lawfulness

of response

Was their response justied by the

evidence?

Distributivejustice

Was the resolution air to allstakeholders?

Economy Did the ocers spend too much ortoo little time on the dispute?

Prospects for

future risk

Did the ocers reduce the potential or

uture problems between the two?


11/14

11

alienation rom the department

hierarchy or hostility toward

the public, but ethnographies

show that good ocers develop

intellectual and moral virtues that

help them weigh the exigencies

o any given situation and makesophisticated judgments about

the best thing to do (Muir 1977,

189224). Providing a orum

that encourages thoughtul

deliberation about these

judgments and the trade-os

they imply would help clariy

priorities among dierent value

systems. Anticipating potential

conficts ahead o time would alsohelp guide discretion in the eld,

as trying to resolve these while

under pressure to act quickly is

very dicult. In the case o the

disgruntled neighbor, while the

ocers were certainly attentive

and polite, a persuasive case could

be made that their gracious and

ecient manner took undue

precedence over eorts tominimize the possibility o uture

confict. In our interviews, some

respondents were surprised that

the ocers seemed satised with

leaving the problem in the same

state they had ound it. There

will certainly be disagreements

about how to best handle

these kinds o disputes, but, by

identiying what the relevantvalues are in a particular context,

exploring their meanings, and

clariying which should take

priority, it should be possible

to justiy some responses as

superior to others. Furthermore,

acknowledging where the

tensions lie between dierent

uses o discretion and identiying

acceptable levels o compromise

would also help inorm decision

making.

Police leaders, especially

chies, rst-line supervisors,

recruit trainers, and eld trainingocers, should play a key role

in this process. Through their

participation, they can help

establish the important public

values that the organization

should pursue and inspire others

to embrace them. What is oten

lost in discussions about sciences

role in governing practice is the

vital contribution a coherentand enlightened philosophy can

make to police work, and yet it

is a valuable leadership trait or

those in authoritative positions

to establish the cultural tone o

their organization by advancing

a view o good policing that is

transparent both to employees

and to the public (Bass 1998).

The next important stepwould be to subject these

standards to empirical testing

beore coming up with a

strategy that allows or them

to be applied in an operational

setting (Mastroski 2007). It

might be possible, or example,

to construct a checklist that

promotes memory recall about

what is important and why andalso helps improve consistency, a

key element o crat (Gatawande

2009; Kritzer 2007). While

guidelines cannot account or

every contingency and will

sometimes not work as intended,

by providing structure they can

increase the likelihood that patrol

ocers will use their discretion

in desirable ways (Kelling 1999).

This approach to mobilizing crat

could be used or a variety o

encounters that are selected or

being particularly problematic

(like various domestic disputes),or because they are commonplace

(trac stops, or example). The

availability o body cameras

easily allows or this kind o

naturalistic observation, as well

as opportunities or supervisors

to give eedback on ocers

perormance beore and ater

the implementation o this

discretionary tool.

Conclusion

I have suggested that we want

our police ocers to act wisely

and well, doing the kind o

job that makes us step back

in admiration at their capacity

to make good judgments.

Undoubtedly there are someocers who prompt this reaction,

so perhaps it makes sense to

take greater advantage o the

insights they have to oer and to

do more to assess their eects.

The evidence-based movement

has captured the attention o

government and generated

excitement about the possibilities

or reorm, so this is a goodtime to use science to cultivate

and test what accumulated

experience has to oer. Oscar

Wilde once remarked that the

proper basis or marriage was

mutual misunderstanding, and

that happiness could not be

ound within its bounds. This


12/14

12

does not have to be the case

or crat and science, with one

looking past the other and

lamenting missed opportunities.

A uller appreciation o the

qualities each brings to the other

promises a much more satisyingand enduring relationship.

Advancing reorm in ways that

police administrators, ocers,

researchers, and ordinary citizens

all care about requires that we

ocus on what can be gained by

strengthening this union and not

on the dierences that divide it.

I would like to thank David

Weisburd and Cynthia Lum or

their thoughtul comments on

the science and crat o policing.

A special note o gratitude

goes to Stephen Mastroski,

the voice inside and outside o

my head always pushing me to

think harder. His comments on

earlier drats on this essay were

invaluable. Finally, I appreciate theassistance o Mary Malina, Police

Foundation communications

director, who oversaw this essays

production. Any errors that

remain are my own doing.

References

Bass, Bernard M. 1998.

Transormational Leadership:

Industry, Military, and Education

Impact. Mahwah, NJ: Lawrence

Erlbaum Associates.

Bayley, David H., and Egon Bittner.

1984. Learning the skills o

policing. Law and Contemporary

Problems47: 3559.

Bayley, David H. 2008. Police reorm:

Who done it? Policing and Society

18: 717.

Birkeland, Sarah, Erin Murphy-

Graham, Carol Weiss. 2005. Good

reasons or ignoring good

evaluation: The case o the drugabuse resistance education

(D.A.R.E.) program. Evaluation

and Program Planning28: 24756.

Bittner, Egon. 1970. The unctions o

the police in modern society. Chevy

Chase, MD: National Institute o

Mental Health.

Bittner, Egon. 1990. Aspects o Police

Work. Boston, MA: Northeastern

University Press.

Bottoms, Anthony, and JusticeTankebe. 2012. Beyond procedural

justice: A dialogic approach to

legitimacy in criminal justice. The

Journal o Criminal Law and

Criminology102: 11970.

Bratton, William J. (with Peter

Knobler). 1998. Turnaround:

How Americas Top Cop Reversed

the Crime Epidemic. New York:

Random House.

Fielding, Nigel. 1984. Policesocialization and police competence.

British Journal o Sociology35:

56890.

Gatawande, Atul. 2009. The Checklist

Maniesto: How to Get Things Right.

New York: Metropolitan Books.

Goldstein, Herman. 1990. Problem-

Oriented Policing. New York:

McGraw-Hill.

Gutting, Gary. 2012. How reliable are

the social sciences? New York Times,

17 May. http://opinionator.blogs.

nytimes.com/2012/05/17/how

-reliable-are-the-social-sciences

/?emc=eta1

Kahneman, Daniel. 2011. Thinking,

Fast and Slow. New York: Farrar,

Straus, and Giroux.

Kelling, George L. 1999. Broken

Windows and Police Discretion.

Washington, DC: National Institute

o Justice. http://www.ncjrs.gov

/pdles1/nij/178259.pd

Klockars, Carl. 1988. The rhetoric o

community policing. In CommunityPolicing: Rhetoric or Reality, eds.

Jack R. Greene and Stephen D.

Mastroski. New York: Praeger.

Kritzer, Herbert M. 2007. Toward a

theorization o crat. Social and

Legal Studies16: 32140.

Lum, Cynthia. 2009. Translating

Police Research into Practice. Ideas

in American Policing, no. 11,

August. Washington, DC: Police

Foundation. http://www.policeoundation.org/pd/Ideas

_Lum.pd.

Lum, Cynthia, Cody Telep,

Christopher Koper, and Julie

Grieco. 2012. Receptivity to

research in policing. Justice Research

and Policy14: 6195.

Marx, Gary T. 2006. Forget big

brother and big corporation: What

about the personal uses o

surveillance technology as seen incases such as Tom I. Voire? Rutgers

Journal o Law and Urban Policy3:

219286.

Manning, Peter K. 1992. Inormation

technologies and the police. In

Modern Policing, eds. Michael

Tonry and Norval Morris. Chicago:

University o Chicago Press.

Maynard-Moody, Steven W., and

Michael C. Musheno. 2003. Cops,

Teachers, Counselors: Stories rom theFront Lines o Public Service. Ann

Arbor: University o Michigan Press.

Mastroski, Stephen D., and Jack R.

Greene. 1993. Community policing

and the rule o law. In Police

Innovation and Control o the Police:

Problems o Law, Order, and

Community, eds. David Weisburd,


13/14

13

Craig Uchida, and Lorraine Green

Mazerrolle. Berlin: Springer-Verlag.

Mastroski, Stephen D. 1996.

Measuring police perormance in

public encounters. In Quantiying

Quality in Policing, Larry T.

Hoover, ed. Washington, DC:Police Executive Research Forum.

Mastroski, Stephen D. 2007. Police

organization and management issues

or the next decade. Paper presented

at the National Institute o Justice

Policing Research Workshop:

Planning or the Future,

Washington, DC, November 2829.

https://www.ncjrs.gov/pdles1

/nij/grants/218584.pd

Mastroski, Stephen D., James J. Willis,and Lauren Revier. 2011. Results

o a survey on high quality policing

in Manassas City, VA. Fairax, VA:

George Mason University,

Department o Criminology,

Law and Society.

Muir, William K., Jr. 1977. Police:

Streetcorner Politicians. Chicago:

University o Chicago Press.

Reiss, Albert, Jr. 1992. Police

organization. In Modern Policing,eds. Michael Tonry and Norval

Morris. Chicago: University o

Chicago Press.

Reiss, Albert, Jr. 1995. The role

o police in crime prevention.

In Integrating Crime Prevention

Strategies: Propensity and

Opportunity, eds. Per-Olo H.

Wiskstrom, Ronald V. Clarke, and

Joan McCord. Stockholm: National

Council or Crime Prevention.

Sampson, Robert J. 2010. Gold

standard myths: Observations on

the experimental turn in quantitative

criminology. Journal o Quantitative

Criminology26: 489500.

Schon, Donald A. 1983. The Refective

Practitioner: How Proessionals Think

in Action. New York: Basic Books.

Sherman, Lawrence W. 1984.

Experiments in police discretion:

Scientic boon or dangerous

knowledge? Law and ContemporaryProblems47: 6181.

Sherman, Lawrence W. 1992. Policing

Domestic Violence: Experiments and

Dilemmas. New York: The Free

Press.

Sherman, Lawrence W. 1998.

Evidence-Based Policing. Ideas in

American Policing, no. 2, July.

Washington, DC: Police Foundation.

http://www.policeoundation.org

/pd/Sherman.pd.Sherman, Lawrence W. 2012. Policing,

theory and evidence. Paper

presented at the annual meeting

o the American Society o

Criminology, Chicago, November.

Skogan, Wesley. 2006. Police and

Community in Chicago: A Tale o

Three Cities. Chicago: University

o Chicago Press.

Skogan, Wesley. 2008. Why reorms

ail. Policing and Society18: 2334.

Taylor, Bruce, Christopher S. Koper,

Daniel J. Woods. 2011. A

randomized controlled trial o

dierent policing strategies at hot

spots o violent crime. Journal

o Experimental Criminology7:

14981.

Thacher, David. 2001. Policing is

not a treatment. Journal o Research

in Crime and Delinquency38:

387415.

Thacher, David. 2006. The normative

case study.American Journal o

Sociology111: 163176.

Thacher, David. 2008. Research or

the ront lines. Policing and Society

18: 4659.

Tyler, Tom R. 2004. Enhancing police

legitimacy. The Annals o the

American Academy o Political and

Social Sciences593: 8499.

Weisburd, David 2008. Place-Based

Policing. Ideas in American Policing,

no. 9, January. Washington, DC:Police Foundation. http://www

.policeoundation.org/pd

/placebasedpolicing.pd.

Weisburd, David, and Peter Neyroud.

2011. Police Science: Toward a New

Paradigm. New Perspectives in

Policing, Executive Sessions on

Policing and Public Saety, January.

Cambridge, MA: Harvard Kennedy

School. http://www.hks.harvard

.edu/var/ezp_site/storage

/ckeditor/le/pds/centers-

programs/programs/criminal

-justice/NPIP-Police%20Science

-TowardaNewParadigm.pd

Willis, James J., Stephen D. Mastroski,

and David Weisburd. 2007. Making

sense o COMPSTAT: A theory-

based analysis o organizational

change in three police departments.

Law and Society Review41: 14788.

Willis, James J., and Stephen D.

Mastroski. 2011. Innovations in

policing: meanings, structures,

processes. Annual Review o Law

and Social Science7: 309334.

Willis, James J., Stephen D. Mastroski,

Jillian Baird, Lauren Revier, Terri

Hines, and Tal Jonathan. 2011.

Results o interviews on high quality

policing in Manassas City, VA.

Fairax, VA: George Mason

University, Department o

Criminology, Law and Society.

Wilson, James Q. 1978. Varieties o

Police Behavior: The Management

o Law and Order in Eight

Communities. Cambridge: Harvard

University Press.


14/14

1201 Connecticut Avenue, NW, Washington, DC 20036-2636(202) 833-1460 Fax (202) 659-9149 E-mail: [email protected]

www.policefoundation.org

POLICE

ABOUT THE POLICE FOUNDATION

The Police Foundation is a national, nonpartisan, nonprot organization dedicated to advancinginnovation and science in policing. As the countrys oldest police research organization, thePolice Foundation has learned that police practices should be based on scientic evidence aboutwhat works best, the paradigm of evidence-based policing. Established in 1970, the foundationhas conducted seminal research in police behavior, policy, and procedure, and works to transfer

to local agencies the best new information about practices for dealing effectively with a rangeof important police operational and administrative concerns. Motivating all of the foundationsefforts is the goal of efcient, humane policing that operates within the framework of democraticprinciples and the highest ideals of the nation.

DIVISION OF RESEARCH, EVALUATION,& PROFESSIONAL SERVICES

Karen L. AmendolaChief Operating Ofcer

David WeisburdSenior Fellow

Garth den Heyer

Senior Research Fellow

Travis Taniguchi

Senior Research Associate

Edwin E. HamiltonProfessional Services Director

Emily OwensResearch Fellow

Adam KaufmanResearch Assistant

Mary SiglerResearch Assistant

Maria ValdovinosResearch & Administrative Coordinator

RESEARCH ADVISORY COMMITTEE

David Weisburd, Chair

Hebrew University and George Mason University

Anthony A. BragaRutgers University and Harvard University

Robin S. EngelUniversity of Cincinnati

Christopher KoperGeorge Mason University

Jerry H. RatcliffeTemple University

BOARD OF DIRECTORS

ChairmanWeldon J. Rougeau

PresidentJames Bueermann

George H. Bohlinger III

Clarence Edwards

Dean Esserman

Paul Helmke

Julie Horney

William H. Hudnut IIIJonathan Knowles

Mark S. Mellman

W. Walter Menninger

Elsie L. Scott

Andrew L. Sonner

Documents

Improving Police: What's Craft Got to Do with It?