police quarterly - university of cincinnati · discover how, if at all, patrolmen in various cities...

http://pqx.sagepub.com/Police Quarterly

http://pqx.sagepub.com/content/11/4/447The online version of this article can be found at:

DOI: 10.1177/1098611108314567

2008 11: 447 originally published online 29 February 2008Police QuarterlyJohn Liederbach and Lawrence F. Travis III

Wilson Redux : Another Look at Varieties of Police Behavior

Published by:

http://www.sagepublications.com

On behalf of:

Police Executive Research Forum Police Section of the Academy of Criminal Justice Sciences

can be found at:Police QuarterlyAdditional services and information for

http://pqx.sagepub.com/cgi/alertsEmail Alerts:

http://pqx.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://pqx.sagepub.com/content/11/4/447.refs.htmlCitations:

at UNIV OF CINCINNATI on September 15, 2011pqx.sagepub.comDownloaded from

http://pqx.sagepub.com/

http://pqx.sagepub.com/content/11/4/447

http://www.sagepublications.com

http://www.policeforum.org/

http://www.acjs.org/

http://pqx.sagepub.com/cgi/alerts

http://pqx.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://pqx.sagepub.com/content/11/4/447.refs.html


447

Authors’ Note: Correspondence concerning this article should be addressed to John Liederbach,Department of Human Services, Criminal Justice Program, 236 Health Center, Bowling Green StateUniversity, Bowling Green, OH 43403-0148; e-mail: [email protected].

Police QuarterlyVolume 11 Number 4

December 2008 447-467© 2008 Sage Publications

10.1177/1098611108314567http://pqx.sagepub.com

hosted athttp://online.sagepub.com

Wilson ReduxAnother Look at Varietiesof Police BehaviorJohn LiederbachBowling Green State UniversityLawrence F. Travis IIIUniversity of Cincinnati

Since the publication of Varieties of Police Behavior more than four decades ago,Wilson’s theory concerning the link between local political culture and police styles hashad a significant and enduring impact on police scholarship. However, there have beensurprisingly few empirical tests of Wilson’s assertions, and of those limited number ofstudies, only a small number provide even partial support for the model. The current studyprovides a test of Wilson’s theory using a unique measure of police style derived fromobservational data rather than official arrest/citation statistics. Perhaps more important,we discuss our findings within the context of previous research that has failed to supportWilson’s theory and present a critical assessment of the requirements of a valid test of thistheory. We argue that Wilson’s model retains an intuitive appeal to those interested inexplaining the relationship between local community context and police behavior.

Keywords: James Q. Wilson; varieties of police behavior; police styles; police andcommunity variation; systematic social observation

In 1968, James Q. Wilson published Varieties of Police Behavior: The Management ofLaw and Order in Eight Communities. This book is a classic of the policing liter-

ature and is widely known and studied. In Varieties, Wilson sets out to investigatesome of the core issues associated with understanding the police including (a) howpolice organizations influence the street-level actions of officers, (b) the factors thatlimit organizational control of individual officer behavior, and (c) how the contex-tual influences of community and political culture contribute to defining certain“organizational styles” exhibited by local police agencies. In Wilson’s (1968) words,

The purpose of this book is to describe how the police patrolman behaves with respectto the more frequently applied laws; to analyze the problems facing the police admin-istrator both in deciding what the patrolman ought to do and in getting him to do it; to



discover how, if at all, patrolmen in various cities differ in performing their functions;and finally to inquire whether—or under what circumstances—such differences as existare based on explicit community decisions. (p. 4)

In the pursuit of these goals, Wilson identified the primary functions of the localpolice as law enforcement, order maintenance, and service delivery. He studied howpolice officers in different departments interacted with citizens noting that theseinteractions varied in terms of frequency and formality. In some communities thepolice interacted with citizens often, whereas in other places police rarely dealt withcitizens. In some communities, police interactions with citizens typically involvedthe application of the law through arrests and/or citations, and in other places theseinteractions were more informal and law enforcement outcomes were rare.

Ultimately, Wilson identified what he described as three distinct “operational styles”of policing based on the frequency and formality of police–citizen interactions.Where police interacted frequently with citizens, and in a formal manner, Wilsonidentified a Legalistic style. In communities where police dealt with citizensfrequently, but were typically more informal and did not invoke the law, he identifieda Service style. Finally, Wilson said a Watchman style characterized communitieswhere the police infrequently encountered citizens.

Four decades after the publication of Varieties, Wilson’s conceptualization anddescription of the police in the United States has become part of what might becalled the “common knowledge” of policing. As Hassell, Zhao, and Maguire (2003)note, almost every textbook on American policing provides an overview of Wilson’sthree operational styles, and Wilson’s contributions have been routinely described as“classic,” or encompassing “perhaps the most important” or “prevailing view” concerningthe organizational influences on police behavior (Carter, 2002, p. 129; Roberg, Novak,& Cordner, 2005, p. 277; Walker & Katz, 2008, p. 369). Moreover, the ideas putforth in Varieties appear to have had a “deep and broad impact,” so much so that thebook still ranks among the most commonly cited works in the police studies literature,and it has obviously contributed to Wilson’s stature among the most influential policingscholars (Cohn, Farrington, & Wright, 1998; Wright & Miller, 1998, p. 249). In short,“everyone” knows that there are three different styles of policing in America.

The question that remains, of course, is what explains these differences? Wilsonobserved that police styles appeared to be related to the characteristics of thecommunity within which the department operated, including the form of local gov-ernment. He argued that police departments were constrained by what he called the“local political culture.” Police activity, according to Wilson, was rarely directed byconscious choices and direct political intervention. Rather, “the prevailing politicalculture creates a ‘zone of indifference’ within which the police are free to act as theysee fit” (Wilson, 1968, p. 233). Wilson found support for the notion that police stylewas a function of local political culture using city and official arrest statistics from1960 for 146 cities with populations ranging from 25,000 to 100,000.

448 Police Quarterly



Given the significant and enduring impact of Varieties, as well as the limitationsassociated with what Wilson characterized as his own “crude and inadequate” test,there have been surprisingly few studies designed to test Wilson’s assertions, and ofthose limited number of studies, only a few provide even partial support for the model(Crank, 1990, 1992; Langworthy, 1985; Pursley, 1976). For example, on the basis ofhis analysis, Langworthy (1985) concluded that Wilson’s theory was more one ofcentral tendency than of constraint. Likewise, Crank’s (1990) test provided onlyqualified support for the notion that local political culture influences the street-levelbehavior of officers. Perhaps even more disconcerting is the fact that the more recenttests to confirm Wilson’s theory have failed (Hassell et al., 2003; Zhao & Hassell,2005; Zhao, He, & Lovrich, 2006). Indeed, Zhao and Hassell (2005) characterizedtheir null findings as indicative of a more general “move toward a falsification ofWilson’s seminal work” (p. 423).

This situation has created an obvious contradiction; that is, the huge imprint that hasbeen left by Wilson’s theory on police scholarship can scarcely be justified by thesomewhat tepid support it has thus far garnered in the empirical literature (Slovak,1996; Zhao et al., 2006). Certainly, there may be valid explanations for the general lackof existing empirical support. For example, the general state of knowledge concerningthe relationship between officer behavior, organizational factors, and the wider com-munity context has been characterized by some as “not well developed” (NationalResearch Council, 2004, p. 156). Others have pointed to a variety of factors that haveworked to alter policing since the publication of Varieties, including the movementtoward police professionalism and the increasing influence of the federal governmentin local policing (Hassell et al., 2003; Zhao & Hassell, 2005; Zhao et al., 2006).

These explanations probably have some merit given both the time elapsed sinceVarieties’ publication and the enormous changes that have occurred in policing overthe course of that time; however, the relative lack of empirical support for Wilson’smodel may also be due in part to the manner in which existing tests have operational-ized his conceptualization of police style. Specifically, previous studies designed to testfor community and/or organizational correlates of police style have exclusively usedofficial arrest/citation statistics as the indicator of police style (see, e.g., Crank, 1990,1992; Langworthy, 1985; Zhao et al., 2006). Though Wilson (1968) himself used offi-cial statistics in his own test, he recognized the inadequacy of the measure:

It must be emphasized, however, that police style is not always best measured by arrestrates, even for high discretion offenses. Albany and Oakland both arrest a lot of drunks,as we have seen, but for quite different reasons, and those reasons are relevant to thelikelihood that any given drunk will be arrested. (p. 272)

This exercise has also confirmed the judgment offered in the beginning of thisstudy—that a full explanation of police style requires first hand observation of thebehavior of the police in order to discover what that style is. (p. 277)

Liederbach, Travis / Wilson Redux 449



Wilson realized that official statistics fail to capture a range of officer behaviorsthat are indicative of different styles of policing. For example, official statistics will notprovide information regarding citizens who the police encounter outside the traditionallaw enforcement realm, including disputants or other types of criminal suspects who arenot arrested and/or cited, service recipients, witnesses or third parties not containedin official reports, and citizens who encounter officers on a casual and/or personalbasis. In short, official statistics cannot directly measure the frequency and formalityof police–citizen interactions because they provide information only on the relativelysmall percentage of encounters in which an officer formally invokes the law.

One objective of the current study is to provide a test of Wilson using a moredirect measure of police style based on data derived from the systematic socialobservation (SSO) of police officers in 20 suburban and small-town agencies. SSOallows for the collection of data that are difficult or impossible to collect throughofficial statistics, especially information that pertains to police–citizen encountersthat would otherwise be of low visibility to the researcher. We are aware of no existingtests of Wilson’s theory that incorporate a measure of police style other than thosederived exclusively from official statistics.

A second and perhaps more important goal of our research is to present a criticalassessment of the requirements of a valid test of Wilson’s model. Such an assessmentappears to be long overdue given the enduring impact of Varieties and the scant levelof existing empirical support for the model. This discussion could be used not onlyto identify the shortcomings of existing tests but also to provide a “roadmap” forfuture research designed to test Wilson’s seminal ideas. In this way, police scholarscould more accurately evaluate the relative merits of one of the most influential theoriesof police behavior and begin to formulate ongoing judgments as to whether the theoryshould continue to hold sway. Before describing the methods used in the currentstudy, we begin with an overview of Wilson’s theory of police style and the studiesdesigned to test its basic assumptions.

Wilson’s Theory and Previous Tests

In Varieties, Wilson established that local police agencies differ among themselvesin terms of the ways in which they approach the tasks of policing, and he identifiedand defined three distinct police styles: legalistic, service, and watchman. The differencesamong these styles are not based on the different activities performed by the agencybecause every police agency must serve the three functions of law enforcement, servicedelivery, and peacekeeping. Rather, police agencies differ in terms of the ways inwhich officers employed by the agency approach police problems. Thus, the distinc-tion is not so much what they do but how they do it (Travis & Langworthy, 2008).

Wilson arrived at these styles using a 2 × 2 typology based on the frequency andformality of police–citizen interactions. For example, officers in a legalistic agency




intervene frequently and formally with citizens. Employees of a legalistic style agencyrely on the formal criminal law as a definition of not only what situations merit theirattention but also in what manner they should intervene. Hence, a legalistic departmentis characterized by officers who issue many citations and make many arrests. Officersin a service style agency also intervene frequently with citizens, but they do so infor-mally. The officers in a service style department treat every citizen complaint asrequiring a police response; they do not rely on the criminal code to define policeissues. As the label suggests, these agencies stress the service delivery function ofpolice. Officers in a watchman style agency do not interact with citizens frequently.These agencies stress the peacekeeping function of policing. Officers in a watchmanstyle department exercise considerable discretion in defining whether a situationrequires a police response and what that response should be; therefore Wilson didnot specifically describe the formality of police–citizen interactions within thesedepartments (Travis & Langworthy, 2008).

Wilson theorized that these styles were related to the characteristics of the popula-tion and the form of local government existing within the city. For his test, Wilson useddata from 146 cities with between 25,000 and 100,000 residents for which he couldobtain measures of government type as well as Uniform Crime Report data. Wilsoneliminated cities that were either very rich or very poor, or had a non-White populationin excess of 5% of the total. By eliminating these cities he hoped to control for the effectof actual crime rates. His measure of political culture included type of local government,professionalism of the manager, and form of local elections. Local political culture wasultimately classified in terms of four categories: partisan mayor-council (or “tradi-tional”), nonpartisan mayor-council, council-manager with professional managers (or“professional”), and council-manager with low levels of professionalism. He opera-tionalized police style using official arrest data for larceny, drunkenness, driving whileintoxicated (DWI), disorderly conduct, and simple assault.

Wilson found support for the theory when the analysis showed higher levels ofarrest for larceny, drunkenness, and DWI for “professional” government cities andhigher levels of arrest for disorderly conduct in partisan government cities. In general,professional government cities demonstrated legalistic styles and partisan citiesshowed watchman styles. In what he called intermediate cities (those with managerswho did not score on measures of professionalism or places with nonpartisanelections) the arrest rates were mixed. These cities demonstrated high rates of arrestfor drunkenness and larceny, and all manager cities (regardless of level of profes-sionalism) showed higher rates of arrest for DWI.

Previous Tests

Studies designed to test Wilson (1968) have focused on the relationship betweenlocal political culture and (a) organizational characteristics and structures, and (b)police style. In terms of tests concerning influences on police organizations, Pursley




(1976) found that the number of police departments with “reform” police chiefs wassignificantly higher in communities with professional governments than in those havingtraditional mayor-council governments. Langworthy (1986) tested the relationshipbetween local political culture and the organizational arrangements of police orga-nizations. His findings did not support Wilson’s theory that police organizations areconstrained by type of local government. More recently, Hassell et al. (2003) foundthat local political culture had no significant impact on the organizational structuresof a sample of 401 large police agencies.

More closely related to our purposes, a few existing studies have specificallytested Wilson’s hypotheses regarding the relationship between local political cultureand police style (Crank, 1990, 1992; Langworthy, 1985; Zhao & Hassell, 2005; Zhaoet al., 2006). As noted previously, all these tests use a measure of style derived fromofficial arrest/citation statistics. Langworthy (1985) used data on 152 large cities(population > 100,000) and found that local government type and police style (DWIand disorderly conduct arrests) are correlated, but that political culture is not sufficientto explain agency style. That is, cities with professional forms of government tendto display legalistic policing, and cities with traditional forms of government tend todisplay a watchman style; however, within group variation was substantial such that“city options tend to be congruent with a particular political culture, but do notappear to be constrained to do so” (p. 98).

Crank (1990) tested the influence of organizational factors, local government, andselected community structural characteristics on the arrest rates for minor crimes(disorderly conduct, trespass, cannabis control, motor vehicle offenses). He found thatthe presence of professional local governments was associated with more legalisticpolice styles. Moreover, environmental factors (e.g., local government type, racial/cultural heterogeneity, economic conditions) appeared to exert more influence on thestyle of “rural” rather than “urban” departments. Similar to his earlier test, Crank (1992)found that local government structure (city manager form) was associated with legalismin terms of the arrest rates for more serious crimes such as aggravated assault, burglary,and theft. As Zhao et al. (2006) point out, however, these tests did not employ Wilson’s(1968) original measure of political culture that includes election type (partisan ornonpartisan).

More recently, Zhao and Hassell (2005) studied the style of approximately 500police agencies that employed 100 or more officers similar to Langworthy (1985).They trichotomized local government type into good government, traditional, andmixed. They concluded that their analysis showed “little support” for a relationshipbetween local political culture and police practices (p. 411). Zhao et al. (2006) testedthe relationship between local political culture and police practices in 281 munici-palities with populations of more than 25,000 using a three-wave panel survey andofficial arrest statistics. Specifically, they tested for a relationship between form oflocal government and the official arrest rates for larceny/DWI (legalistic style) anddrunkenness/disorderly conduct (watchman style). They found no evidence of an




effect of local government on police practices, and they identified a need for addi-tional tests of Wilson since their analyses revealed “little evidence” to support theapplication of Wilson’s theory to contemporary policing (p. 577).

The Present Study

The research reported here relies on data collected during direct observations ofpolice officers conducted by a research team from the University of Cincinnati,Division of Criminal Justice, as part of a larger project funded by the NationalInstitute of Justice.1 The police departments included in the present study are locatedin Southwestern Ohio within the greater Cincinnati Metropolitan area, but not thecity of Cincinnati (see Table 1). The departments employ an average of fewer than20 sworn officers (386 total) and serve a combined population of approximately219,000. The jurisdictions include 15 suburban communities located in HamiltonCounty (Ohio) and four communities located in Clermont County, which bordersHamilton County to the east and is more geographically isolated from Cincinnati’surban core. The remaining study site straddles the borders of three adjacent counties(Hamilton, Clermont, and Warren counties).

The study sites are all small in terms of their population compared with moreurbanized central cities and most are homogeneous in terms of their racial composi-tion (more than 90% White on average); however, they are fairly diverse in terms ofthe other demographic and structural characteristics listed in Table 1. In purelydescriptive terms, these sites include two affluent residential suburbs (AmberleyVillage, Terrace Park), five middle-class residential suburbs (Colerain Twp., DelhiTwp., Loveland, Milford, and Forest Park), two middle-class suburbs that are highlycommercialized (Blue Ash and Sharonville), six working-class suburbs (Cheviot,Deer Park, Reading, Arlington Hts., Fairfax, and Lockland), and five small townsthat are more sparsely populated and rural in character than the suburban jurisdic-tions (Amelia Village, Goshen Twp., Harrison, Felicity, and Williamsburg).

Data and Methods

The methodology used in the field was systematic social observation (Mastrofskiet al., 1998). The main procedures used in SSO include the development of instru-ments to collect and record observations systematically, and the subsequent investi-gation of those activities through direct observation (Reiss, 1971). SSO differs fromthe qualitative and ethnographic methodologies used by researchers to study patrolofficers in the 1960s and 1970s in terms of (a) the use of predesigned protocols anddata collection instruments, (b) the employment of a large number of observers, and(c) the inclusion of large samples of officers (National Research Council, 2004).




454

Tabl

e 1

The

Stu

dy S

ites

Med

ian

(%)

(%)

Popu

latio

n (%

) (%

) L

and

(%)

No

Swor

n H

ouse

hold

W

hite

O

wne

r C

hang

e N

on-

(%)

Use

M

edia

n (%

) Age

(%

) Age

(%

) C

olle

ge

H.S

. O

ffic

ers

Popu

latio

nIn

com

eC

olla

rO

ccup

ied

1990

-200

0W

hite

Mar

ried

Res

iden

tial

Age

<21

>65

Gra

duat

eD

egre

e

Am

berl

ey V

illag

e15

3,42

5$8

1,49

260

.698

.010

.213

.175

.291

48.2

24.6

24.4

63.5

5.3

Am

elia

42,

752

$44,

900

29.0

54.4

49.8

3.7

58.1

n/a

29.1

32.3

7.6

13.8

16.5

Arl

ingt

on H

ts.

1389

9$3

0,28

814

.956

.5–1

7.1

8.1

44.9

4934

.630

.112

.24.

631

.8B

lue

Ash

3412

,513

$61,

591

53.2

74.3

5.5

13.6

59.6

4239

.428

.214

.248

.79.

2C

hevi

ot23

9,01

5$3

5,15

027

.461

.6–6

.23.

750

.275

35.5

25.6

17.8

16.4

22.8

Col

erai

n Tw

p.24

60,1

44$4

9,96

029

.281

.3n/

a12

.258

.577

35.8

31.0

11.9

18.8

16.4

Dee

r Pa

rk11

5,98

2$3

9,69

229

.171

.9–3

.23.

950

.578

38.0

24.4

20.3

16.9

20.0

Del

hi T

wp.

3230

,104

$55,

052

31.9

84.9

n/a

2.3

62.4

8636

.732

.714

.723

.711

.3Fa

irfa

x12

1,93

8$4

1,41

831

.880

.8–4

.53.

652

.545

36.3

30.8

14.7

15.7

26.1

Felic

ity2

922

$20,

781

22.1

44.8

7.7

3.1

48.2

n/a

31.1

34.9

12.3

6.7

42.9

Fore

st P

ark

4019

,463

$49,

298

34.3

62.5

4.6

63.9

52.1

5533

.930

.79.

126

.211

.7G

oshe

n Tw

p.7

13,6

63$4

6,31

418

.687

.2n/

a1.

863

.8n/

a34

.232

.58.

48.

424

.1H

arri

son

217,

487

$46,

107

25.5

72.0

–0.4

2.2

57.9

n/a

32.4

35.1

9.3

11.9

22.0

Loc

klan

d18

3,70

7$2

8,29

217

.847

.9–1

4.9

30.1

40.5

4735

.628

.011

.98.

628

.6L

ovel

and

2311

,677

$52,

738

38.5

74.6

16.9

5.1

61.2

8235

.832

.411

.035

.010

.4M

ilfor

d14

6,28

4$3

1,92

331

.749

.811

.05.

550

.2n/

a39

.125

.721

.323

.922

.1R

eadi

ng18

11,2

92$3

9,14

025

.159

.2–6

.28.

649

.863

37.6

26.7

16.0

15.9

21.0

Shar

onvi

lle56

13,8

04$4

7,05

540

.863

.64.

912

.553

.238

39.1

23.6

17.1

33.8

12.2

Terr

ace

Park

142,

273

$95,

530

59.9

95.0

6.6

1.8

73.5

9738

.738

.012

.471

.01.

5W

illia

msb

urg

52,

358

$37,

115

20.1

60.0

1.6

1.4

51.5

n/a

33.4

31.2

10.9

10.4

26.8

Sour

ce:2

000

U.S

. Cen

sus:

HU

D S

tate

of

the

Citi

es D

ata

Syst

ems

(SO

CD

S).

n/a

=no

t app

licab

le.



Observed officers were accompanied by trained observers who recorded every-thing the officers did during their entire shift, including the activities that the officersperformed and information regarding the nature of their interactions with citizens.2

The present study uses a framework similar to that outlined by Mastrofski et al.(1998) to systematically organize and record the observer’s field notes into reliablecoded data. Coding instruments were developed to record information regarding theofficer’s interactions with citizens including (a) the encounter/activity instrumentand (b) the citizen instrument.

The encounter/activity instrument accounted for every minute of the observedofficer’s shift time, whether they were in direct contact with citizens or performingtasks that did not include citizens, such as motor patrol, report writing, or driving enroute to and from locations. A citizen instrument was completed for every citizenencountered by the observed officers. The citizen instrument provided informationconcerning the nature and character of the interaction including the citizen’s demo-graphic characteristics, demeanor, offenses (if any), types of problems encountered,and other situational factors related to each encounter. These situational factorsincluded requests made by citizens to the officer, police law enforcement actions(e.g., arrests, citations, searches, use of force), officer requests made to citizens, andfactors surrounding any disputes that occurred among citizens.

Data were collected over a 14-month period between April 1999 and May 2000.The research team randomly selected shifts to be observed within the 20 agencies tocomplete 2.5 observations per month per department over the course of the 14-monthproject.3 Observations were made with 213 individual officers employed by thesampled agencies. A total of 583 observations were completed encompassing 4,813 hrof observation, or the equivalent of over 611 8-hr shifts.

Advantages and Disadvantages of Observational Data

In terms of the present study, the primary advantage of observational data is thatit allows the researcher to collect information on citizens who interact with thepolice and who are not arrested and/or cited, including criminal suspects/disputantswho were not arrested and/or cited, witnesses and/or third parties, victims, servicerecipients, and citizens who encounter police on a casual basis or as friends. Theinclusion of these citizens improves on existing methodologies that have usedofficial arrest/citation records as an indicator of police style in at least two ways.First, the observational data provide a more valid measure of “frequency” sinceinformation on police–citizen encounters that fall outside the realm of traditional lawenforcement are included. This advantage is especially evident in cases in whichpolice interact with citizens informally and/or those encounters that involve citizenswho are primarily recipients of police services rather than criminal suspects. Second,the observational data allow for a measure of “formality” that is relative to all otherobserved police–citizen interactions, and hence controls for frequency of interaction.




In contrast, official arrest/citation statistics cannot provide this type of measurebecause they do not contain information on situations wherein officers choose not toarrest and/or cite citizens who have violated the law.

Although observational data can be an especially effective means to garner informa-tion on citizen encounters that would otherwise be of low visibility to the researcher, themethod does lend itself to the danger of reactivity. An officer and/or citizen may changehis or her behavior simply because of the presence of the observer. The research teamendeavored to limit the danger of reactivity by several means: (a) The research teampromised confidentiality to individual officers who were observed as is required by thefederal sponsoring agency. Confidentiality agreements were signed by all observers inaccordance with these requirements. (b) A certain level of confidence and familiaritybetween the observers and the officers was established. All observers were instructed tooutline the confidential nature of the information they obtained to the officer at thebeginning of each shift, and they told officers that they were free to view the observer’snotes at any time during the observation. (c) Observers were instructed to avoid takingnotes while in the presence of citizens, so as to avoid citizen inquiries and/or influenc-ing the events that were observed. If necessary, observers would record informationimmediately after the encounter occurred, or while the officer performed routine patrol.Observers were instructed to indicate whether or not they believed observed officersand/or citizens had altered their behavior because of the observer’s presence. Indicationsof reactivity on the part of either the observed officer and/or citizens that were apparentto the observer occurred very infrequently. Observers indicated that some sort of reac-tivity occurred in 34 of the 17,480 observed officer activities and citizen interactionsrecorded over the course of the study.

Observational data also have a potential problem with reliability in terms of main-taining consistent coding procedures across observers. The research team endeav-ored to limit these concerns through observer training courses conducted prior to andduring the study period. As Reiss (1971) suggests, observers viewed videotapes topractice coding typical police activities. In addition to these training sessions,observer data were reviewed and cleaned on an ongoing basis during the length ofthe project to ensure that all observers were following the coding standards set forthduring the training sessions, thus providing a continuing check on the reliability ofthe observation data.

Measure of Police Style: The Frequency andFormality of Police–Citizen Interaction

We developed a measure of police style based on the observed frequency andformality of police–citizen interaction in the 20 jurisdictions (see Table 2). We initiallycalculated frequency as the total number of police–citizen encounters within eachjurisdiction per standardized 8-hr shift.4 This measure of frequency includes not onlythose encounters that included an arrest and/or citation, but also all other types of




police–citizen encounters that occurred during the observed shift. Following Wilson’s2 × 2 conceptualization of police style, we subsequently calculated the averagenumber of citizen interactions across the 20 sites. Jurisdictions in which police–citizenencounters exceeded the median were ranked “high” on frequency with the remainderbeing ranked “low.”

Similarly, we calculated formality in terms of the number of encounters thatinvolved an arrest and/or citation in the 20 jurisdictions per standardized 8-hr shift.To obtain a measure of formality that was standardized relative to frequency, wedivided the number of encounters with an arrest and/or citation per shift by totalencounters per shift for each department. Once again following Wilson’s 2 × 2scheme, we ranked the study sites relative to the median in terms of “high” and“low.” Table 2 presents the rankings of the 20 study sites in terms of frequency/for-mality and corresponding police style, including five “legalistic” agencies (high fre-quency/high formality) and five service agencies (high frequency/low formality).The remaining 10 agencies were defined as “watchman” because they exhibitedinfrequent citizen interaction relative to the other agencies; however because we hada measure of formality based on outcomes of citizen encounters, we were able toidentify two different types of “watchman”: (a) a traditional Wilsonian “watchman”with low frequency and low formality and (b) a “formal watchman” where interac-tions with citizens were infrequent, but tended toward formal outcomes.

Findings

We sought to identify correlates of police–citizen interaction based on our classi-fication of the agencies in terms of frequency and formality. Initially, we investigatedfor any association between the local political culture of the 20 communities andpolice style. We collected information pertaining to the local political culture of eachof the communities including form of local government and type of local election.Table 3 presents a classification of the 20 communities in terms of local political cul-ture.5 As Table 3 indicates, four communities conform to Wilson’s professional typeof local political culture with council-manager forms of government and nonpartisanelections (Amberley Village, Forest Park, Loveland, and Milford). Likewise, sixcommunities conform to Wilson’s traditional type of local political culture withmayor-council forms of government and partisan elections (Cheviot, Deer Park,Harrison, Lockland, Reading, and Sharonville). Of the remaining communities, sixexhibited a “mixed” political culture with mayor-council forms of government andnonpartisan elections (Amelia, Arlington Hts., Felicity, Fairfax, Terrace Park, andWilliamsburg). There were four communities that did not readily fit Wilson’sscheme. The city of Blue Ash has a council-manager form of local government withpartisan elections. Colerain, Delhi, and Goshen are townships that are governed byeither a professional administrator or a township trustee.6




458

Tabl

e 2

Pol

ice

Styl

e:F

requ

ency

and

For

mal

ity

of C

itiz

en I

nter

acti

ons

Enc

ount

ers

Obs

erve

d To

tal

Freq

uenc

y E

ncou

nter

s w

/arr

est/c

ite

Form

ality

Sh

ifts

aE

ncou

nter

sFr

eque

ncyb

Ran

kw

/arr

est/c

itepe

r Sh

ift

Form

ality

cR

ank

Styl

e

Col

erai

n Tw

p.33

.00

230

6.96

HIG

H46

1.39

39.2

002

HIG

HL

EG

AL

IST

ICFo

rest

Par

k32

.42

184

5.67

HIG

H40

1.23

38.2

176

HIG

HL

EG

AL

IST

ICG

oshe

n Tw

p.22

.31

146

6.54

HIG

H38

1.70

32.2

604

HIG

HL

EG

AL

IST

ICH

arri

son

30.7

920

26.

56H

IGH

662.

1435

.326

7H

IGH

LE

GA

LIS

TIC

Shar

onvi

lle32

.64

205

6.28

HIG

H43

1.31

74.2

097

HIG

HL

EG

AL

IST

ICD

eer

Park

29.8

916

65.

55H

IGH

240.

8029

.144

6L

OW

SER

VIC

ED

elhi

Tw

p.33

.63

199

5.91

HIG

H34

1.01

10.1

710

LO

WSE

RV

ICE

Felic

ity21

.27

156

7.33

HIG

H10

0.47

01.0

641

LO

WSE

RV

ICE

Lov

elan

d32

.06

200

6.23

HIG

H34

1.06

05.1

702

LO

WSE

RV

ICE

Rea

ding

32.3

620

76.

39H

IGH

391.

2051

.188

5L

OW

SER

VIC

EA

mbe

rley

Vill

age

28.4

793

3.26

LO

W7

0.24

58.0

753

LO

WW

AT

CH

Arl

ingt

on H

ts.

30.4

012

54.

11L

OW

170.

5592

.136

0L

OW

WA

TC

H

Blu

e A

sh30

.51

168

5.50

LO

W32

1.04

88.1

906

LO

WW

AT

CH

Milf

ord

32.7

415

74.

79L

OW

200.

6108

.127

5L

OW

WA

TC

H

Will

iam

sbur

g28

.95

143

4.93

LO

W21

0.72

53.1

471

LO

WW

AT

CH

Am

elia

31.9

712

33.

84L

OW

280.

8758

.228

0H

IGH

WA

TC

H (

form

al)

Che

viot

30.2

215

95.

26L

OW

311.

0258

.195

0H

IGH

WA

TC

H (

form

al)

Fair

fax

32.9

116

14.

89L

OW

631.

9143

.391

4H

IGH

WA

TC

H (

form

al)

Loc

klan

d32

.02

156

4.87

LO

W43

1.34

29.2

757

HIG

HW

AT

CH

(fo

rmal

)Te

rrac

e Pa

rk33

.27

150

4.50

LO

W37

1.11

21.2

471

HIG

HW

AT

CH

(fo

rmal

)

a. 8

-hr

shif

t equ

ival

ent.

b. E

ncou

nter

s pe

r st

anda

rdiz

ed 8

-hr

shif

t.c.

Enc

ount

ers

resu

lting

in a

n ar

rest

and

/or c

ite p

er s

tand

ardi

zed

shif

t / E

ncou

nter

s pe

r sta

ndar

dize

d 8-

hr s

hift

(per

cent

age

of e

ncou

nter

s re

sulti

ng in

form

al a

ctio

n).



The right-most column of Table 3 lists the previously identified style of the agen-cies and can be used to provide a crude examination as to whether or not the stylesare related to political culture in our sample. For example, Wilson suggests that aprofessional local political culture tends to be correlated with either a legalistic orservice style; however, in our sample two of the four communities with a profes-sional local political culture exhibited a watchman style, one agency was legalistic,and the remaining agency exhibited a service style. Likewise, Wilson suggests thata traditional local political culture tends to be correlated with the watchman style,but only two of the six communities with a traditional political culture in our sam-ple exhibited a watchman style, two were legalistic, and the remaining two agenciesexhibited a service style. Wilson found no consistent patterns for cities with a mixedor intermediate local political culture in his test. In our sample, these communitiestended to be “low” in terms of police–citizen interaction, and hence watchman.There were no patterns in terms of formality, as three of the six mixed communitieswere “formal” watchman, two were low formality Wilsonian watchman, and oneexhibited a service style. The four communities that did not fit Wilson’s political cul-ture scheme evidenced no pattern in terms of police style.

We next sought to identify correlates of police style using local political cultureand its indicators (election type, type of local government), as well as the general


Table 3Police Style and Political Culture

Government Type Election Type Political Culture Style

Amberley Village Council-manager Nonpartisan PROFESSIONAL WATCHForest Park Council-manager Nonpartisan PROFESSIONAL LEGALISTICLoveland Council-manager Nonpartisan PROFESSIONAL SERVICEMilford Council-manager Nonpartisan PROFESSIONAL WATCHCheviot Mayor-council Partisan TRADITIONAL WATCH (formal)Deer Park Mayor-council Partisan TRADITIONAL SERVICEHarrison Mayor-council Partisan TRADITIONAL LEGALISTICLockland Mayor-council Partisan TRADITIONAL WATCH (formal)Reading Mayor-council Partisan TRADITIONAL SERVICESharonville Mayor-council Partisan TRADITIONAL LEGALISTICAmelia Mayor-council Nonpartisan MIXED WATCH (formal)Arlington Hts. Mayor-council Nonpartisan MIXED WATCHFairfax Mayor-council Nonpartisan MIXED WATCH (formal)Felicity Mayor-council Nonpartisan MIXED SERVICETerrace Park Mayor-council Nonpartisan MIXED WATCH (formal)Williamsburg Mayor-council Nonpartisan MIXED WATCHBlue Ash Council-manager Partisan OTHER WATCHColerain Twp. Professional administrator Nonpartisan OTHER LEGALISTICDelhi Twp. Professional administrator Nonpartisan OTHER SERVICEGoshen Twp. Township trustee Nonpartisan OTHER LEGALISTIC



community characteristics presented in Table 1 for the 20 communities. Chi-squareanalyses were performed on the data to discern the degree of association betweenpolice style and these community characteristics (see Table 4). As Table 4 shows,type of election (partisan or nonpartisan), type of government (council-manager ormayor-council), and local political culture (professional, traditional, or mixed) werenot significantly associated with police style (legalistic, service, watchman).

Of the remaining 13 community characteristics, only population (χ2 = 8.800,p < .05) and median household income (χ2 = 6.800, p < .05) were significantly associatedwith police style. In terms of population, all five legalistic departments (frequentpolice–citizen interaction) were above the median for population, and 8 of the 10watchman departments (infrequent police–citizen interaction) were below the medianfor population. In terms of median household income, all five legalistic departmentswere above the median, and 7 of the 10 watchman departments were below the median.As these findings indicate, population and median household income were highlycorrelated in our sample, with 8 of the 10 largest communities in terms of populationhaving higher than median incomes. An additional chi-square analysis was performedusing “formal” watchman as a fourth distinct style, which resulted in population beingthe only community factor significantly associated with style.


Table 4Community Characteristics and Police Style*

df Na χ2 p

Population 2 20 8.800 .012*Median household income 2 20 6.800 .033*Number of sworn officers 2 20 3.600 .165Median age 2 20 3.600 .165Political culture 4 16 3.593 .464Greater than 65 (%) 2 20 2.400 .301Married (%) 2 20 2.400 .301Residential (%) 2 20 2.400 .301Type of election 2 20 0.952 .621Owner occupied (%) 2 20 0.800 .670No h.s. diploma (%) 2 20 0.800 .670College grad. (%) 2 20 0.800 .670Less than 21 (%) 2 20 0.800 .670Non-White (%) 2 20 0.400 .819White-collar (%) 2 20 0.400 .819Population change (1990-2000) 2 17 0.275 .871Type of local government 2 17 0.234 .889

a. Table incorporates “formal watchman” into the “watchman” category.*p < .05.



Discussion

For almost 40 years, one conception of local policing has dominated thinking aboutwhat police agencies do and how they do it. Few empirical tests of this conceptual-ization have been reported, and of those none completely supports the dominantmodel. This article adds to that body of work reporting a test of the model derived fromthe observation of officers from a sample of small-town and suburban police depart-ments. In contrast to previous tests, the present study uses a unique measure of stylederived from the observational data. Our results, like those of earlier tests of Wilson’s(1968) model, failed to provide support for the hypotheses that local political cultureis related to police style. Local government characteristics such as partisan electionsand type of local government were not correlated with either frequency of interaction,formality of interaction, or police style.

We did identify variation in the frequency and formality of police–citizen interactionssuggesting the existence of different police styles. We found a significant relationshipbetween population size and style of policing. The larger communities tended towardthe legalistic style whereas smaller communities tended toward the watchman style.Since population and median household income were highly correlated in our sample,the less wealthy communities—which also tended to be among the smallest in oursample—were more likely to exhibit a watchman style.

In his description of the various styles, Wilson describes expected links betweenorganizational characteristics and police style. He suggests that police style will beassociated with specialization in the department and with the level of pay and benefitsoffered to officers. In short, the ability of the organization to reward or punish officerbehavior influences the ability of the police administrator to direct that behavior. Whilewe did not have the data required to test for this relationship, anecdotal informationsuggests that the smaller, poorer agencies had little specialization and paid relativelylow police salaries. It is interesting to note that larger communities (and normallylarger police agencies with more specialization) tended toward the legalistic style ofpolicing. Unfortunately we were unable to directly assess these issues with the dataavailable.

Still, overall our findings mean that once again, Wilson’s model was not supportedby an attempt to replicate his original test of the theory. There are several possibleexplanations for this, including the possibility that Wilson’s explanation of policeorganizational styles is wrong. In fact, others who have recently conducted tests thathave failed to support Wilson’s work have intimated that the accumulation of mostlynegative findings in the existing literature has moved us closer to a falsification ofWilson’s ideas (Zhao et al., 2006; Zhao & Hassell, 2005). Certainly, findings fromthis study could be used as additional evidence in favor of abandoning what has beenone of the most influential theories of police behavior.

We would argue, however, that a summary dismissal of Wilson (1968) based on theevidence provided by existing tests would be premature on several grounds. First, we




believe that no one, ourselves included, has managed to faithfully replicate Wilson’soriginal test. When Wilson first subjected his theory to an empirical test, he recog-nized that his observations might be limited to the eight communities that had beenstudied in depth for the book (p. 271). He suggested that measuring political culturewould require extensive surveys of the public and that measures of police policieswould be even more difficult. He stated that police agency style is not always bestmeasured by arrests since what matters is not the number or rate of arrests, butrather, the reasons that such arrests are made (p. 273). Nonetheless, arrest rates foroffenses that allowed high levels of police discretion were his measure of agencystyle. Wilson attempted to control for differences in the true crime rate by limitinghis sample to medium-sized, middle class, predominantly White cities. He used afour-category scheme to operationalize local government type.

Given these parameters, no one has fully replicated the original test. For instance,Langworthy (1985) studied cities with a population of 100,000 or more and wasunable to measure the professionalism of city managers. He did not control formedian family income or for percentage non-White. Zhao and Hassell (2005) studiedpolice agencies (not cities) employing 100 or more officers. These agencies tend toserve municipalities with populations of 50,000 or more. They, like Langworthy (1985),trichotomized local government type into good government, traditional, and mixed.In this study, we examined police styles in midwestern towns and suburbs with com-paratively small populations and police organizations that employed relatively fewofficers. Owing to the constraints associated with conducting large-scale observa-tional studies, there were only 20 communities in our sample, and it could be arguedthat all of them lack the cultural and racial heterogeneity needed to discern substantivedifferences in police styles. Though we believe our measure of police style moredirectly indicates the frequency and formality of police–citizen interaction, it is notthe measure that was used by Wilson.

Each of these studies contained important departures from Wilson (1968),whether in the selection of communities, sampled agencies, or in the measurementof key variables. It may well be that structural characteristics such as population, thepercentage of non-White residents, and aggregate income levels are not simply correlatesof crime rates as originally hypothesized by Wilson, but also correlates of politicalculture as well. Beyond that, Wilson’s original description of political culture andhow it influenced police style included discussions of population homogeneity inrace, class, and income as well as police organizational characteristics such as officersalary, specialization, centralization, and civilianization. Neither Wilson nor any ofthose following his first empirical test has included these factors in the analysis.

Though the absence of any true replications of Wilson (1968) makes us hesitant tosupport notions of a movement toward falsification, we would also argue that negativejudgments as to the original validity of Wilson’s work seem misplaced given the timethat has elapsed since the initial publication of Varieties. As others have alreadynoted, large-scale changes have occurred within the political, social, and occupational




context surrounding policing (Hassell et al., 2003; Zhao & Hassell, 2005; Zhao et al.,2006). These changes have probably altered the manner in which policing is accom-plished since the 1960s, so much so that our ability to “recapture” what Wilsonobserved four decades ago has probably been irrevocably lost.

For example, Hassell et al. (2003, p. 244) highlight changes that have occurred in therelationship between police and the local political culture, whereby progressive-erareforms designed to limit political influences on police organizations have clearlysolidified since the 1960s. It may also be that Wilson’s (1968) measurement of localpolitical culture is too simplistic for the study of contemporary political arrangements(Hassell et al., 2003). In terms of the social context, it is likely that public expectationsof the police have become more homogenized with the advent of mass culture depictionsof policing and their wide-scale dissemination through television, movies, and otherforms of media. Likewise, the advent of the “information age” and associated growth incommunications technologies may have reduced the isolation, and hence uniqueness,of many smaller police agencies. DWI, for example, is probably not the “highly dis-cretionary” situation that it was in earlier years because changes to the sociolegaldefinition of DWI appear to have altered public perceptions concerning the relativeseriousness of the offense. Finally, police styles have almost certainly been alteredby changes to the occupation itself, including (a) growth in the number of womanand minority officers, (b) more demanding educational and training requirements,(c) the advent of nationwide recruitment of officers, and (d) the steadily expandinginfluence of state and federal governments through funding incentive programs andthe spread of “unfunded mandates.”

The development and spread of higher education for law enforcement personneland the standardization of police training curricula also work to reduce variation inpolice behavior across jurisdictions. Contemporary police officers are likely to berecruited from outside the jurisdiction in which they will serve and be trained undera statewide curriculum. When Varieties was published, it was more common forpolice agencies to recruit locally and to provide agency-specific, idiosyncratic training.So too, changes in the preparation of city officials, especially the growth in the numberof “professional (college trained)” city managers has greatly reduced the likelihoodof finding variation in types of local government structure. Jurisdictions that havemanager-council forms of government overwhelmingly employ what Wilson definedin 1968 as “professional” managers.

Given these and countless other shifts within the context that surrounds police work,it is probably too much to expect that Wilson’s conceptualization would survive thetest of time unscathed. Nor should it be surprising that a 40-year-old model cannotadequately explain policing styles in contemporary American society. In this respect,current failures to replicate Wilson (1968) should probably not be interpreted asadditional fodder toward the ultimate falsification of his theory, but rather, as evidenceattesting to the fact that things have indeed changed, and as a call to action for contem-porary police scholars to set on the task of updating Wilson’s theory and explaining howthese changes impact the behavior of today’s police.




To this end, Wilson’s broad ideas should retain an intuitive—albeit quantitativelychallenged—appeal. Wilson himself recognized the complexities associated withstudying the relationship between political culture and police style, as well as thecrudity of his own measures. In describing the connection between local politicalculture and police styles, Wilson (1968) had this say,

The most important way in which political culture affects police behavior is through thechoice of police administrator and the molding of the expectations that govern his role.Just as the most important decision a school board makes it its choice of superintendent,so the most important police decision a city council makes (or approves) is the selectionof chief. In some communities, it is expected that he will be the “best man available”; inothers it is that he will be the “deserving local fellow” or the man “closest to the party.”And once in office, the chief will confront a zone of community indifference to hispolices of varying dimensions . . . the political culture acts as a filter, different for eachcommunity, that screens out certain complaints and demands, leaving the chief free toignore them, and passes through (or even amplifies) others. (pp. 233-324)

Thus, it is likely that this “filtering” process—whereby local community normsand values are distilled in a manner that is unique to each community—makes itexceedingly difficult to understand the connection between communities, politicalculture, and police behavior in purely quantitative terms. The study of local politicalculture, Wilson (1968, p. 277) wrote, requires “firsthand” knowledge of the communityand a full understanding of police style requires “firsthand observation.” In short,Wilson described a qualitative phenomenon best studied with qualitative methods.Attempts to test or validate Wilson’s theory with limited quantitative measures arenot likely to capture the nuanced nature of local political culture and the linksbetween that culture and police practice.

One of the important limits of the extant research is the failure, as in our own case,to measure the mechanisms through which local political culture can be transmittedto individual officers. Wilson devoted considerable attention to the organizationaland structural characteristics of the different types of police departments. His modelspecifically states that the link between local political culture and police style is indirect.Future research should attempt to specify the paths through which local politicalculture influences police behavior.

Laura Huey recently studied police practices on “skid row” in three locations:Edinburgh, Scotland; San Francisco, California; and Vancouver, British Columbia.Her ethnographic study involved observation and interview techniques and revealedimportant differences in how the police dealt with citizens on skid row in the threelocations. Huey (2007) concludes that “perhaps the most important conclusion thatI took away from this research: the idea that communities articulate their valuesthrough their policing practices” (p. 210). Forty years after the publication ofVarieties, Huey’s research indicates that local variation in police practice remains.




In our own study, anecdotal and experiential knowledge of the local communitieshelps to explain the differences in style. For instance: the palpable suspicion of anaffluent suburb; the busy transience of commercial hubs; the emptiness of some ruraltowns. Likewise, so does firsthand knowledge of the activities officers typically performin different agencies, like controlling weekend crowds of unruly juveniles, enforcingtraffic codes along a local stretch of interstate highway, or the casual greeting offamiliar citizens along a main street. This firsthand knowledge can often provide morepowerful explanations than correlational analyses. Of course, these explanations arecolored by our understanding of Wilson’s original model and may reflect more ofour application of Wilson’s theory than an independent confirmation of the theory.

At the end of the day what remains are two observations. First, whether the productof local political culture or some other factor, most observers of police, and the Americanpolice in particular, accept that there is variation between local police departments.Second, both the understanding of that variation and its explanation are likely to bequalitative phenomena for which we have not yet developed sufficiently precisemeasures to enable sophisticated quantitative analyses. To understand variation inlocal policing, we may need to devote more time to the refinement of the theorybefore we can accomplish empirical testing.

Notes

1. “Policing in a Community Context: An Observational Study of Suburban, Rural, and UrbanPolicing” (Grant #98-IJ-0063) included 21 police agencies in the Greater Cincinnati area. Data collectedduring observations of the Clermont Co. Sheriff’s Office are not included in this study.

2. A total of 43 observers participated in the project. All observers were affiliated with the Universityof Cincinnati either pursuing their PhD degree (27.9%), masters degree (41.8%), or undergraduatedegrees (30.2%).

3. The observation schedules were constructed on a monthly basis. Agencies that used a standard 8-hrshift were randomly assigned ride dates and shift times to complete the goal of 2.5 observations per month.Agencies that did not employ a standard 8-hr shift (e.g., 10- or 12-hr shifts) were randomly scheduled foreither one or two observations per month on a rotating basis to obtain the 8-hr shift equivalent of 2.5 shiftsper month. The choice to randomly select shifts to be observed (rather than officers) was primarily drivenby practical concerns. The observed agencies routinely employed only a few officers on patrol during anygiven shift, and these assignments were often made no more than 1 week in advance. Thus, observationswere conducted with officers who were assigned to patrol on randomly selected dates and shifts, and theresearch team reviewed the observation data on an ongoing basis to ensure that a representative sample ofofficers were observed within each sampled agency.

4. Encounters often involve multiple citizens. Table 2 discusses encounters as a whole, rather thanindividual citizens.

5. Wilson’s measure of political culture included indicators of the “professionalism” of the city managerin terms of (a) attainment of a college degree and (b) ICMA membership. In consideration of earlierfindings indicating that virtually all city administrators today possess a college degree (see, e.g., Zhao, 1996)and consistent with Zhao et al.’s (2006) recent test of Wilson, we excluded college degree attainment from ourmeasure of political culture. Likewise, there was no variability in our sample in terms of ICMA membership,as all of the city managers in our sample were members.




6. Township jurisdictions include all municipalities and unincorporated areas in a given townshiplying within a particular county. This is a popular policing arrangement in several midwestern states,including Ohio, Michigan, and Wisconsin (Falcone, Wells, & Weisheit, 2002).

References

Carter, D. L. (2002). The police and the community (7th ed.). Upper Saddle River, NJ: Prentice Hall.Cohn, E. G., Farrington, D. P., & Wright, R. A. (1998). Evaluating criminology and criminal justice.

Westport, CT: Greenwood.Crank, J. (1990). The influence of environmental and organizational factors on police style in urban and

rural environments. Journal of Research in Crime and Delinquency, 27, 166-189.Crank, J. (1992). Police style and legally serious crime: A contextual analysis of municipal police departments.

Journal of Criminal Justice, 20, 401-412.Falcone, D., Wells, L. E., & Weisheit, R. A. (2002). Small-town police department. Policing: An

International Journal of Police Strategies & Management, 25, 371-384.Hassell, K. D., Zhao, J., & Maguire, E. R. (2003). Structural arrangements in large municipal police organizations:

Revisiting Wilson’s theory of local political culture. Policing: An International Journal of Police Strategies &Management, 26, 231-250.

Huey, L. (2007). Negotiating demands: The politics of skid row policing in Edinburgh, San Francisco, andVancouver. Buffalo, NY: University of Toronto Press.

Langworthy, R. H. (1985). Research note: Wilson’s theory of police behavior: A replication of theconstraint theory. Justice Quarterly, 2, 89-98.

Langworthy, R. H. (1986). The structure of police organizations. New York: Praeger.Mastrofski, S. D., Parks, R. B., Worden, R. E., DeJong, C., Snipes, J. B., & Terrill, W. (1998). Systematic

observation of public police: Applying field research methods to policy issues. Washington, DC:National Institute of Justice.

National Research Council. (2004). Fairness and effectiveness in policing: The evidence. Washington,DC: National Academies Press.

Pursley, R. (1976). Community characteristics and policy implementations: Some exploratory findingsabout two categories of municipal departments. Journal of Criminal Justice, 4, 291-301.

Reiss, A. J. (1971). Systematic observation of natural phenomena. In H. L. Costner (Ed.), Sociologicalmethodology (pp. 3-33). San Francisco: Jossey-Bass.

Roberg, R., Novak, K., & Cordner, G. (2005). Police and society (3rd ed.). Los Angeles: Roxbury.Slovak, J. S. (1996). Styles of urban policing: Organization, environment, and police styles in selected

American cities. New York: New York University Press.Travis, L. F., III, & Langworthy, R. H. (2008). Policing in America: A balance of forces. Upper Saddle

River, NJ: Pearson Prentice Hall.Walker, S., & Katz, C. M. (2008). The police in America. Boston: McGraw-Hill.Wilson, J. Q. (1968). Varieties of police behavior: The management of law and order in eight communities.

Cambridge, MA: Harvard University Press.Wright, R. A., & Miller, J. M. (1998). The most cited scholars and works in police studies. Policing: An

International Journal of Police Strategies & Management, 21, 240-254.Zhao, J. (1996). Why police organizations change: A study of community oriented policing. Washington,

DC: Police Executive Research Forum.Zhao, J., & Hassell, K. D. (2005). Policing styles and organizational priorities: Retesting Wilson’s theory

of local political culture. Police Quarterly, 8, 411-430.Zhao, J., He, N., & Lovrich, N. (2006). Effect of local political culture on police behaviors in the 1990s:

A retest of Wilson’s theory in more contemporary times. Journal of Criminal Justice, 34, 569-578.




John Liederbach is an associate professor in the Criminal Justice Program at Bowling Green StateUniversity. His primary research interests include the study of police behavior across community types, sub-urban and rural policing, and medical crime. He has published in numerous professional journals, includingJustice Quarterly, Policing: An International Journal of Police Strategies and Management, Criminal JusticeReview, and the American Journal of Criminal Justice. He is also coauthor of the forthcoming book PolicePatrol Allocation & Deployment (Prentice Hall).

Lawrence F. Travis III is a professor of criminal justice and director of the Center for Criminal JusticeResearch in the Division of Criminal Justice at the University of Cincinnati. He formally worked as theresearch director of the Oregon State Board of Parole, and as a research analyst with the NationalParole Institutes. He is the author of Introduction to Criminal Justice (Cincinnati, OH: Anderson),coauthor with Robert Langworthy of Policing in America (Englewood Cliffs, NJ: Prentice Hall), and editorof Policing: An International Journal of Police Strategies and Management. He has published on a varietyof criminal justice topics.




police quarterly - university of cincinnati · discover how, if at all, patrolmen in various cities...

Documents