Journal of Criminal Justice 34 (2006) 569 – 578
The effect of local political culture on policing behaviors in the
1990s: A retest of Wilson’s theory in more contemporary times
Jihong Zhao a,⁎, Ni He b , Nicholas Lovrich c
a
Department of Criminal Justice, University of Nebraska at Omaha, Omaha, NE 68182, United States
b
College of Criminal Justice, Northeastern University, Boston, MA 02115, United States
c
Department of Political Science, Washington State University, Pullman, WA 99164, United States
Abstract
In research on policing, James Q. Wilson was among the first scholars to suggest that local political culture constitutes a significant
factor in explaining variation among law enforcement agency practices. Almost forty years after the publication of Varieties of Police
Behavior, a classic study of police organizational behavior, Wilson’s monograph remains the basis of a widely held theory used to
explain variation in police agency behaviors. More specifically, Wilson (1968) identified three distinctive styles of policing: the
legalistic, the watchman, and the service styles. In his empirical work with these styles of policing, Wilson argued that local political
culture was the major determinant of variation in policing styles. The purpose of this study was to retest the validity of Wilson’s
argument in today’s policing environment. Using panel data collected among police agencies across the U.S. surveyed in 1993, 1996,
and 2000, the authors found that there was little evidence to support the application of Wilson’s theory to the practices of
contemporary police organizations.
© 2006 Elsevier Ltd. All rights reserved.
Introduction
Decentralization is a defining feature of American
democracy (e.g., Gabriel, 1956) and the principle of federalism is intended to serve as an institutional guarantee of
decentralization in American political life (Conlan, 1998,
pp. 309–316). A good example of this decentralization is
found in the great variety of local government policies and
special districts serving local government residents across
the country (Lorch, 2001, pp. 227–266). Political scientists had long argued that distinctive political cultures
were present at the local level. For example, Elazar (1975,
pp. 13–16) identified three principal historical sources of
local political culture; in his own words he noted, “Three
⁎ Corresponding author. Tel.: +1 402 554 3928.
E-mail address: jzhao@mail.unomaha.edu (J. Zhao).
0047-2352/$ - see front matter © 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcrimjus.2006.09.011
aspects of political culture stand out as particularly influential in shaping the operation of the American political
system.” They are: (1) the set of citizen perceptions of
what goals are properly pursued through politics, (2) the
kinds of people who become active in government and
politics based on their established traditions, and (3) the
actual way in which the art of governance is practiced as
reflected in the structure and operation of local government (Elazar, 1975; also see Elazar, 1970, 2002). In this
study, the authors examined the effect of the third principal source of local political culture on the variation of
local government agency practices identified by both
Daniel Elazar and James Q. Wilson as critical to the conduct of local government business. Government structures
are clearly known to have a strong impact on policy
outcomes in the area of management styles (Williams &
Adrian, 1963), equal employment opportunity efforts
570
J. Zhao et al. / Journal of Criminal Justice 34 (2006) 569–578
(Salzstein, 1986; Warner, Steel, & Lovrich, 1989), and
crime rates (Stucky, 2003). James Q. Wilson applied
that same line of reasoning to policing practices.
After almost forty years, it is fair to say that Wilson’s
(1968) book, Varieties of Police Behavior (hereafter
referred to as Varieties), continues to be one of the most
influential scholarly works in American policing. Not
surprisingly, a cursory review of the police literature
showed that almost every textbook on American police
featured a description of Wilson’s three styles of
policing (e.g., Bartollas & Hahn, 1999; Carter &
Radelet, 1999; Fyfe, Greene, Walsh, Wilson, &
McLaren, 1997; Langworthy & Travis, 2003; Peak,
2003; Roberg, Crank, & Kuykendall, 2000; Walker &
Katz, 2003). The social science citation index listed 620
citations of Wilson’s book in the period 1980 to 2004.
Citation and page analyses indicated that Wilson’s
scholarship was among the most prominent and widely
cited in the police and criminal justice literature and his
book, Varieties, had been ranked among the most often
cited contributions to the literature in the field of criminal justice (Cohn, Farrington, & Wright, 1998; Wright
& Miller, 1998).
In the 1970s and 1980s, several highly regarded scholars
noted the importance of the book to serious research on the
police. For example, in the late 1970s, Dorothy Guyot
(1977) argued that Varieties was the lone exception to the
general lack of credible empirical research on police
organizations. Langworthy (1986, p. 32) suggested that
Wilson’s theory of police behavior “remains the only
empirically derived theory of police organization to date.”
Similarly, Slovak (1986, p. 5) lamented that “there is a very
real sense in which the promise offered by Wilson’s
original analysis has gone unfulfilled.” In their recent
review of research on policing, Maguire and Uchida (2000)
found that the influence of Variety had remained very
strong over the decades since original publication.
The purpose of this article is to reexamine Wilson’s
theory about the relationship between local political
culture (variation in local government structure) and
priorities of policing (law enforcement, order maintenance, and provision of services). Different from all
previous studies, this research was based upon panel
survey data collected among law enforcement executives
heading 281 U.S. municipal police departments surveyed
in first 1993, in again 1996, and yet again in 2000. In
addition, a tandem survey of municipal clerks was sent to
each police agency’s city government in case of those
three years. In the analysis to follow, the authors are going
to use the same variables specified in Wilson’s book,
Varieties, to measure local political culture throughout the
empirical analysis reported here.
Literature review
The decentralization resulting from federalism in the
American political system creates a multilevel of local
government agencies in charge of a wide array of public
services provided to local communities. For example, in the
area of police services, agency practices differ widely
across communities; some agencies focus heavily on
traditional law enforcement goals exclusively, while others
are engaged in efforts to address social disorder on a wide
array of socioeconomic fronts. Over the three past decades,
several perspectives have emerged as to why variations in
public service delivery system have taken the different
patterns present in U.S. cities. Elazar (1975), for example,
suggested that the answer could be found in the history of
settlement of local communities. He identified three
distinctive strains of American political culture, namely:
(1) individualistic, (2) moralistic, and (3) traditionalistic.
Each strain of political culture has its own perspective
regarding the role of government, the function of
bureaucracy, and patterns of belief among local residents
about the proper goals of political activity. Localities differ
in the mix of political culture strains present and those
different mixes are associated with different types of local
politics and local government services (see Elazar, 1998).
Other political scientists had demonstrated that the
elements of local government structure such as the gender
and ethnicity of the chief administrator, the composition
of the municipal council, and the method of election of
public officials were predictors of variation in services
among municipalities (e.g., Williams & Adrian, 1963).
Lineberry and Fowler (1967) for example, identified the
local government structure from the least reformed
(mayor-council, partisan election with all council members elected from districts) to most reformed government
(council-manager, nonpartisan with all council members
elected at-large). Further, they found that the allocation of
financial resources differed significantly between the least
reformed versus the most reformed government. Recently, Stucky (2003) reported that crime patterns varied
systematically by local government structures (also see
Salzstein, 1986; Warner et al., 1989).
Three styles of policing
The historical roots of Wilson’s Varieties can be traced
back to the scholarship being produced in political science
in the early 1960s. At that time the documentation of the
relationship between local governmental institutional
arrangements, prevailing political values, and the resulting practices of government was a focus of scholarly
inquiry in the discipline (see Dye, 2002, pp. 11–29; Orren
J. Zhao et al. / Journal of Criminal Justice 34 (2006) 569–578
& Skrowronek, 1995, pp. 296–317; Seidelman &
Harpham, 1985, pp. 7–18). Carried out with the help of
Harvard graduate students in government, Wilson’s
(1968) seminal work was based on observations collected
from field trips to eight police departments located in
several states. After reviewing the results of extensive
interviews with officers of all ranks, intensive field work
in the communities in question, and the collection of
departmental official records, Wilson identified three
distinctive styles of policing–namely, the legalistic, the
watchman, and the service-oriented styles.
Police departments manifesting a legalistic style of
policing reflected the organizational values of a bureaucratic approach to policing. Professionalism and impartiality were viewed as the most essential elements of
organizational operation, and strict enforcement of the law
was a top priority. In this regard, Wilson observed “the
legalistic police style encourages patrolmen to take a law
enforcement view of the situation they encounter…” (1968,
p. 188). Wilson found that officers working in a legalistic
style police department tended to be highly productive in
their law enforcement activities as viewed from the standpoint of enforcement activity. For example, these officers
had high arrest rates of juvenile offenders and tended to
issue a large number of traffic tickets to motorists. The
oraizational structure of legalistic departments was typically strictly hierarchical and decidedly specialized. Here,
decision making was overwhelmingly top–down, with
relatively little input solicited from rank-and-file offiers.
Wilson referred to the legalistic style of policing as
“institutional” because it emphasized organizational
mechaisms and functional rationality in achieving its goals.
In contrast, the watchman style of policing stressed the
order-maintenance function of law enforcement. Primary
attention in this setting was given to controlling social
disorder in the local community. Wilson (1968) suggested
that the watchman style of policing was similar to the
“communal” model of policing wherein police officers
acted as “craftsmen” in the maintenance of civic order and
were deeply involved with the community they serve. He
found that, “In some communities, the police in dealing
with situations that do not involve serious crime act as if
order maintenance rather than law enforcement were their
principal function” (Wilson, 1968, p. 140). In addition,
Wilson noted the following with respect to these agencies:
“The police are watchman-like not simply in emphasizing
order over law enforcement but also in judging the
seriousness of infractions less by what the law says about
them than by their immediate and personal consequences…” (Wilson, 1968, p. 141).
Those police departments which manifested a service
orientation were inclined to emphasize community
571
residents’ satisfaction as a major organizational goal.
Accordingly, police officers took all requests for both law
enforcement and order maintenance seriously. Crimes
such as burglaries and robberies took precedence over
minor infractions of the law, and “arrests are avoided
when possible but there will be frequent use of informal
nonarrest sanctions…” (Wilson, 1968, p. 201). A good
public relationship with the community, and responsiveness to community needs, were the major concerns of all
ranks within the service style of policing law enforcement
agency.
The predictor of policing styles: local political
culture
After identifying these three styles of policing,
Wilson (1968) argued that the major determinant of
variation in police behavior was the prevailing local
political culture:
Thus, police work is carried out under the influence
of a political culture…By political culture is meant
those widely shared expectations as to how issues
will be raised, governmental objectives determined,
and power for their attainment assembled; it is an
understanding of what makes a government legitimate. (pp. 228–235)
Empirically, Wilson operationalized the concept of
political culture as represented by the type of local
government in existence where his subject police agencies
were located. Four types of local government were
identified: (1) high-professional council-manager, (2)
low-professional council-manager, (3) nonpartisan
mayor-council, and (4) partisan mayor-council. More
specifically, cities with high-professional council-manager regimes were termed “good government” settings while
cities with partisan mayor-council regimes were termed
“traditional” government settings. Wilson clearly linked
the “good government” form of local government and
political culture with the legalistic style of policing. Not
surprisingly, the emergence and development of the city
manager position was a parallel phenomenon to the
emergence of the bureaucratic model of policing-sharing
the assumptions of the politics and administration
dichotomy and the propriety of independent professional
control of public organizations.
Wilson stated that the watchman style of policing
was associated with the partisan mayor-council form of
government and its underlying political culture. An
important feature of this policing style was that the
police department was continually open to the influence
of local politics. Finally, a department with a service
572
J. Zhao et al. / Journal of Criminal Justice 34 (2006) 569–578
orientation was linked to both low-professional councilmanager and non-partisan mayor-council governments,
wherein a strong sense of public service duty prevails
and egalitarian values predominant. Nonpartisanship,
collective efforts, and community service work were
strong elements of this political culture, and these values
dictated a high concern for citizen satisfaction for the
local police operating in the environment of such a
political culture.
Wilson tested the hypothesized theoretical linkage
between political culture and variation of policing style by
examining the types of local government identified and
the arrest rates for specific crimes in corresponding police
departments. Two principal data sets were used in
Wilson’s analyses. The first data set concerning the
types of local government institutional arrangements was
obtained from the International City Management
Association (ICMA). The second data set involved arrest
rates used as indicators of variation in police behavior, and
these data were collected from arrest rates maintained by
the Federal Bureau of Investigation. Wilson controlled for
size of the city, median family income of the city
population, and the percent of non-White population; 146
police agencies were included in his empirical analysis.
Wilson (1968) hypothesized as follows:
…the more partisan the political system, the more
politicians represent small geographic constituencies,
and the more nonprofessional the executive head of the
government, the more likely the city will have a political culture favorable to the watchman police style.
By contrast, cities electing nonpartisan officials at large
and vesting executive authority in a highly professional
city manager will more likely have a political culture
favoring the legalistic police style. (p. 272)
The findings reported by Wilson indicated a clear
distinction between legalistic and watchman departments
as judged from arrest rate patterns (disorderly conduct
versus larceny and driving while intoxicated). The
empirical results documented in this analysis supported
Wilson’s hypothesis: “In sum, the theory that the political
culture of a community constrains law enforcement styles
survives the crude and inadequate statistical tests that
available data permit” (Wilson, 1968, p. 276). In addition,
the findings derived from his analysis of ICMA and FBI
UCR data supported his hypothesis that the service style
of policing emphasized both the law enforcement and
order maintenance functions of policing.
It was interesting to note that the historical development of American policing in the first seven decades of
the twentieth century witnessed the steady transformation
from the watchman style of policing to a professional
force that began with the push from the progressive
movement and a group of reform-minded police administrators such as August Vollmer at Berkeley, California
during the 1930s, and O. W. Wilson’s reform of the
Chicago Police Department during the 1960s (Rubinstein,
1973; Walker, 1977). Kelling and Moore (1988) succinctly characterized the history of police reform as falling
into three eras: the political, the reform, and the community eras. The political era captured the characteristics of
the watchman style of policing in which the authority of
policing was derived from the local community and
politicians, while the authorization of policing in the
reform era was clearly based on the rule of law.
Since the publication of Varieties, several scholars
attempted to replicate Wilson’s findings. First, Langworthy (1985) examined 152 police agencies serving
city populations over 100,000 in an attempt to demonstrate the viability of Wilson’s observations. Using FBI
1975 arrest data in the analysis, he found only partial
support for Wilson’s theory. Langworthy wrote in this
regard: “The expectation that arrest rates for larceny and
driving while intoxicated would be higher in ‘good’
government cities than in traditional cities is supported”
(Langworthy, 1985, p. 97). The socioeconomic characteristics of these 152 cities, however, were not controlled for in his analysis; hence these findings represent
only a limited test of Wilson’s ideas.
Using a sample of small and medium-sized municipal
police departments in Illinois, Crank (1990) tested
competing factors (e.g., racial-cultural differentiation,
economic conditions, organizational environment, and
local political culture) using the number of police arrests
reported by the State Police of Illinois for offenses such
as trespass, disorderly conduct, cannabis control, and
motor vehicle offenses. Crank’s findings revealed that
the presence of a city-manager style of government was
associated with significant increases in arrests for disorderly conduct, motor vehicle offenses, and cannabis
control. Crank’s test of police behavior, however, did not
employ the original measures of local political culture
developed by Wilson. For example, local partisan
election was not used to distinguish between different
styles of local government structure (e.g., commissioncouncil style of government). Similarly, Pursley (1976)
found that the number of police departments with
“reform” police chiefs was significantly higher in the
council-manager type of governments than in mayorcouncil governments.
Recently, using data derived from a 1998 survey of
large police agencies, researchers Hassell, Zhao, and
Maguire (2003) also examined the relationship between
styles of policing and law enforcement organizational
J. Zhao et al. / Journal of Criminal Justice 34 (2006) 569–578
arrangements. According to Wilson, formalization (distribution of authority, senior management versus first line
supervisor and number of rules), vertical differentiation
(number of ranks), and functional differentiation (number
of special units) should differ among the three categories
of police agencies. For example, legalistic police departments are more likely to be formalized, centralized, and
functionally limited than watchman police agencies. In
this test of Wilson’s predictions, the empirical findings
failed to support Wilson’s argument that structural
arrangements among these three styles of policing vary
predictably (Hassell et al., 2003).
The several studies summarized above found varying
degrees of support for the salience of Wilson’s
categories and the utility of his theory linking political
culture to styles of policing. Given the influence of
Wilson’s Varieties and the continuing widespread reporting of these linkages, it was indeed important to
retest his theory using more contemporary data.
Fortunately, an appropriate set of panel data captured
at three points in time during the 1990s was available for
such an analysis.
Methodology
Data sources
The information concerning local political culture
used in this analysis was obtained through three waves
of surveys of municipal agencies. The Division of Governmental Studies and Services (DGSS) at Washington
State University conducted a series of agency surveys in
1993, in 1996, and again in 2000. The staff of DGSS
conducted mail surveys directed to the same set of 281
municipalities located in forty-seven states in roughly
three-year intervals since 1978. The cities in the original
sample were selected from among those municipalities
initially surveyed in a nationwide survey of cities of
over 25,000 population conducted by the International
City Management Association in 1969. In the 2000
round of the periodic survey, 233 cities (82.9 percent) in
the sample participated. In year 1996, 244 (86.8 percent)
cities completed and returned survey questionnaires. In
1993, the survey responses of 241 (85.8 percent) cities
were available for analysis. The consistently high return
rates yielded a core sample of 185 (65.85 percent) cities
which returned completed survey instruments in all
three national surveys.
Uniform Crime Report (UCR) data published by the
Federal Bureau of Investigation provides the information
required on police arrests. UCR data are collected as part
of a nationwide effort to gather data on crimes reported to
573
police from approximately 17,000 state, county, and city
law enforcement agencies voluntarily reporting reported
crimes and police arrests in each jurisdiction. The data
used here included the 1993, 1996, and 2000 UCR data
on police arrests in the cities included in the current
study. Only police agencies that reported complete
twelve-month crime statistics in a given year were
selected for the analysis. Wilson used the same categories of UCR arrest data some forty years ago. Lastly, data
used for the four demographic characteristics in each city
(i.e., city population, percentage of minority population,
percentage of workforce unemployed, and percentage of
young people from fifteen years to twenty-four years
old) were derived from the 1990 and 2000 U.S. Census
counts.
Measures
Dependent variables
Four dependent variables were used in this study: (1)
larceny arrest rate per 1,000 population; (2) drunkenness
arrest rate per 1,000 population; (3) DWI arrest rate per
1,000 population; and (4) disorderly conduct arrest rate
per 1,000 population. In fixed-effect ordinary least
square regression models using panel data, each of the
four dependent variables was regressed on a set of
independent variables (see Appendix A for model
specifications).
Independent variables
In Varieties, Wilson identified three distinctive types
of local political culture reflected in their structural
arrangements for governance. More specifically, type of
local mayoral election, relationship between city chief
administrator and city council, previous experience of
the administrator, and educational attainment of the
administrator were all important characteristics used to
determine the three types of prevailing local political
culture. An earlier study on the influence of local
political culture on police innovations showed that over
96 percent of city chief administrators had college
degrees (using a sample of 281 cities with populations
ranging from 25,000 to 3,000,000) (Zhao, 1996). This
finding suggested that unlike the situation forty years
ago, virtually all city administrators today possess a
college degree. As a consequence of this fact, this
particular item was excluded from the measure of local
political culture.
Following Wilson’s coding method, the three types
of local political culture were operationalized as: (1)
574
J. Zhao et al. / Journal of Criminal Justice 34 (2006) 569–578
professional government: council-manager form of
government, nonpartisan election, and the manager
has previous experience in city management (which
corresponds to legalistic style police agencies); (2) traditional government: mayor-council form of government and the mayoral election is partisan (which
corresponds to watchman style police agencies); and
(3) mixed type of government: mayor-council form of
government where mayoral election is nonpartisan or
council-manager government where the election is
partisan (corresponding to service style police agencies).
In this study, traditional government was treated as the
reference group.
Control variables
In his study, Wilson controlled for the socioeconomic
conditions in the cities under investigation. He included
three aggregate demographic control measures: size,
economic conditions, and racial composition. Wilson
(1968, p. 272) argued that “controlling for such factors
should eliminate any large differences in the true, but
necessarily unknown, crime rate because crime varies
significantly with city size, social class, and racial
composition.”
Similar variables were used to control for the effects
of socioeconomic factors. Four control variables were
included in the analysis: (1) city (population) size; (2)
percentage of minorities in a city (percentage of nonWhite); (3) percentage of unemployed; and (4) percentage of persons aged fifteen to twenty-four. Demographic
data for all of these variables were obtained from the
1990 and 2000 decennial census counts of the U.S.
population. Since the census data were available for
each decade, the authors created a measure estimating
the change in socioeconomic characteristics taking place
over the period 1990 to 2000. The time-varying socioeconomic variables were estimated via a linear trend
calculation (also see Kovandzic, Sloan, & Vieraitis,
2002).
In addition to these control variables, a total of fortysix state dummy (the first listed state in the data set,
Alabama, was used as the reference group) and two
years of observation dummy variables (i.e., 1996 and
2000, 1993 was the reference group) were included in
the statistical analyses. Consideration of state and
regional variations was important to incorporate in the
analysis because numerous studies had demonstrated
important geographic effects on the structure and
operations of municipal police departments (Maguire,
Kuhns, Uchida, & Cox, 1997; Warner et al., 1989; Zhao
& Lovrich, 1998).
Findings
This section reports the primary findings of the
replication of Wilson’s theory of local political culture
determinants of policing styles in U.S. cities. Descriptive statistics for all variables included in the analysis are
shown in Table 1. Comparing across all four dependent
variables in each of the three time periods, the arrest
rates (per 1,000 population) were consistently ranked
from high to low in the following order: larceny,
drunkenness, DWI, and disorderly conduct. Although
ANOVA tests revealed some overall differences in the
four categories of arrest rates across three waves of data,
more in-depth comparisons based on post hoc Bonferroni tests suggested more specific patterns. Statistically
significant between-years (i.e., year 2000 compared to
year 1996; and year 2000 compared to year 1993) differences were found in the larceny, DWI, and disorderly
conduct arrest rates. The only statistically significant
difference between drunkenness arrest rates was found
in the year 2000 and year 1993 comparison.
The majority of the municipalities included in the
sample had either a professional council-manager type of
government (varying from 39 percent to 44 percent in
three waves of data), or a mixed type of government type
Table 1
Descriptive statistics
Variables
Dependent variables
(per 1,000 pop.)
Larceny rate⁎ ,a,c
Drunkenness rate⁎ ,c
DWI rate⁎ ,a,c
Disorderly rate⁎ ,a,c
Independent variables
Professional
government (1 = yes)
Traditional
government (1 = yes)
Mixed type
government (1 = yes)
2000
1996
1993
Mean (S.D.)
Mean (S.D.)
Mean (S.D.)
5.64 (3.32)
5.03 (4.32)
4.21 (3.39)
2.66 (3.05)
7.71 (4.32)
3.67 (4.91)
2.53 (3.80)
0.23 (0.96)
7.91 (4.11)
2.70 (10.77)
1.63 (4.46)
0.64 (2.31)
0.41 (0.49)
0.39 (0.49)
0.44 (0.50)
0.13 (0.32)
0.14 (0.32)
0.14 (0.34)
0.46 (0.50)
0.47 (0.50)
0.42 (0.49)
Control variables
Population (x 100,000) 1.88 (3.39)
1.90 (3.00)
1.81 (3.17)
Percent minority⁎ ,b,c
38.71 (22.90) 34.88 (21.88) 29.90 (20.95)
Percent unemployed
6.68 (2.88)
6.80 (2.74)
6.81 (3.03)
Percent
15.58 (5.34) 15.84 (4.83) 16.14 (5.11)
15–24 years old
⁎ Anova test, p b .05.
a
Post hoc Bonferroni test, 2000 and 1996, p b .05.
b
Post hoc Bonferroni test, 1996 and 1993, p b .05.
c
Post hoc Bonferroni test, 2000 and 1993, p b .05.
J. Zhao et al. / Journal of Criminal Justice 34 (2006) 569–578
575
Table 2
Fixed-effect panel data analyses predicting legalistic and watchman policing outcomes: using least square regression with robust standard errors
(robust standard errors in parentheses)
Law enforcement oriented
Professional government (1 = yes)
Mixed type government (1 = yes)
Population (x 100,000)
Percent minority
Percent unemployed
Percent 15–24 years old
Year 1996
Year 2000
(State dummies omitted)
Adjusted R2
n
Order maintenance oriented
Larceny
DWI
Drunkenness
Disorder
b
b
b
b
− 0.015 (0.491)
− 0.364 (0.444)
− 0.000 (0.000)
− 0.003 (0.011)
0.127 (0.068)
0.009 (0.027)
− 0.237 (0.345)
− 2.196⁎ (0.340)
0.506 (0.513)
− 0.007 (0.443)
− 0.000 (0.000)
− 0.033⁎ (0.012)
0.0210 (0.086)
0.129⁎ (0.043)
1.097⁎ (0.510)
3.430⁎ (0.549)
0.949 (1.110)
1.513 (1.090)
−0.000 (0.000)
−0.067 (0.039)
1.311 (0.857)
0.037 (0.109)
0.357 (1.015)
4.082⁎ (0.849)
− 0.769 (0.995)
− 0.741 (1.022)
0.000 (0.000)
− 0.012 (0.011)
0.172⁎ (0.064)
0.086⁎ (0.033)
− 0.172 (0.224)
2.304⁎ (0.279)
.524
535
.298
474
.335
357
.379
371
⁎ p b .05.
(varying from 42 percent to 47 percent in three waves of
data). The average city population in the sample varied
from 181,000 (in 1993) to 190,000 (in 1996) and to
188,000 (in 2000). The mean percentage minority population in the sample varied from about 30 percent (in
1993) to about 39 percent (in 2000), a statistically
significant increase. The mean percentage of workforce
unemployment remained relatively stable, at slightly less
than 7 percent in the time periods ranging from 1993 to
2000. The average population of young people (age
fifteen to twenty-four) also remained rather stable in the
same time period, averaging about 16 percent.
Table 2 reports the impact of political culture on two
major categories of police arrests: law enforcement
oriented (larceny and DWI) and order maintenancerelated (drunkenness and disorderly conduct). According to Wilson (1968, p. 274), legalistic police agencies,
which were located in communities with professional
political cultures, should have higher arrest rates for the
first two offense types. Watchman-style police departments, which were located in communities with traditional political cultures, should have higher arrest rates
on the latter two offense types. The results of these
analyses suggested that political culture, as measured by
the forms of municipal government, had no statistically
significant impact on either law enforcement oriented or
order maintenance oriented police arrests.
In contrast, significant effects were found in selected
social demographic variables. For example, the percentage of young people living in a given city was positively
related to DWI and disorderly conduct-related police
arrests in that city. Unemployment rate was positively
associated with disorderly conduct-related arrests, but
not with the other types of arrests. In addition, the per-
centage of minority population was negatively associated with DWI-related arrests, but that variable was
unrelated to the other types of arrests under investigation.
Among the two year dummy variables included in the
analyses, year 2000 was a statistically significant predictor for all four types of police arrest rates. Controlling
for all the other variables, there were more DWI–,
drunkenness–, and disorder conduct–related arrests in
year 2000 compared to those of 1993. Larceny arrest
rates for the year 2000, however, were significantly
lower that those of 1993. Overall, with the addition of
state dummy variables, the regression models explained
between 30 percent and 52 percent of the variation in the
dependent variables.
Discussion and conclusion
The purpose of this study was to reassess the utility of
Wilson’s theory of the linkage between local political
culture and styles of policing as evidenced in aggregate
arrest rates for various reported crimes. Longitudinal
agency survey data collected in 1993, 1996, and 2000
from 185 U.S. cities were utilized in the analysis. In
addition, the authors controlled for both time and place
(state) in the statistical model, and the authors used
essentially the same variables as those used by Wilson
nearly forty years ago. Overall, the findings did not find
support for Wilson’s theory of local political culture in
explaining today’s police organizational behavior. There
was no evidence of the effect of local government culture
as specified by James Q. Wilson on arrest rates among
these many contemporary American police agencies.
Why is political culture no longer a strong predictor
of patterns of organizational behavior among large
576
J. Zhao et al. / Journal of Criminal Justice 34 (2006) 569–578
police agencies? The authors would like to offer three
speculations that center on the changes in the character
of police work taking place over the course of the past
forty years. The first speculation concerns the dramatic
increase in crime rates occurring shortly after the completion of Wilson’s study in the mid-1960s. The
historical trend of crime rates in the U.S. was that
crime incidents per 100,000 population began to climb
rapidly in the late 1960s. Based on the Uniform Crime
Report of the FBI, the rate of increase was even faster in
the 1970s and 1980s than it had been in the late 1960s.
Fear of crime, particularly violent street crime captured
the attention of ordinary Americans (e.g., Wilson, 1985).
Police agencies across the nation witnessed politicians
at all levels of government using the fear of crime in
their political appeals, and “getting tough on crime”
once elected to office (Lyons, 2002; Scheingold, 1984).
Studies conducted in the 1990s had documented the fact
that police agencies, large and small alike, consistently
rated crime control as their top priority followed by
social disorder regardless of their local crime rate and
regardless of social demographic factors (Zhao, He, &
Lovrich, 2003). Similarly, scholars of American policing had long argued that law enforcement was the core
mission of contemporary American police agencies
(Bittner, 1970; Cao, 2004; Manning, 1988).
The second speculation concerned change in the
organizational arrangements in place within contemporary police agencies. In the mid-1960s, specialization
(number of special units) and formalization (rules and
policies) were rather limited among police organizations. Consequently, Wilson found that police departments characterized as watchman style did not have
much formalization and specialization compared to their
counterparts of legalistic style. Since the period in which
Wilson conducted his studies, police departments had
changed significantly in this regard, as a result of both
external and internal pressures. Formalization, for example, is a major feature of organizational life in
contemporary law enforcement, virtually all agencies
maintained detailed policy and procedure documents. In
fact, research studies consistently point to formalization
as a major trait of contemporary American law enforcement agencies (Langworthy, 1986; Zhao, 1996).
In the Law Enforcement Management and Administrative Statistics (1999) survey, virtually all of the large
police agencies reported having internal policies on use
of force, code of conduct and appearance, the handling of
domestic disputes, hot pursuits, etc. In terms of external
pressures, the data used for this article indicated that
around 20 percent of municipal police departments
surveyed in 1993, 1996, and 2000 were under court order
to promote equal employment opportunity by hiring
additional minority and female officers. Organizational
change in this area is not a luxury, but rather more of a
necessity for contemporary police agencies. The authors
can use the Omaha Police Department as a case in point
to illustrate how much has changed in this area over the
years. The Omaha Police Department has been transformed from a watchman style of policing operation to a
professional style of policing organization without a
corresponding change in local government structure.
According to the Wilson classification of policing
styles, the Omaha Police Department should adhere to
the watchman style (mayor-council type of government
with partisan election). In fact, however, the Omaha P.D.
is highly differentiated in operations by having a gang
unit, organized crime unit, traffic unit, pawn unit, and
even a nuisance unit. At the same time, the department is
highly formalized featuring extensive policies on use of
force, racial profiling, hot pursuit, etc. In addition,
Wilson observed in the watchman style police departments he studied that educational requirements were
minimal: “In none of the three cities has a high school
diploma been required for entrance into the force and
many officers could not meet such a requirement were it
imposed” (Wilson, 1968, p. 152). Today, over 50 percent of police officers in the Omaha Police Department
possess a college degree. Finally, Wilson found that
police officers working in watchman style departments
tended to be poorly paid: “Patrolmen are locally recruited, paid low salaries, expected to have second jobs”
(Wilson, 1968, p. 151). The entry-level salary for an
Omaha police officer is the highest in the metropolitan
area (featuring over fifteen agencies).
The third speculation is related to the expanding
influence of the federal government in law enforcement
policies and practices. There has been significant federal
government involvement with respect to how police
work should be conducted. Prior to Wilson’s study, the
influence of the federal government on local law
enforcement agencies was rather minimal. Since the
1970s, however, the federal government has played a
much more important role in influencing the organizational behavior of police agencies across the entire
country. Walker (1999) observed that in the late 1960s,
American police agencies emerged in the national
spotlight. Presidential commissions and the Law Enforcement Assistance Administration (LEAA) collectively played a major part in professionalizing law
enforcement agencies. In 1973, the American Bar
Association (ABA) published its Standards Relating to
the Urban Police Function. This was a reflection of a
growing body of research on policing and the emergence
J. Zhao et al. / Journal of Criminal Justice 34 (2006) 569–578
of a deeper level of understanding of the role that police
departments should play in a democratic society on the
part of major stakeholder groups in American society
(Walker, 1999, p. 35).
In the most recent concrete example of such an
understanding, the Crime Control Act of 1994 emphasized the systematic redeployment of police resources into
community policing activities. Over 11,000 law enforcement agencies had received federal funding (around $7
billion) distributed through the Office of Community
Oriented Policing Services. Along with those funds, these
agencies have had to adapt themselves to federal
requirements for building police-community partnerships. All of these federal and associated state efforts to
promote community policing have contributed in a major
way to the standardization of police work, which in turn
reduces the influence of local political culture on police
policies and practices.
This article calls into question one of the most classic
pieces of scholarship in the policing. Wilson’s important
demonstration of the creative application of empirical
methods to the development of theory in policing is
routinely cited in textbooks and in scholarly writings
about the police. This article makes the argument that
new research is needed to examine the effect of local
political culture on contemporary policing in the U.S.,
and notes that a theory which once could be relied upon
to explain variation of police practices across local
government jurisdictions can no longer serve that
function. Such is likely the case in a good deal of social
sciences, and social scientists would be wise to revisit
their favorite theories in the way that was done here with
Wilson’s classic work. This analysis was not intended to
demean the foundational work of James Q. Wilson, but
rather to deepen the authors’ understanding of the
limitations of that work and demonstrate the wisdom of
the periodic replication of classic social science
scholarship (see Cao, 2003).
Appendix A. Statistical model
The authors used pooled cross-sectional time-series
(or panel data) analysis in the study reported here. Based
on a set of well-established criteria from the field of
econometrics (cf. Baltagi, 2001; Judge, Hill, Griffiths,
Lutkepohl, & Lee, 1988), the authors specified fixedeffects model estimation as the most appropriate technique for the study. The Breusch and Pagan Lagrangian
multiplier test for random effects and the Hausman
specification test were examined systematically. The
Hausman test results indicated that coefficients associated with fixed-effect and random-effects models were
577
significantly different from one another. The Breusch
and Pagan test results weighed in favor of the fixed-effect
model.
The general model can be written as:
Yit ¼ bVXit þ ai þ gt þ eit
Eai ¼ Eeit ¼ 0
Eai2 ¼ ra2; Eaiaj ¼ 0 f or i p j
Eeitejs ¼ re2; if i p j and t p s ðEeitejs ¼ 0; otherwiseÞ
where Yit was the dependent variable for city i at year t.
The symbol Xit represented a set of explanatory
variables, and εit represented the error term. The statespecific effect was measured by αi (forty-six state
dummies were used out of a total of forty-seven states
included in the sample, Alabama was the reference
category). The time-specific component γt represented
two year dummy variables (1996 and 2000, 1993 was
the reference category) that controlled for the unknown
factors affecting the variation in policing styles that were
not accounted for by the other variables. It was also
assumed that αi,γt, and εit were uncorrelated with Xit
(Baltagi, 2001; Hsiao, 1995, 2003; Judge et al., 1988).
References
Baltagi, B. H. (2001). Economic analysis of panel data (2nd ed.). New
York: John Wiley and Sons.
Bartollas, C., & Hahn, L. (1999). Police in America. Boston: Allyn and
Bacon.
Bittner, E. (1970). The functions of police in modern society.
Washington, DC: National Institute of Mental Health.
Cao, L. (2003). Reflection on integration and elaboration in theory
construction. Crime and Criminal Justice International, 1, 125−147.
Cao, L. (2004). Three cases that have changed American policing.
Political Science and Law, 22, 50−56.
Carter, D., & Radelet, L. (1999). The police and the community. Upper
Saddle River, NJ: Prentice-Hall.
Cohn, E. G., Farrington, D. P., & Wright, A. R. (1998). Evaluating
criminal justice and criminology. Westport, CT: Greenwood Press.
Conlan, T. (1998). From new federalism to devolution: Twenty-five
years of intergovernmental reform. Washington, DC: Brookings
Institute Press.
Crank, J. P. (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.
Dye, T. R. (2002). Understanding public policy (10th ed.). Upper
Saddle River, NJ: Prentice Hall.
Elazar, D. J. (1970). Cities of the prairie. New York: Basic Books.
Elazar, D. J. (1975). The American cultural matrix. In D. J. Elazar &
J. Zikmund (Eds.), The ecology of American political culture:
Readings (pp. 13−42). New York: Thomas Y. Crowell Company.
Elazar, D. J. (1998). Forward. In D. Schleicher & B. Swodlow (Eds.),
Federalism and political culture (pp. ix−xv). New Brunswick, NJ:
Transaction.
Elazar, D. J. (2002). The closing of the metropolitan frontier: Cities of
the prairie revisited. New Brunswick, NJ: Transaction.
578
J. Zhao et al. / Journal of Criminal Justice 34 (2006) 569–578
Fyfe, J., Greene, J., Walsh, W., Wilson, O. W., & McLaren, R. (1997).
Police administration (5th ed.). New York: McGraw-Hill.
Gabriel, R. H. (1956). The course of American democratic thought
(2nd ed.). New York: Ronal Press.
Guyot, D. (1977). Police departments under social science scrutiny.
Journal of Criminal Justice, 5, 105−118.
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 and Management, 26, 231−250.
Hsiao, C. (1995). Panel analysis for metric data. In G. Arminger, C.
Clogg, & M. Sobel (Eds.), Handbook of statistical modeling for
the social and behavioral sciences (pp. 361−400). New York:
Plenum Press.
Hsiao, C. (2003). Analysis of panel data (2nd ed.). Cambridge, UK:
University of Cambridge Press.
Judge, G., Hill, R. C., Griffiths, W. E., Lutkepohl, H., & Lee, T. C.
(1988). Introduction to the theory and practice of econometrics
(2nd ed.). New York: John Wiley and Sons.
Kelling, G., & Moore, M. (1988). From political to reform to community:
The evolving strategy of police. In J. R. Greene & S. Mastrofski
(Eds.), Community policing: Rhetoric or reality (pp. 1−26). New
York: Praeger.
Kovandzic, T., Sloan, J., & Vieraitis, L. (2002). Unintended consequences
of politically popular sentencing policy: The homicide promoting
effects of ‘three strikes’ in U.S. cities (1980–1990). Criminology and
Public Policy, 1, 399−424.
Langworthy, R. (1985). Research note: Wilson’s theory of police behavior:
A replication of the constraint theory. Justice Quarterly, 2, 89−98.
Langworthy, R. (1986). The structure of police organizations. New
York: Praeger.
Langworthy, R., & Travis, L. (2003). Policing in America: A balance
of forces (3rd ed.). Upper Saddle River, NJ: Prentice Hall.
Law Enforcement Management and Administrative Statistics. (1999).
Data for individual state and local agencies with 100 or more
officers. Washington, DC: U.S. Department of Justice.
Lineberry, R., & Fowler, E. (1967). Reformism and public policies in
American cities. American Political Science Review, 61, 701−716.
Lorch, R. S. (2001). State and local politics: The great entanglement
(6th ed.). Upper Saddle River, NJ: Prentice Hall.
Lyons, W. (2002). The politics of community policing: Rearranging
the power to punish. Ann Arbor: University of Michigan Press.
Maguire, E. R., Kuhns, J. B., Uchida, C. D., & Cox, S. (1997). Patterns
of community policing in nonurban America. Journal of Research
in Crime and Delinquency, 34, 368−394.
Maguire, E. R., & Uchida, C. D. (2000). Measurement and explanation in the comparative study of American police organizations. In
D. Duffee (Ed.), Measurement and analysis of crime and justice
(pp. 491−557). Washington, DC: U.S. Department of Justice.
Manning, P. (1988). Community policing as a drama of control. In J. R.
Greene & S. Mastrofski (Eds.), Community policing: Rhetoric or
reality (pp. 27−46). New York: Praeger.
Orren, K., & Skrowronek, S. (1995). Order and time in institutional
study: A brief for the historical approach. In J. Farr, J. S. Dryzek, &
S. T. Leonard (Eds.), Political science in history research
programs and political traditions (pp. 296−317). New York:
Cambridge University Press.
Peak, K. J. (2003). Policing America: Methods, issues, challenges
(4th ed.). Upper Saddle River, NJ: Prentice Hall.
Pursley, R. (1976). Community characteristics and policy implementations: Some exploratory findings about two categories of
municipal police chiefs. Journal of Criminal Justice, 4, 291−301.
Roberg, R., Crank, J. P., & Kuykendall, J. (2000). Police and society
(2nd ed.). Los Angeles: Roxbury.
Rubinstein, J. (1973). City police. New York: Farrar, Straus, and Ginoux.
Salzstein, G. H. (1986). Female mayors and women in municipal jobs.
American Journal of Political Science, 30, 140−164.
Scheingold, S. A. (1984). The politics of law and order. New York:
Longman.
Seidelman, R., & Harpham, E. J. (1985). Disenchanted realists political
science and the American crisis, 1884–1984. Albany: State University
of New York Press.
Slovak, J. S. (1986). Styles of urban policing: Organization, environment,
and police styles in selected American cities. New York: New York
University Press.
Stucky, T. (2003). Local politics and violent crime in United States
cities. Criminology, 41, 1101−1135.
Walker, S. (1977). A critical history of police reform. Lexington, MA:
D.C. Heath.
Walker, S. (1999). Police in America: An introduction (4th ed.).
Boston: McGraw-Hill.
Walker, S., & Katz, C. M. (2003). Police in America: An introduction
(5th ed.). Boston: McGraw-Hill.
Warner, R., Steel, B., & Lovrich, N. (1989). Conditions associated
with the advent of representative bureaucracy: The case of women
in policing. Social Science Quarterly, 70, 562−578.
Williams, O., & Adrian, C. (1963). Four cities. Philadelphia: University
of Philadelphia Press.
Wilson, J. Q. (1968). Varieties of police behavior. New York: Harvard
University Press.
Wilson, J. Q. (1985). Thinking about crime (2nd ed.). New York:
Vintage Books.
Wright, R. A., & Miller, J. M. (1998). The most cited scholars and
works in police studies. Policing: An International Journal of
Police Strategies and Management, 21, 240−254.
Zhao, J. (1996). Why police organizations change: A study of community
policing. Washington, DC: Police Executive Research Forum.
Zhao, J., He, N., & Lovrich, N. (2003). Community policing: Is it
changing the basic functions of policing in the 1990s? A national
follow-up study. Justice Quarterly, 20, 697−724.
Zhao, J., & Lovrich, N. (1998). Determinants of minority employment
in American municipal police agencies: The representation of
African American officers. Journal of Criminal Justice, 26,
267−277.