UNIVERSITY OF CINCINNATI

Transcription

1 U UNIVERSITY OF CINCINNATI Date: I,, hereby submit this original work as part of the requirements for the degree of: in It is entitled: Student Signature: This work and its defense approved by: Committee Chair: Approval of the electronic document: I have reviewed the Thesis/Dissertation in its final electronic format and certify that it is an accurate copy of the document reviewed and approved by the committee. Committee Chair signature:

2 Policing in the Suburbs: Assessing Wilson s Theory of Local Political Culture A dissertation submitted to the Division of Research and Advanced Studies of the University of Cincinnati In partial fulfillment of the requirements for the degree of Doctor of Philosophy (Ph.D.) in the School of Criminal Justice of the College of Education, Criminal Justice, and Human Services July 2009 by Theresa Ervin Conover M.P.A. University of Cincinnati, 1997 B.U.P., University of Cincinnati, 1991 Dissertation Committee: Dr. James Frank (chair) Dr. Lawrence Travis Dr. Edward Latessa Dr. Ken Novak

3 Abstract Since its original publication in 1968, James Wilson s Varieties of Police Behavior and the introduction of political culture has dominated the policing literature as an explanation of the relationship between community characteristics and police behavior. Three archetypal styles were identified: legalistic, watchman, and service. Wilson posited that a community s policing style can be predicted by their form of local government. While support was found in his initial test, subsequent tests have been unable to garner support. This dissertation starts by using Wilson variables with data from a systematic social observation (SSO) study and measures policing style by frequency and formality of contacts rather than arrests. The study was conducted in 20 suburban communities in southwest Ohio. Alternative variables were introduced and their hypothesized relationship with frequency and formality discussed. The data on encounters were broken out into three subgroups: number of all, full and full encounters with probable cause and each were analyzed accordingly. Associations were found for crime indicator, non-residential land use and assessed value per capita in the alternative models using chi-square. There was no overwhelming support for either the Wilson or alternative variables to predict style. Next, variables representing both Wilson and alternatives were used to predict ratio level measures of frequency and formality. The Formality/Frequency Index represents the amount of formality exhibited in each respective community. In the end, political culture was not found to support the frequency of contacts (# of encounters) or formality (# arrests and/or citations). Crime indicator was the only variable to be a significant predictor of the Formality/Frequency Index (full encounter model) in any of the alternative models presented. This relationship is tenuous at best due to the small sample size. ii

4 It is important to note that this dissertation was not centered on testing Wilson s theory of political culture -- the focus was on explaining the relationship between community characteristics and police behavior. However, it did use the theory as a benchmark, as a point of reference in the quest to identify these characteristics. This dissertation also discussed the state of the research which has addressed Wilson. It also has given plausible explanation as to why political culture may have been an appropriate measure in its day, but may no longer have explanatory power. Alternative community-level variables were presented that were hypothesized to explain variation in police behavior vis-à-vis frequency and formality of policecitizen encounters in suburbs. Some support for these alternative hypotheses was found, though the effect was not consistent. Ultimately these hypothesized relationships did not hold utilizing higher level statistical analyses. There is work to be done in addressing the variation found in police-citizen contacts between communities. The challenge for future research is to tie the acknowledgement that policing varies by community with characteristics that explain why. iii

5 iv

6 Table of Contents CHAPTER ONE: PROBLEM STATEMENT... 1 Introduction... 1 Suburbanization and the study of smaller communities... 2 Local Political Culture and the Introduction of Policing Styles... 3 Alternative Variables... 5 Importance of Study... 5 CHAPTER TWO: COMMUNITY VARIATION AND... 8 POLICING STYLES... 8 Wilson s Varieties of Police Behavior... 8 Police Style Limitations of Existing Theory CHAPTER THREE: REVIEW OF THE LITERATURE Tests of Local Political Culture Langworthy Zhao and Hassell Crank Liederbach and Travis Towards a New Community Police Theory Alternative Community Characteristics CHAPTER FOUR: METHODOLOGY Research Questions Description of Study Sites Data Sources Measurement of Study Variables Wilson s independent variables Estimated Daytime Population Land Use Crime Indicator Voting Behavior Tax Base /Resource Capacity The Funnel Dependent Variables: Measurement Issues with UCR Arrest Data Dependent Variable: Frequency and Formality Dependent Variable: Policing Style Dependent Variable: Eight-hour shift equivalent Dependent Variable: Formality/Frequency Index Analytical Approach Strengths and Limitations of Current Research Summary CHAPTER 5: FINDINGS Wilson s Political Culture Model Police Style The Alternative Model v

7 Frequency: Police-Citizen Encounters Formality: Arrest and/or Citation Computation of Police Agency Policing Style Research Questions Findings Bivariate Relationships Chi-square Analysis Analysis of Variance OLS Regression Summary CHAPTER 6: DISCUSSION Additions to scholarly research Reconceptualizing Wilson s Theory Why have studies repeatedly failed to find support for political culture? Convergence of reforms and other institutional responses Discussion of Findings Limitations of this study Future Research vi

8 List of Tables Table 4.1 Demographic Characteristics of Communities 44 Table 4.2 Description of Independent and Dependent Variables. 52 Table 4.3 Dependent variables and descriptions of datasets 60 Table 4.4 Approach to Analyses.. 64 Table 5.1 Descriptive Statistics. 68 Table 5.2 Additional Resources Income Tax, Police Levies, and Grants.. 71 Table 5.3 Frequency and Formality of Police-Citizen Encounters per 8-hour shift.. 73 Table 5.4 Computation of Police Agency Style: All Encounters.. 77 Table 5.5 Computation of Police Agency Style: Full Encounters. 78 Table 5.6 Computation of Police Agency Style: Full Encounters with Probable Cause Table 5.7 Policing Styles Moving Through the Funnel. 80 Table 5.8 Bivariate Relationships: Community Characteristics by Frequency and Formality Measures. 83 Table 5.9 Chi Square Analysis: All Encounters. 87 Table 5.10 Chi Square Analysis: Full Encounters.. 88 Table 5.11 Chi Square Analysis: Full Encounters with Probable Cause.. 89 Table 5.12 Analysis of Variance: All Encounters. 91 Table 5.13 Analysis of Variance: Full Encounters 91 Table 5.14 Analysis of Variance: Full Encounters with Probable Cause. 92 Table 5.15 Zero Order Relationships for Wilson Independent Variables. 94 Table 5.16 Zero Order Relationships for Alternative Variables 95 Table 5.17 OLS Regression: All Encounters 98 Table 5.18 OLS Regression: Full Encounters.. 99 Table 5.19 OLS Regression: Full Encounters with Probable Cause 100 Table 5.20 OLS Analysis of All Encounters per 8-hour shift equivalent. 103 Table 5.21 OLS Analysis of Full Encounters per 8-hour shift equivalent 104 Table 5.22 OLS Analysis of Full Encounters with probable cause per 8-hour shift equivalent. 105 Table 5.23 OLS Analysis of Arrest, Cite or Both per 8-hour shift equivalent. 106 List of Figures Figure 4.1 Police-Citizen Encounters to Arrest or Citation Funnel 57 Figure 5.1 Frequency/Formality Table.. 75 vii

9 CHAPTER ONE: PROBLEM STATEMENT Introduction A widely held assumption in the criminal justice field is that police agencies are a reflection of the communities they serve. One of the central tenets of the American experience is the exercise of local control over government functions including the police, rather than a centralized state authority. Goldstein (1977) commented on this relationship: Viewing the police primarily as an agency of municipal government is a way of emphasizing the fact that each community has the opportunity to make its own judgments as to what its police force should do. Implicit in this approach is the belief that most of the non-criminal functions police now perform are not inappropriate tasks if a community concludes that the police agency is the logical administrative unit in which to house them (1977:33). Does community influence police behavior -- are certain community characteristics associated with varying police behavior? Generally, research on this question has been couched in terms of Wilson (1968) and the examination of political culture as the mechanism of influence on police style. Wilson proposed a theory of local political culture for American policing over 40 years ago. His work has played a prominent role in the examination of the relationship between the community and police organizational styles in criminal justice research. As hypothesized, variations in police style should be observed across police departments and can be explained by a community s political culture. This work is widely cited, however, empirical tests of the theory are few and those that have been undertaken have generally not been supportive. Although there are several theories that address community characteristics and their impact on police behavior this study starts with Wilson s (1968) premise as it relates to political 1

10 culture and police style and moves on to examine other community characteristics that may influence police-citizen interactions. Research in this area has been described as providing mixed support (Crank 1990, 1992; Langworthy 1985; and Slovak 1986) and more recently results have leaned toward falsification of the theory (Zhao and Hassell 2005; and Zhao, He and Lovrich 2006) or towards acceptance as intuitive appeal (Liederbach and Travis 2008). The purpose of this study is to increase the understanding of the relationship between community and policing. This is done in three ways: (1) by examining Wilson using a different measure of style in a suburban context; (2) by providing alternative community variables to predict policing style; and (3) using alternative variables to shed light on the relationship between other community characteristics and the frequency and formality of police-citizen encounters. Suburbanization and the study of smaller communities A substantial shift in population has occurred since the time of Wilson s writing. The 1970 census indicated the watershed in the geographical distribution of U.S. residents: more people lived in suburbs than in central cities or rural areas (Baumgartner 1988). By 1980, approximately 45 percent of U.S. residents lived in suburbs, compared with 30 percent in cities and 25 percent in remaining areas (Baumgartner 1988). By 2000, half of the U.S. population called the suburbs home (U.S. Census Bureau 2002). Given that half of the U.S. population lives in suburbs, it is important to understand what community characteristics are associated with policing. Wilson s qualitative investigation of policing styles included suburban and urban communities with populations ranging from 25,532 to 1.3 million. His empirical research included only large communities with populations ranging from 25,000 to 100,000. The majority of studies which have tested Wilson have included urban areas (cf. Liederbach and 2

11 Travis 2008). Since support for political culture is an unresolved issue in the urban communities it has been used to study, and the importance of suburbs, it is only natural to extend Wilson s inquiry to suburbs. Since a potential issue in studying smaller areas is that arrest data is not readily available via the FBI Uniform Crime Report (UCR) this study utilizes recent data collected through systematic social observations (SSO). Most of the existing research relies on the initial measures utilized in Wilson s original test on national data. Wilson uses a substitute measure for political culture a modification of a form of government variable. He issues caution about the crudity of measures and the crude and inadequate statistical tests employed in the analysis of this data. The agencies in Wilson s sample were urban. If the theory is externally valid the same hypotheses should be supported in smaller, less urban communities. This study starts with a test of Wilson s theory on smaller, suburban police departments using systematic social observation and then proposes an alternative approach to understanding police behavior in terms of the frequency and formality of police-citizen interaction. Local Political Culture and the Introduction of Policing Styles In recent years, Wilson s theory has come under criticism because it has been unable to support the hypotheses originally made. In the most recent analyses (Zhao and Hassell 2005; Zhao, He and Lovrich 2006; and Liederbach and Travis 2008) demographic variables, not political culture, were found to be significant predictors of styles of policing. Wilson, without question has dominated community-level inquiry as it relates to policing. Wilson s (1968) central premise was the reliance on political culture and its influence in defining the police style in communities. In short, he was interested in the role of local politics in defining the role of police in the community. Order maintenance functions of the police, 3

12 including family quarrels, juvenile disturbances, domestic disturbances, prostitution, and gambling were the focus of Wilson s discussion. He found that police behavior differed from community to community and illustrated three archetypal police styles: service, legalistic, and watchman. Wilson (1968) associated certain community characteristics with the type of police style the community exhibited. Watchman style departments were found in areas where politicians appealed to working and lower middle class constituencies. Emphasis was given to maintaining a low tax rate with a low level of public services. Industrial growth was absent in these communities. The service style department was found in areas where politicians regarded politics as the provision of public services and amenities to persons clearly able to afford them (250). It was believed that people were willing to pay for programs and services that increased their quality of life beyond the provision of basic public services. The legalistic style was exhibited by communities employing a strong, professionally oriented city manager recruited from outside the city. In his narrative (the majority of the book) he touches on factors such as industrial base, level of public services, quality of life issues, and taxation among others in the descriptions of communities he studied. However, the analysis of national data (the results of which are oft cited) include only the variables of political culture without addressing those other community factors that he apparently determines are important in his narrative. It may be that conceptualizing and creating measures for those variables would have been too time consuming and costly. Since they are mentioned in the narrative it warrants attention to at least examine concepts of those characteristics of community as they relate to policing styles. 4

13 Studies that have empirically tested Wilson s theory of local political culture have cast doubt on its ability to predict styles of policing using the community characteristics originally employed (Swanson 1978; Langworthy 1985; Crank 1990, 1992; Zhao and Hassell 2005; and Zhao, He and Lovrich 2006). These studies have included cities with populations ranging from 25,000 to 1 million and cities with police agencies employing at least 100 sworn full-time officers. More recently data from suburban communities have been used to test Wilson s theory (Liederbach and Travis 2008). Alternative Variables In a move to more broadly understand the characteristics of communities that impact police behavior, alternative variables representing variations in a community s physical context are included in this dissertation. These variables are included to gauge the differences in land use, community wealth, road networks, voting behavior and population density which are hypothesized to structure the opportunity for police citizen encounters. These variables will be first used to predict police style and then to predict overall levels of police-citizen contact in various subgroups of frequency and formality. Importance of Study This dissertation attempts to assess the theory of local political culture first by testing Wilson utilizing data on smaller, suburban communities acquired through systematic social observation (SSO). Policing style (watchman, legalistic, service) is not operationalized using arrest data, rather a composite of frequency and formality of contacts which comport with the styles identified by Wilson. Liederbach and Travis (2008) were the first to operationalize style in this manner. Next, since SSO data provide rich detail about the frequency and formality of 5

14 police-citizen encounters those outcomes are also identified and discussed. This approach diverges from earlier efforts by acknowledging that there are different types of encounters only some of which may result in a formal law enforcement response: the analogy of a funnel is presented. Wilson s political culture theory is addressed but is not the sole focus of this dissertation. Starting with the new operationalization of policing style the dissertation moves on to examine alternative community-level variables which may be likely to provide opportunities for police-citizen contact. Finally, the alternative variables are used to examine the frequency and formality of police-citizen contacts. This study seeks to achieve the following objectives in addressing the question as to how community characteristics may influence police behavior either through style or frequency and formality of police-citizen encounters: 1. To utilize the frequency and formality of police contacts with citizens to define police agency style; 2. To test Wilson s (1968) theory of local political culture to predict policing style in suburban communities using SSO rather than arrest data; 3. To examine the relationship between community characteristics and policing style by including alternative community-level variables to account for community variation; 4. To examine the relationship between alternative community characteristics and the frequency and formality of police-citizen interaction; and 5. To increase scholarly knowledge about the relationship between community characteristics and policing in the suburbs. In the quest to develop an explanation for the relationship between community and policing styles, specifically, and the frequency/formality of contacts, generally, this study starts with Wilson s premise, past attempts to test it and proposes alternative variables to account for differences in community composition and physical structure. Chapter 1 presented the problem to be addressed by this dissertation. Chapter 2 lays out Wilson s theory detailing the hypothesized relationship between communities and policing 6

15 styles. Chapter 3 reviews the extant literature that has tested the theory of local political culture and provides alternative measures to extend Wilson s theory to the 21 st century. Chapter 4 details the methods to be carried out in this research. Chapter 5 provides the results of statistical analyses comparing the Wilson and the alternative variables. Finally, Chapter 6 discusses the findings and their importance to scholarly research. 7

16 CHAPTER TWO: COMMUNITY VARIATION AND POLICING STYLES Wilson s Varieties of Police Behavior Varieties of Police Behavior was written out of an interest in the administration of important public services and how that administration is affected by the nature of the tasks to be performed and the political context in which the organization exists (1968:x). Wilson was interested in how community decisions affected police functions. These decisions included legal and organizational constraints on individual officers. In his narrative, he is primarily focused on law enforcement and order maintenance functions; service functions were omitted. The objective was to: describe the behavior of patrolmen discharging their routine law-enforcing and order-maintaining functions, to explain how that behavior is determined by the organizational and legal constraints under which patrolmen work, to discover the extent to which it varies among police departments, and to determine insofar as the evidence permits what accounts for these differences and especially how local politics contributes to them (p. 10). In the end, eight communities were described in the exploratory study: Albany, Amsterdam, Brighton, Nassau, Newburg, and Syracuse, NY; Highland Park, IL; and Oakland, CA. Wilson defines political culture as 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 (p. 233). The prevailing political culture creates a zone of indifference within which the police are free to act as they see fit (p. 233). Wilson (1968) noted that political culture, though not cohesively defined as it is not easily measured, determines police behavior as discussed above. Ultimately, police in a community 8

17 reflect the values and desires/expectations of its citizenry and define what the police do (police role). Having been the seminal study in this area, other scholars have attempted to define political culture in varying ways. Wilson (1968) used a variable for form of government which is a substitute for political culture; it includes four (4) attributes: mayor-council/partisan ballot; mayor-council/non-partisan ballot; council-manager/low professionalism ; and council-manager/high professionalism. He had access to non-published data concerning the educational attainment of the city managers in his study. This education variable has not been replicated in any subsequent studies. Wilson (1968) laid out the following hypotheses as they pertained to political culture : Cities with professionalized good government regimes would be likely to have legalistic police forces that made more arrests than other cities in law enforcement situations (that is, for larceny and driving while intoxicated); that cities with partisan, mayor-council regimes would be more likely to have watchman-style police forces that would make more arrests in order maintenance situations where the action is typically police-initiated (that is, for disorderly conduct); that in order maintenance situations where the police response is citizen-invoked (that is, assault), there would be relatively little difference between police departments, and thus between cities, because arrests in these cases require citizen cooperation and therefore depend more on the characteristics of the citizens than on the characteristics of the police; and finally that drunkenness arrests might or might not differ, depending on whether the police saw such matters as problems of law enforcement or order maintenance, but that professionalized cities would probably show a higher arrest rate because they were more likely to take the law enforcement point of view (Wilson 1978:274). 9

18 The dependent variable, policing style, is measured as a proxy for arrest rates for highly discretionary offenses (larceny, simple assaults, drunkenness, disorderly conduct, and driving while intoxicated). He cautions that arrest rates may not be the most valid way to measure policing styles. Communities may have higher arrest rates for certain offenses, but have very different reasons for why this is so. The prime concept he was trying to tap was those offenses which provided officers the widest amount of discretion for enforcement. Police Style Wilson (1968) describes three archetypal police styles: legalistic, watchman and service. The following discussion includes a description of behaviors manifested in each departmental style. Briefly, the communities where each departmental style may take shape are somewhat dependent on the populations served. Communities exhibiting diverse populations and high residential instability are more synonymous with either legalistic or watchman departments, while homogenous communities tend to orient toward the service style. Watchman Style The watchman style identified by Wilson (1968) is similar to the type of police found in the United States at the beginning of the 19 th century. The principal function of the watchman department is order maintenance. This type of department has a predisposition to avoid formal involvement, since their main function is to control public disorder rather than to enforce the law (p. 148). Officers tend to overlook common minor violations, especially juvenile and traffic offenses. In describing the types of situations officers in a watchman department engage in, it might be more worthy to examine those instances where they tend to steer clear. According to Wilson, a certain level of vice and gambling was tolerated as it was considered a private matter, 10

19 juveniles were treated informally or ignored, motorists were left alone, and private disputes involving friends and family were treated informally, and business disagreements required a civil response. As such, officers tended to follow the path of least resistance. Public intoxication and disorderly conduct offenses were frequently used to maintain order in public places. The community impacts police response in the watchman departments largely through the actions of the individual officers. Unlike the legalistic style department where one community standard was perceived, responses here are individualized based on the group characteristics exhibited by the citizen(s) involved in the encounter. Typically, watchman departments operate in heterogeneous communities with little population stability (Langworthy and Travis 1999). Further they state: Citizen intervention is not frequent in watchman style agencies. There are fewer proactive contacts with citizens. Officers maintain substantial discretion in addressing the problem and if, and how, to respond. The peacekeeping function is emphasized. The role played by these officers is of arbiter or judge (Langworthy and Travis 1999:164). Wilson (1968) predicted that officers in watchman agencies engaged in work routines that would put them infrequent contact with the citizenry. Using the law for order maintenance purposes resulted in few formal sanctions such as arrest or citation. It is expected that there would be infrequent contacts and informality when dealing with citizens. Legalistic Style In the legalistic police department, situations are generally handled as law enforcement rather than order maintenance matters. Formal action (arrests and issuance of citations) is a key component in this departmental style. Under the law enforcement view, many traffic tickets are issued, a high proportion of juveniles are detained and arrested, a high number of misdemeanor 11

20 arrests are made, shoplifters are prosecuted, and action is taken against vice. In the watchman department the standard of conduct depended on the individual, in contrast, the legalistic department acts as if there is a single standard of community conduct (Wilson 1968). There is an emphasis to distinguish the police administration from the politics in the delivery of police services. Although both the legalistic and watchman departmental styles tend to be found in communities with diverse populations and population instability, the deciding factor is the influence the community has over its police organization. The difference is determined by the strength and involvement of the local government (Langworthy and Travis 1999). In communities with a strong, involved local government (elected officials, professional administrators, and police chief) the legalistic style emerges, as it is believed the best way to serve the constituency is by applying the law equally. On the other hand, in those communities where the strong, involved local leadership is missing, the applicability of the law is determined by the individual officer and the style of policing exhibited is watchman (order maintenance). Langworthy and Travis (1999) elaborate on Wilson s legalistic style: Officers in legalistic agencies intervene frequently and formally with citizens. Criminal law is relied upon to determine situations requiring intervention as well as the manner of intervention. The form of intervention is to enforce the law either by issuing citations or making arrests, therefore, this type of department is characterized by officers who issue many citations and make many arrests. Officers interact with citizens as ministers of the law (Langworthy and Travis 1999:163). Departments exhibiting the legalistic style have a high number of contacts with citizens and make more arrests and issue more citations. In these departments it would be expected that there would be frequent and formal contact with citizens. 12

21 Service Style Service department officers intervene frequently but not formally with citizens. These types of departments are generally found in middle class communities where there is perceived to be a high level of apparent agreement on the need for and definition of public order and no administrative demand for a legalistic style (Wilson 1978:200). The police in these departments protect the public order from minor threats (i.e. unruly teenagers and outsiders ). Juveniles are generally treated informally and suspicious persons are observed. The few family quarrels that may arise are addressed as private matters requiring decorous conduct. Illegal activity in these communities should be low since it would not be tolerated by the public and it should be fairly easy to identify by the police. This reduces the amount of crime that would be committed by residents of that particular community. Low crime rates provide police with more opportunity to focus on traffic, juveniles and service (Wilson 1968). Following Wilson s (1968) hypotheses, it would be expected that there would be many police encounters with citizens which result in few formal sanctions in service style departments. In these types of communities police can act as if their task were to estimate the market for police services and to produce a product that meets the demand (Wilson 1968:200). Generally there are few dispatch calls and the citizenry expects customer service. For minor infractions arrests are avoided when possible. Although there will be increased rates for traffic tickets and juvenile referrals to family court, these rates will be lower than in legalistic departments. There will be frequent use of informal, non-arrest sanctions (warnings). There should also be contacts with the public that are outside of the criminal justice sphere (house checks, door unlocking, etc.). 13

22 Langworthy and Travis (1999) comment on the service style: Officers in service style agencies intervene frequently, but informally with citizens. There are many contacts with citizens but few arrests or citations. Service delivery is stressed in this type of department. Citizen problems are responded to as officers handle the problems generally by means other than invoking the law. Warnings are more likely than arrests or citations. Resolving citizen problems is the primary task in the service agency. Police act as gatekeepers who often direct citizens to relevant services (Langworthy and Travis 1999:163). According to Wilson, a department that exhibits high frequency of contact and high formality of response would be considered a legalistic department. The watchman department would be expected to exhibit low frequency of contact and low formality of response. A service department would be recognized by high frequency of contact and low formality of response. Empirical Test Control variables in Wilson s study included: community s population size, median family income, and non-white population. Using 1960 data, cities with populations between 25,000 and 100,000 persons, median family annual income between $5,000 and $7,500, and contained a nonwhite population less than or equal to five percent were included in the sample (n=146). Wilson believed that controlling for these variables 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 (Wilson 1968:272). He eliminated high poverty and high minority cities believing that communities without these extremes allowed police have ample opportunity to focus on the types of discretionary offenses, rather than more serious crimes. Wilson tested the following hypotheses: 1. Cities with professionalized governance would be likely to have legalistic police departments where there were a higher number of arrests for law enforcement situations (larceny and driving while intoxicated); 14

23 2. Cities with partisan, mayor-council regimes would exhibit the watchman style of policing where more order maintenance arrests were initiated by officers (disorderly conduct); 3. There would be relatively little difference in arrest rates between police departments, and thus between cities, for citizen-initiated, order maintenance situations. Wilson presented tables which examined arrest rates for five offenses by community political system. The data presented four categories of community political systems and provided the total arrest rates for each type of city by each offense type. The findings were derived from this simple analysis. 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 1968:276). Limitations of Existing Theory The section above described Wilson s theory as it was formulated in The following discussion will elaborate on issues that have changed which may have an impact on his formulation and measures used to test the theory. There are several issues relevant to the variables used by Wilson. First, since the 1960s the council manager government has become the most popular form of local government structure (ICMA 2006). In 1960, 38 percent of American cities with populations over 5,000 operated under a council-manager system (ICMA 1960). By 2006, 49.1 percent of municipalities with populations over 2,500 operated under the council-manager form of government (ICMA 2006). This form of government is more prevalent in cities with populations of 25,000 or more (63 percent). Since more cities have adopted the council-manager form of government variation among types of government diminishes. In addition, research suggests that there is a convergence of the mayor-council and city manager governmental structures (DeSantis and Renner 2002). Wilson also had access to unpublished 15

24 data for educational levels of the city managers in his original study; these data have not been available in subsequent studies. Second, although Wilson attempted to account for the zone of indifference that gives officers the latitude to act in discretionary situations by using an arrest measure for the dependent variable it does not account for the opportunity that some offenses may present. For example, large cities are heterogeneous having a mix of activities (residential, commercial, etc.) and there may be ample cases of drunkenness for each large city at the aggregate level; however, suburbs tend to be more homogenous. Some suburbs may not present the opportunity structure for particular order maintenance offenses as measured by Wilson. Some communities do not have commercial areas that contain bars or liquor establishments; others have little if any commercial areas. A high number of arrests would not be expected to be found for drunkenness, disorderly conduct, or DWI offenses in these communities. Also, many suburbs do not have large amounts of public open space (sidewalks, public parks) that cities do, therefore, the on view or policeinitiated order maintenance offenses included by Wilson (drunkenness and disorderly conduct) may not be viewable by officers as frequently as they are in urban areas. In addition, the raw data used by Wilson do not reflect those discretionary situations where officers had the opportunity to take action but did not. Wilson s data reflect only arrests made for particular offenses. Similar to the opportunity issue above, the composition of many suburbs is homogeneous in terms of race, income, and other factors. Compared to the large cities included in his original study, the composition of suburban communities may be different. Berry (1972) notes that the socio-economic status (SES) of major central cities will have average status scores because they are very large polities spanning a wide range of community areas of all status levels within them, 16

25 and because they have a diversified economic structure with a broad range of job types (p. 24). In short, the differences in the values of Wilson s control variables in cities get averaged out. The homogeneity of suburbs within the current study should provide greater overall variation to allow comparisons across communities. Although Wilson s narrative of the eight cities under study focused on vice, gambling and prostitution offenses, his empirical test using national data include five offenses (larceny, DWI, drunkenness, simple assault and disorderly conduct) to depict order maintenance and law enforcement arrests. He selects these offenses to highlight the wide discretionary latitude officers have in enforcing the law. What is not known are the number of situations where the police used their discretion to handle the situation short of arrest. All that is known are the number of arrests that resulted for the five offense types which Wilson uses to represent order maintenance and law enforcement activities. In the following chapter existing research on Wilson s theory and an introduction of alternative variables and their hypothesized relationship between community characteristics and frequency and formality of police-citizen encounters are discussed. 17

26 CHAPTER THREE: REVIEW OF THE LITERATURE Tests of Local Political Culture The inability of empirical tests to find consistent support for the theory of local political culture may be rooted in the conceptualization of political culture. The following section covers the tests of Wilson s original research on the topic. There have been four noted empirical tests of Wilson (Langworthy 1985; Zhao and Hassell 2005; Zhao, He and Lovrich 2006 and Liederbach and Travis 2008) where the authors have set out to test Wilson. Other scholars have used political culture as a competing variable in the examination of legalistic police behavior for non-legally serious offenses (Crank 1990). Langworthy 1985 Langworthy (1985), using 1975 FBI arrest and International City Manager Association (ICMA) data, was one of the first to test Wilson s theory. The sample included 152 cities with populations over 100,000 (1970 data). There was more variation in these data than the original because there was no attempt to eliminate cities at either extreme in terms of size, income or race as done by Wilson. Since detailed information regarding education and experience of city managers was not available for this period [Wilson had used unpublished data sets from the ICMA (Langworthy 1985)] Langworthy presented a three-fold political culture typology rather than four. His typology included good, mixed and traditional governments. Good government cities were characterized by council-manager forms of government and nonpartisan elections. Traditional cities were represented by mayor-council forms of government and 18

27 partisan elections. Mixed government cities were those that either conducted partisan elections or had an established mayor-council form of government. Langworthy employed analysis of variance (ANOVA) and box plots in his analysis of data. The ANOVA results supported Wilson s original hypotheses. First, good governments had higher arrest rates for larceny and DWI than other cities. Second, traditional governments had higher arrest rates for disorderly conduct. Third, arrest rates for simple assault were similar across city typologies. Fourth, good government cities had higher arrest rates for drunkenness. In addition to ANOVA, a visual examination utilizing box plots was presented. Arrests for larceny, DWI and simple assault were nearly identical across the three government typologies. Drunkenness and disorderly conduct arrest rates exhibited differences. Traditional government cities exhibited less variation in drunkenness arrests. Langworthy states that the constraining effect of political culture (government type) on drunkenness arrests is evident in traditional cities. There was less variability in disorderly conduct arrests in good and mixed government cities. Overlap in arrest rates was found across city government types for different arrest types. Ultimately, Langworthy found that arrest rates for all but two offenses were similar across government types suggesting that Wilson s theory was one more of tendency than constraint. Zhao and Hassell 2005 Another study examining large municipal police agencies was conducted by Zhao and Hassell. There were 304 communities with large agencies in their sample. Large was defined as 100 or more full-time sworn officers. Using a coding method similar to Wilson s, political culture was operationalized as follows: (a) professional government (legalistic); (b) traditional government (watchman); and (c) mixed type of government (service). Five 19

28 control variables were included: city size, percent of population non-white, percent of population below the poverty line, percent of population aged 15 to 24, and geographic region. Wilson included only the first three variables in his analysis; unlike the current authors, he measured median family income and not percent of population below the poverty line. The dependent variable, arrest rates per 10,000 population, was measured for the following offenses: larceny, driving while intoxicated (DWI), drunkenness, and disorderly conduct identical to Wilson. According to Wilson s hypotheses, communities with professional political cultures are more likely to have higher arrest rates for law enforcement activities (larceny and DWI). Communities with a watchman orientation are more likely to have higher arrests for drunkenness and disorderly conduct. For larceny arrests, the authors found the effect of professional agencies to be significant, however, opposite the hypothesized direction: professional (legalistic) departments had fewer arrests for larceny than those departments located in cities with a traditional (watchman) political culture. Percent minority and percent below the poverty line were stronger predictors than political culture. The regression analysis for DWI arrest rates also found that control variables (percent minority, percentage below the poverty line and geographic region) were stronger predictors than political culture. Wilson also hypothesized that watchman style agencies were more likely to focus on drunkenness in their cities. The regression analysis for drunkenness arrest rates indicated that the local political culture had no effect, however, the three control variables (percent minority, percentage below the poverty line and percent of population ages 15-24) were significant. Additionally, agencies in the south were more likely to have higher arrest rates for drunkenness. 20

29 The model which examines the relationship between political culture and the arrest rate for disorderly conduct finds that similar to larceny arrests, political culture is significant, but opposite the hypothesized direction. Fewer disorderly conduct arrests were made in professional cities than in cities that operate with traditional (watchman) political culture. The only control variable found to be significantly related to disorderly arrest was percent minority population. The authors state that (o)verall, our findings indicate a move toward a falsification of Wilson s theory of local political culture in explaining today s (emphasis original) police organizational behavior (Zhao and Hassell 2005:243). Reasons given to support their contention include the policing reform movement that gained acceptance and momentum after the time of Wilson s writing and the influence of federal funding of police agencies. The following studies employ a measure of political culture in non-urban areas. In the following article Crank (1990) studied political culture as but one measure in determining police outcomes and he examines different offense codes for smaller communities. Liederbach and Travis (2008) use systematic social observation in suburban communities. Crank 1990 Crank (1990) examined urban and rural municipal police departments in Illinois communities with populations ranging from 2,612 to 97,030. Communities smaller than 2,500 population were excluded from the study because details on community structure were not available. In all, there were 284 communities (161 urban, 123 rural). Similar to Wilson, Crank used arrest statistics as a measure of police style. Non-index UCR offenses reported to the Illinois State Police for the year 1986 were used for the following offenses (these differ from Wilson s): trespass, disorderly conduct, motor vehicle, and cannabis control. These offenses 21

30 were thought to be illustrative of high discretionary offenses which allowed wide latitude for police to enforce with the potential outcome of arrest, thus reflecting the role of agency style to influence their decision making. Demographic data were obtained through the 1980 U.S. Census and governmental structure variables were obtained from the ICMA. Policing styles in urban and rural communities were compared using controls for culturalracial heterogeneity, economic conditions and city governance. Crank points out that the factors influencing police styles vary between urban and rural areas. This is the first study to utilize smaller community arrest data and suggests that results may be important in expanding the understanding of policing styles in non-urban settings. For the four offenses included in his analysis arrest rates were positively associated with the city manager form of government, supporting Wilson s hypothesis, albeit with different offense categories than the ones Wilson originally used. In rural police departments, Crank (1990) found a positive relationship between trespass, disorderly conduct, motor vehicle-related and cannabis control arrests and city manager type, whereas he found the same relationship for only vehicle related arrests in urban areas. Liederbach and Travis 2008 Liederbach and Travis tested Wilson s theory with twenty suburban communities in the Midwest using systematic social observation (SSO) data. Populations of the communities ranged from 899 to 60,192. Frequency and formality of police-citizen interactions were used as a style measure in lieu of arrests. Using Chi-square analysis the authors did not find support for Wilson s theory. They did, however, find a significant association between population and median household income, and policing style. Liederbach and Travis do not align themselves with other scholars who suggest a movement toward falsification of the theory. They discuss the 22

31 various reasons theirs and other findings did not find support for Wilson. Their explanations include the absence of exact replications of Wilson s original test and changes in political, social and occupational environments. Progressive-era reforms which started the drive toward police and city administration professionalism have been widely implemented since the late 1960s the time of Wilson s writing. Officers have lost much of their discretion in arrest decisions for such offenses as DUI and domestic violence due to legal changes as many states and localities have mandated arrest policies for these offenses. Changes in the policing field including the composition of staff, education and training requirements, non-local recruitment of officers, and increased intergovernmental relations (IGR) have all been hypothesized to explain why Wilson s theory has not been supported. Liederbach and Travis confirm that there are differences between communities and policing style. All said, the authors do not ring the death knell for Wilson s theory, instead they state that the theory should retain an intuitive albeit quantitatively challenged appeal. There is no clear consensus in support of Wilson s theory today. Indeed, in recent years there has been a lack of support for the theory. Much of this trend toward falsification of his theory has been pegged to the changes in legislation and departmental policy that structures officer discretion. These changes occurred after Wilson s original writing. In order to address the issue of decreased discretion in officer decision making, the current study uses measures of frequency and formality in police-citizen encounters and identifies alternative community characteristics hypothesized to correlate with police-citizen interaction and ultimately styles of policing identified by Wilson. Police style operationalization has been improved through the use of available data provided by systematic social observation. Other issues which may explain the 23