Measuring the Performance of the Police: The Perspective of the Public

Executive Summary

The aim of this project is to review and critically assess the current survey methods used to measure police performance in common law jurisdictions. Specifically, an emphasis is placed on the questions that are asked on different public opinion and community surveys in Canada and internationally, with the ultimate goal of recommending better approaches to conducting such surveys. Other police performance measures, such as operational metrics, are also touched upon, albeit to a lesser extent.

A comprehensive and systematic literature review of published research of Canadian, as well as international literature in the field of police performance measurement through surveys was conducted. Surveys on all levels of geography – national, provincial, or municipal – were considered for review. Particular attention was paid to analytical research that looked into the actual questions asked on surveys, their meaning and whether they measure what they are supposed to measure. Analysis focused on public views of the police, and a considerable part of the present paper addresses the issue of satisfaction, trust, and confidence in the police.

Police undertake a broad array of work. Other than the traditionally-assigned tasks of pursuing, arresting and charging criminals, preventing crime from occurring, and dealing with traffic-related offences and accidents, police are further expected to resolve various conflicts in their communities, reduce or prevent social disorder, and construct and maintain community relations. It is important to understand that the measurement of police performance is a complicated task that has multiple dimensions (Coleman, 2012; Maguire, 2003; Moore and Braga, 2003). There is no single measure that will be even remotely close to measuring the performance of everything the police does.

Both direct and indirect measures need to be accounted for when attempting to measure police performance. Direct measures of police performance commonly used include crime rates, number of arrests and fines issued, clearance rates, and calls for service response time. Some indirect measures of police performance include surveys, direct observations of social behaviour, situational studies and independent testing.

Measurements of police performance through public opinion polling include: 1) general questions on satisfaction with police and 2) specific questions on police performance. The general questions on satisfaction with police asked on surveys is supposed to be the simplest and quickest way to measure the overall level of satisfaction of citizens with the police. It is important to ask these types of questions because: a) they provide a quick indicator for the overall support for police among citizens; b) they carry implications for the support constituents give to police work; and c) a decrease in the perceived legitimacy of the police could potentially lead to non-compliance with the authority of the police and increased crime rates (LaFree, 1998; Tyler, 1990).

The generalized questions on satisfaction with police tend to be too general to tease out specific information on what it is that citizens favour about police services. More importantly, it is impossible to tease out that about which citizens are unhappy. Other methodological issues associated with generalized questions include unstandardized, inconsistent, and at times poor choice of wording or questions and response categories making comparisons across time and place impossible. There seems to be confusion when it comes to the meaning of the terms “favourable views of,” “confidence in” and “trust of” the police; these terms tend to be used interchangeably on surveys, when in fact they measure rather different public sentiments.

When it comes to specific questions on police performance, there are many examples of these types of questions that are being asked of police in contemporary public opinion surveys in Canada and the Western World. Unfortunately, these questions are used in a more or less arbitrary manner without being properly tested for validity and reliability. Ground work on the meaning of the questions, whether they measure what they are supposed to measure (reliability), and how well they measure it (validity), is largely absent in the literature.

There are three notable studies that attempt to conceptualize and test specific measures of police performance measured through public opinion polling: Maguire and Johnson (2010) in the U.S. who derived questions from theoretical constructs and tested them against Mastrofski’s six dimensions of policing as a service industry;^Footnote1 Coleman (2012) in Canada who consulted policing stakeholders in order to derive and test the factors that may constitute police performance; and Jackson et al (2011) in the U.K. and Europe who derived, tested and implemented European indicators of trust in judicial systems.

In Canada, only one national survey administered by Statistics Canada asks six questions on police performance. This constitutes the only source of public opinion police performance metrics that is comparable across time and place in Canada. The majority of Canadian municipal police services commission their own annual or bi-annual public opinion community surveys that include numerous general and specific police performance questions. However, the questions tend to be unstandardized, inconsistent, with varying response category thus making them incomparable across time and place.

More research into either testing the existing questions or creating new standardized questions is required to improve the measurement of public satisfaction with the services provided by the police in Canada. A project similar to the creation of European indicators of trust in justice (Jackson et al, 2011) that involves discussion among experts, proper testing and re-testing of measures of police performance needs to take place. In the absence of valid, reliable, and standardized indicators, police services and policy makers will continue to ask different, sometimes vague questions that are not comparable. Such inconsistent measurement of police performance runs the risk of poorly evaluating the performance or police and policing policies and practices across Canada, leading to inefficient and ineffective policing and ultimately compromising the public safety of Canadians.

Introduction

In 2012, the number of police officers in Canada reached 69,539, representing eight consecutive years of growth and the highest police officer strength since 1981. Total spending on policing was over $12.5 billion in 2012, more than double its 1997 level (Statistics Canada, 2014).

The reported crime rate in Canada has been dropping steadily over the past two decades, from its peak in the early 1990s. However, notwithstanding declining crime rates, police have been increasingly called upon to address a wide range of non‑criminal calls for service, including social and mental health incidents (Statistics Canada, 2011).

Growing policing costs are not sustainable for many jurisdictions, particularly given competing priorities for public funds, such as education and healthcare. Provinces and municipalities, as well as policing stakeholders, have voiced their concerns on this issue as they are finding it increasingly difficult to justify paying these levels for police services (Brennan, 2014; CBC, 2013).

In an environment of limited resources, it becomes increasingly important to measure and evaluate the performance of policing services to ensure the performance of the service is in line with the objectives set out for it. Public Safety Canada led the coordination of a national conference entitled the Summit on the Economics of Policing: Strengthening Canada’s Policing Advantage, in January 2013, in Ottawa. The purpose of the Summit was to increase awareness of the challenges and opportunities facing policing, provide practical information on improving efficiency and effectiveness, and strengthen the foundation for innovation and reform in Canadian policing.

The emerging debate on how to best achieve long-term, sustainable levels of policing has included a discussion of targeting and refining the application of policing resources to ensure police service delivery is both efficient and effective. It is recognized that, in order to assess the activities and performance of police, as well as how and where resources should be applied, objective performance metrics are required.

The Police Information and Statistics Committee (POLIS)^Footnote2 is responsible for collecting national level crime and policing data. As such, POLIS ensures the integrity and comparability of many basic police performance measures through facilitating the collection and analysis of data by police services and Statistics Canada. This information includes trends in police-reported crime, satisfaction with the police, numbers of police employees and spending on policing, amongst other measures. In 2011, members of POLIS became aware that a generally increasing trend in the difference between the rate of self-reported criminal victimization and police-reported crime was being observed. This concerned POLIS members because it could mean that the value of traditional police-reported crime statistics as a measure of crime in Canadian society, or as a performance measure for policing, was being eroded. It could also indicate that the growing gap might partly be due to falling confidence in the police by members of the public. In order to collect data to address these concerns, Public Safety Canada consulted with POLIS members in order to reform the questions on the General Social Survey, Victimization Module, that related to ‘reason for not reporting to police.’

Preliminary literature scans indicated that there have been studies of how police performance can be measured using polling, such as through self-reported victimization or questions on confidence in the police, but that there were significant information gaps in how to undertake a refined analysis of this type of data, a lack of Canada-specific material, and a lack of easily accessible summaries that could be useful to police services. To this end, Public Safety Canada commissioned a report on the views of Canadian police boards on the use of performance metrics (see Kiedrowski et al, 2013). Among other findings, the results of the study indicated that uniformity and consistency in performance measurement frameworks among Canadian police jurisdictions was lacking. Further, the study indicated that “[w]hen valid and reliable quantitative performance metrics were included they were most often operational indictors” (Kiedrowski et al, 2013: II).

While operational performance metrics are indeed very useful indicators of police performance, they are prone to various measurement errors. Other performance indicators, such as the views of the public on the performance of the police, are often not present in the police performance frameworks. These metrics can only be measured through questions asked on public opinion surveys, such as the ones assessing the levels of people’s general confidence or trust in the police or specific issues that people trust the police to address.

Numerous polling questions related to policing are currently asked in Canada and around the world on polls, all of which measure similar concepts (see illustrative examples of these questions in the Appendix). However, the wording and categorization of the questions are different across Canada, making comparing the questions across times and places exceedingly difficult. Further, the validity and reliability of these questions has rarely been tested. It is the intent of this paper to discuss those metrics and to contribute further to the pool of knowledge on police performance metrics gathered through public opinion polling.

Project Objective

The aim of this project is to review and critically assess the current survey methods used to measure police performance in common law jurisdictions. Specifically, an emphasis is placed on the questions that are asked on different public opinion and community surveys in Canada and internationally, with the ultimate goal of recommending better approaches to conducting such surveys. Other police performance measures, such as operational metrics, are also touched upon, albeit to a lesser extent.

Methodology

A comprehensive and systematic literature review of published research of Canadian, as well as international literature in the field of police performance measurement through surveys was conducted. Surveys on all levels of geography – national, provincial, or municipal – were considered for review. Particular attention was paid to analytical research that looked into the actual questions asked on surveys, their meaning and whether they measure what they are supposed to measure. Analysis focused on public views of the police, and a considerable part of the present paper addresses the issue of satisfaction, trust, and confidence in the police.

This review follows the methodological standards of source retrieval identified by the Campbell Collaboration in Searching for Studies: A Guide to Information Retrieval for Campbell Systematic Reviews. Relevant articles for this review were identified mainly through electronic databases using keyword searches, but were augmented by a hand search strategy developed from the initial search of the electronic databases.

In addition to keyword searches on databases, a search for surveys on websites of the police organizations representing the major Canadian Census Metropolitan Areas (CMA) and the Royal Canadian Mounted Police (RCMP) was conducted. Relevant surveys, if available, were analyzed and discussed.

Finally, a search on Google was conducted using the same keywords proposed for search on databases. Hand search through the gray literature found using Google helped identify further literature that is relevant for analysis.

Proposed Databases and Keywords for Literature Review:

Databases

• Academic Search Premier
• Criminology: a SAGE Full-text Collection
• EBSCO databases
• IngentaConnect
• Journal Storage (JSTOR)
• Social Science Research Network (SSRN)
• Social Sciences Citation Index (SocINDEX)
• Theses Canada
• Google

Keywords

• Search 1: (polic* AND measur* OR performance) AND (survey)
• Search 2: (polic* AND measur* OR performance) AND (public opinion)
• Search 3: (law enforcement AND measur* OR performance) AND (survey)
• Search 4: (law enforcement AND measur* OR performance) AND (public opinion)

Performance Measurement

According to Neely et al (1995), performance measurement may be defined as the process of quantifying the efficiency or effectiveness of an action, while a performance measure is the metric used in the quantification process.

The words “quantifying” and “metrics” may suggest to some that performance measurement must include some type of counting. A sign hanging in Albert Einstein’s office refutes that suggestion, proposing that measurement is much more than just counting, particularly with respect to achieving qualitative objectives: “The problem is that not everything that counts can be counted, and not everything that can be counted counts.” Whitaker et al (1982) discusses how best to measure, where he claims that observing is not measuring, as measuring requires the comparison of an observation to a standard of measure.

In both the private and public sectors, performance measurement is considered a necessary component of good management. Experts in the field of management consulting have commented on the concept of performance measurement, with Peter Drucker, the well-known management guru stating: “If you can’t measure it, you can’t manage it.” Michael A. Lebowitz (2010: 1) expounded on that idea during a speech: “There's an old saying that if you don't know where you want to go, any road will take you there…this saying is mistaken. If you don’t know where you want to go, no road will take you there. In other words, you need an understanding of the goal.”

Organizations may benefit from these ideas by establishing objectives; without knowing what the organization wants to achieve, it is difficult to know when the objective has been reached. As a corollary of this practice, organizations should consider how success will be measured in order to be able to know that the objectives have been met. Converting these concepts into action is not as easy as it may sound. It is important that organizations understand that measuring performance is a complex endeavour, usually involving much more than counting outputs.

In his book, Basic Issues in Police Performance (1982), Gordon Whitaker and colleagues describe some of the considerations of such an endeavour. First, before even determining which measures to use, there has to be a commitment by management to support the implementation and management of a performance measurement framework; without such a commitment, it is unlikely that scarce resources would be dedicated to these efforts. Second, data collection is frequently difficult to undertake in a policing context for a number of reasons, for example: those collecting the data may not be qualified or trained properly; crime victims may not know the proper terminology to use when reporting crime, thereby skewing the reporting statistics; or there may be subjective bias in reporting by police where, for example, the data indicates poor performance. Third, the validity of the data may be questionable. For example, police-reported crime rates depend not only on the occurrence of crime, but also on citizens’ reporting of crime, which may result in a reported rise in crime, when in reality, the ‘rise’ was attributable to better reporting. Fourth, being measured may in and of itself affect performance. For example, if performance is being measured by the number of arrests made or the number of complaints, it would be natural for a police officer to change his or her behaviour to ensure ‘good’ results, when this might not actually be ‘normal’ practice for the officer. Finally, efficiency performance measures can be complicated as they depend upon the relationship among inputs, outputs and outcomes. These components are often difficult to define and isolate for any given police service or officer, especially when there is a wide variety of responsibilities and tasks to accomplish.

Police Performance Measurement

Measurement of police performance has been attempted since the establishment of the institution of modern policing in the mid-nineteenth century (Maguire and Uchida, 2000). The intent of measuring performance was to track how well the institution was performing the duties assigned to it. By and large, the measures were done on a local community or municipal level, and focused strictly on the inputs, activities, and outputs of police work (Maguire, 2003).

The initial measurements of police performance involved direct measures only: crime rates; number of arrests and fines; clearance rates; and response times. It was not until the late 1930’s during the so-called “professionalization of policing” that the importance of citizens’ views on police performance was recognized in the Unites States. The surveys asking for the opinion of the general public on the matter became popular and formed a part of performance measurement models in the 1960’s and 1970’s.

Police undertake a broad array of work. Other than the traditionally-assigned tasks of pursuing, arresting and charging criminals, preventing crime from occurring, and dealing with traffic-related offences and accidents, police are further expected to resolve various conflicts in their communities, reduce or prevent social disorder, and construct and maintain community relations. Having recognized the broad spectrum of tasks assigned to the police, it is important to understand that the measurement of police performance is a complicated task that has multiple dimensions (Coleman, 2012; Leckie, 2012; Kiedrowski et al, 2013; Maguire, 2003; Moore and Braga, 2003). There is no single measure that will be even remotely close to measuring the performance of everything the police do. The process of measuring the success of the police work is far from straightforward.

Both direct and indirect measures need to be accounted for when attempting to measure police performance. Each of the possible measures has strengths and weaknesses that are important to understand prior to opening a larger discussion on which ones to adopt. The direct measures of police performance, which are sometimes referred to as traditional measures throughout this report, are very powerful measures that can provide an objective way to measure police performance, when used correctly.

Crime rates, along with the more sophisticated measure of the Crime Severity Index (Statistics Canada, 2009), are probably the oldest and the most traditional measures of the success of police work. The reasoning behind using crime rates as a measure of success is reflective of the common belief that reducing crime is a primary task of the police. Indeed, a reduction in crime rates can be an indication of the excellent work being done by the police. However, there are at least three problems with using crime rates or the Crime Severity Index as the main measure for police performance. First, the work of the police is, by far, not the only influence on crime rates. An array of factors such as levels of education, rates of poverty, the age of the local population, rates of addiction or other mental health problems, the number of transient residents, etc. influence the level of crime in a community (Gomes, 2007). Unless statistically controlled for and weighted by these and other factors, crime rates alone cannot, and should not, be used as the sole performance measure of police work.

Second, crime that is reported to and recorded by the police is heavily dependent on the processes of police departments that receive and record the report. The behaviors of police officers receiving the complaint, whether and how the complaint was recorded in the system, the departmental practices that encourage or prevent reporting, if the police services collect data on crime reported to non-police entities in their jurisdiction, or statistical reporting practices can all have a significant influence on crime rate statistics.

Finally, it is well established that not all crime is reported to the police. Victims of crime might be reluctant to report incidents to the police for a number of reasons, including fear of retaliation from the offender, embarrassment, fear of deportation in the case of immigrants, fear of secondary victimization through the interaction with the criminal justice system, thinking that the damage was not big enough or that the police would not be able to help. Regardless of the reasons for not reporting crimes to the police, unreported crime would greatly influence the true rate of crime. In this sense, it is important to use victimization surveys, such as the Victimization cycle of the General Social Survey (GSS) in Canada, to supplement the police-reported crime rate with the rate of unreported crime.

The number of arrests and fines issued are other traditional measures of police performance. Similar to crime rates, the number of arrests and fines issued by the police is a measure of performance because arresting criminals and enforcing the law by giving out fines for infractions is seen as one of the primary outcomes of police work.

Using arrest and infraction data as a primary performance measurement for police work can be problematic for several reasons. First, the definition of arrest can vary among different agencies. Various studies in the U.S., for example, show that uniformity in how police agencies define arrest is lacking, therefore making comparisons between agencies close to impossible (Sherman, 1980a; Sherman, 1980b). Second, an arrest is but one of the options for action for police officers in a difficult situation. In this sense, an arrest could be seen as a failure by a police officer to employ other strategies to resolve the situation in another manner. For example, petty crimes such a theft of small, inexpensive items in a supermarket or possession of a small amount of marijuana could result in issuance of a warning rather than an arrest. For youthful offenders, in certain cases, informal restitution arrangements or transfer to parental responsibility might also be an outcome of police involvement.

Similar to arrests, the infractions given out may vary greatly across police jurisdictions due to differences in operational police policies. Some jurisdictions may be seen as harsher enforcers of traffic laws, whereas others may be seen as more lenient. Some jurisdictions may decide to invest more in roadside alcohol-testing program, which will result in a greater number of infractions and charges. Provincial and territorial traffic laws vary as well, paving the way for variations in enforcing the laws.

Lastly, arrests are time and place sensitive. For example, arrest and infraction data spiked in Toronto during the G8 and in Huntsville during the G20 summits in 2010.^Footnote3 This spike in the number of arrests should not be seen as a measure of police performance because these are special, international events that saw a great number of protesters and have skewed the arrest and infraction data in the province of Ontario for 2010.

Clearance rates, which are the proportion of crimes solved by a police jurisdiction in a given period of time, are yet another traditional measure of police performance that has been used widely. In an ideal type of scenario, this indicator portrays the efficiency of police to counter crime in their assigned areas. The trouble with using the clearance rate as a measurement of performance is that it is prone to definition and measurement errors, making cross-comparisons difficult (Maguire, 2003; Riedel and Jarvis, 1999). For example, smaller communities may have a lower crime rate, which means police officers have fewer opportunities to solve crimes. On an individual basis, these officers would not fare well in statistical comparisons with their counterparts serving higher crime communities since their clearance rates would naturally be lower. Further, this measurement is less meaningful if it cannot be replicated or put into the context of other police services. A cross-sectional comparison that would control for number of people, neighbourhoods, businesses, and crime rates might be useful to partially address this issue (Whitaker et al, 1982).

Clearance rates are also fairly simple to manipulate in the case where a certain police jurisdiction wishes to produce a higher rate (Cordner, 1989). There may, for example, be a bias related to who is collecting the data. In particular, police have a vested interest in providing data that portrays them in a favourable light (Gomes, 2007; Lithopoulos and Rigakos, 2005; Paré and Ouimet, 2003; and Whitaker et al, 1982), yet police services themselves are the source of most data used to measure their own performance.

Call for service response times have been used in the past to evaluate performance. These indicators are not as popular as the other three traditional performance measures, and are also prone to various errors. Response times may vary according to the size of geographic area of community served, resources allocated to policing, location of the closest dispatching station, etc. Further, not all incidents require a police response, and not all the citizens calling the police are necessarily interested in police appearing at the scene of the incident. Interestingly, research has also shown that it is not the actual response time but the perception of the right response time that matters with regard to citizens’ satisfaction with the time it takes police to respond (Hatry, 1999). Put differently, it is often the case that providing specific or expected time of response by dispatchers is more important than a unit that responds more quickly (Maguire, 2003). In this sense, using a customer or citizen measure of satisfaction with response time is probably a better measure of police performance than using actual response times.

It is nearly impossible to correct for the human errors or at-source manipulation of data of the direct measures of police performance. Errors in recording arrest data, for example, cannot be corrected unless the error is detected early in the process. Likewise, manipulation of clearance rates by a police jurisdiction cannot be corrected unless the data is verified by an independent authority. However, it may be possible to adjust the measures through the application of various statistical controls and weights (Maguire, 2004). Depending on the availability of data, the size and the social conditions of the community served by the police could be controlled for during the analysis.

Nevertheless, it is evident that measuring police performance using the traditional measures alone is not sufficient. To measure the police’s ability to arrest criminals only would be to completely disregard the other important work done by the police. The multi-dimensional nature of police’s work requires a multi-dimensional approach to the measurement of their performance. Researchers have thus resorted to indirect measures of police performance such as surveys, direct observations, and situational studies and independent testing. These are briefly summarized below.

Surveys can be a rich source of indicators that measure police performance. They may be administered by police jurisdictions, other state or provincial law enforcement authorities, academics, public opinion polling companies, or national institutions collecting statistical data. Surveys can be a very powerful tool to collect data that would supplement the traditional measures of police performance. The types of questions asked on surveys depend on which dimension of police work the researcher is attempting to measure. The wording of questions and their placement within the survey influence the answers of the respondents.

There are different types of surveys that researchers can employ in order to derive these indicators.

• General community or national surveys usually ask questions about how the public views police performance, often resembling customer satisfaction surveys. Issues covered may include satisfaction with the work done by the police, the effectiveness with which police deal with social disorder and crime, whether individuals feel protected by the police, etc.
• Contact or victimization surveys target individuals who were in contact with the police, or who were victims of crimes and dealt with the police following their incident. These kinds of surveys usually ask questions regarding individuals’ experiences dealing with the law enforcement authorities.
• Employee surveys are mostly internal surveys asking police employees about their work. They can be used to assess the levels of job satisfaction and the integrity of police employees.

Direct observations of social behavior or certain phenomenon is a research method that is used widely in criminology. For example, social disorder was observed by researchers in Chicago, where occurrences of the phenomenon were recorded on camera, coded, and analyzed (Sampson and Raudenbush, 1999). In police performance research, Mastrofski et al (1998) recorded and analyzed encounters between police officers and individuals to evaluate the behavior of police officers.

Situational studies/independent testing is yet another method that can be used to assess the performance of police. In one example of this method, an American news network, ABC News, handed over 40 wallets that were supposedly found on the street to police officers in New York City and Los Angeles. All the wallets were returned with full contents in this police integrity test (ABC News, 2001, cited in Maguire, 2004). There were other studies conducted using this technique in the U.S., a notable example of which is by the Police Complaint Center, a Florida-based organization that conducts inquiries into police misconduct (Maguire, 2004). Although not widely practiced in academic research, probably due to ethical concerns and costs, the technique provides an objective alternative to other popular methods.

Dimensions of Police Work and Their Measurement

The multi-dimensional nature of police work naturally calls for a multi-dimensional approach to police performance measurement. Some academics suggest that police performance should be measured by multiple indicators, ranging from the individual victims’ accounts (Neyroud, 2008) to more formalized public satisfaction surveys. Before discussing the actual performance measures, a short description of the dimensions of police work should be provided. One of the recent and most discussed examples in the academic literature on the multi-dimensional framework of police work can be found in Moore and Braga (2003: 17-24). They identified seven dimensions of police work.

These seven dimensions of police work can be used when setting up a framework to measure the performance of the police (Keidrowski et al, 2013). There is not one single performance measure that can assess all of the dimensions of police work at the same time; neither is there one perfect measure that can assess a single dimension of police work (Gallagher et al, 2001). The measuring instrument needs to be chosen depending on the dimension that needs to be measured. Each of the dimensions should be measured using an array of indicators, including the traditional direct measures of crime rates, arrests and infractions, clearance rates, and response times, but supplemented with indirect measures through surveys, observations, or situational studies.

1. Reduce crime and victimization.  This is thought of as the primary and a most important dimension of police work.  Police are meant to protect citizens and their well-being, as well as reduce both the actual and perceived danger of victimization and crime in the community.
   
   The performance of the police in the dimension of reducing crime and victimization has been historically measured through police-reported crimes that are compiled into the Uniform Crime Reporting Survey (UCR) in Canada. It has basic information on the type of offence committed, the characteristics of the offender (if known), the damage, and the victim.  Higher rates of reporting crimes to police could potentially mean that victims possess higher levels of confidence in the police. However, crimes go unreported for many other reasons, such as crimes being dealt with in another way, the crime was not important enough to report, or simply that the victim did not want to deal with the law enforcement personnel (Moore and Braga, 2003).  Further, since not all crime is reported to the police (Moore and Braga, 2003), the UCR might not be a perfect measure of crime and victimization rates in a community. Victimization surveys, such as the Victimization cycle of the GSS in Canada, can be used to complete the portrait of crime in a community since they capture the crime that is not reported to the police. Further, even if police effectiveness could be accurately measured by the crime rate, a lowering of that rate would not necessarily translate into public confidence in police effectiveness since the confidence is shaped by other factors along with reducing crime rate.
   
   
2. Call offenders to account. This dimension of police work is related to the first, in that they are both achieved through actions of the police with the ultimate goal of controlling crime and reducing victimization. However, an important difference between the two is that for many members of the community, justice may be seen as served when those who are guilty of a crime are brought to justice by being apprehended, fined, and/or jailed. In other words, the idea here is that criminals need to be punished for their actions. Yet to other members of a community, it may seem unjust that individuals are locked up for a long time for actions that they may commit in future. It should further be noted that modern-day policing strategies may combat crime without necessarily calling offenders to account, through reducing criminal opportunities or disrupting criminal enterprises. An example of such a strategy includes preventative policing where police educate the community to lock their cars and homes or not drive under the influence of drugs or alcohol, which in turn, reduces crime and victimization but does not necessarily result in additional arrests.
   
   The dimension of calling offenders to account has usually been measured through the direct measure of clearance rates. The measure is supposed to assess how effective the police are at dealing with crime by arresting offenders and clearing cases. Aside from the methodological troubles associated with clearance rates that are described in the previous section, Moore and Braga (2003) point out that oftentimes, it is not enough for the police to think that they have solved a crime if the rest of the judicial system – prosecutors, courts, and juries – disagrees. In other words, clearance rates may be telling only a part of the story with regards to police effectiveness in battling crime and calling offenders to account. A more complete picture could arise if one looks into the quality of police investigations and arrests (Moore and Braga, 2003: 42), where three factors are taken into account: the level of professional expertise with which police conduct investigations and make arrests; the degree to which the investigations and arrests can sustain legal scrutiny; and the extent to which the work of police can result in actual convictions. Moreover, calling offenders to account could also mean issuing a fine or a warning for certain types of offences rather than resorting to an arrest.
   
   While developing such measures of performance is not a straightforward task, there is evidence from research done in the U.S. that increasing police’s professional expertise and the quality of investigations does translate into higher levels of conviction (Eck, 1992). Ultimately, it is up to police jurisdictions to make the effort of including such performance measures when assessing their performance. They can be included as part of their routine evaluation of services, or as case studies of their investigative and arresting procedures.
   
   
3. Reduce fear and enhance personal security. This dimension of police work is somewhat controversial because it would seem that if police are successful in achieving good results on the first two dimensions, common sense would dictate that the fear of crime should decline and sense of security should increase in a community. However, past research has shown that fear of crime in a community is not necessarily explained by actual rates of crime (Moore and Trajanowicz, 1988; Skogan and Hartnett, 1997). Fear of crime can result from perceptions of social disorder in neighbourhoods (e.g., public drunkenness, open prostitution, and noisy crowds of people). Further, crime images and stories portrayed by the media can contribute at least as much to people’s sense of fear regarding crime in their communities, as the actual risk of victimization in their neighborhoods.  Still, reducing fear of crime is an important aspect of police work because if the police are only successful in dealing with the crime but leave citizens feeling unsafe, their job has only been partially accomplished.
   
   Unlike the previous two dimensions of policing, the measurement of reduce fear of crime and enhance personal security is not a straightforward process because of the subjective nature of this dimension. Fear of crime and sense of personal security are often associated with citizens’ subjective perceptions of their environment, which is not something the police can always influence or control. Depending on the change in the actual surroundings or a change in a person’s perceptions of their surroundings, the sense of fear and perception of security can be somewhat of a fluid construct that can change from one day to another (Moore and Braga, 2003). Moreover, a sense of fear can be experienced differently by two people living in the same neighbourhood.
   
   Still, it is important to measure the performance of the police in their efforts to reduce the fear of crime and enhance personal security. There are numerous examples of surveys that already enquire about people’s fear of crime in their neighbourhood, including walking alone at night, whether their level of fear has increased or decreased compared to last year, etc. Further, some surveys ask questions about citizens’ initiatives to protect themselves and improve their security. These measures can act as indirect measures of police performance in reducing the fear of crime when, for example, there is a noticeable shift from individual strategies of self-defence to the collective efforts of battling crime.
   
   
4. Ensure civility in public spaces. A lot of crime occurs in private spaces, although privacy is well-protected by constitutions in the Western world. The police are either invited or legally authorized in certain situations, such as calls for domestic violence, assault, or vandalism, to enter these spaces to protect citizens and to enforce the law. However, the public space is also shared by the citizens of a community who desire to feel safe while present in that space. It is the responsibility of the police to ensure the safety and well-being of citizens in the public space.
   
   Similar to the measurement of fear of crime and sense of personal security, police performance in ensuring citizens’ level of comfort and sense of security in a public space needs to be properly measured. A rate of crime that occurs in public spaces and clearance rates of such crimes are possible direct measures of this dimension. The levels of neighbourhood disorder and situational and observational studies of crimes in public spaces, albeit not common, are some possible indirect measures. Finally, citizens’ perception of fear of crime and their personal sense of comfort and security in public spaces could be measured through surveys.
   
   
5. Use force and authority fairly, efficiently, and effectively. Police officers are given the mandate to use force and authority if necessary, to complete the objectives of their jobs. However, whether police use force fairly has been a subject of debate for decades. Police make efforts to train their officers and implement policies that minimize the use of force in the protection of public safety. There is an expectation placed on police to use authority and force in a way that is fair and only when absolutely necessary. In Canada, an example of such an approach to the use of force can be found in the RCMP’s Incident Management/Intervention Model (IMIM), which provides guidance and direction to RCMP officers regarding the appropriate use of force in certain situations.
   
   Moore and Braga (2003) identify four measurable concepts of the dimension of use of force and authority fairly, efficiently, and effectively. First, the concept of fair distribution of police protection efforts among all constituents that is based on necessity and not on favourability can be measured through a rigorous examination of the ways in which police allocate their resources. Second, the degree of distribution of the inconvenience caused by police efforts to battle crime can be measured through revisions of police service policies with regard to fair distribution of efforts, as well as mechanisms of receiving complaints from the public about police. Third, the extent to which police actually avoid excessive use of force and authority can be measured though actual public complaints about excessive use of force, as well as successful and unsuccessful civic lawsuits against the police. Review of the number of unnecessary arrests and the frequency of excessive use of force associated with the arrests can be helpful measurements too.  Further, surveys of citizens who have had contact with the police could tease out citizens’ experiences about the levels of use of force and authority. Fourth, the concept of overall police legitimacy and support for police among the members of the community can be measured through surveys where people are asked about their overall support for police, as well as perceptions pertaining to police fairness and legitimacy in certain situations. This approach is more useful when appropriate controls are included in the analysis, such as whether the respondent had previous contact with police or has been a victim of crime.
   
   In addition to Moore and Braga’s suggestions for measuring the level of police use of force, police could record the suspect’s levels of force or violence and the police response to it during the interaction. Such information could then be compared against the police service’s policies on use of force, operational data, or public complaints against the police. While this data could provide some valuable insights into the police performance on use of force, there is a danger of inconsistent recording of a suspect’s use of force data by the police due to the subjective and perceptional nature of the issue, which would introduce potential bias in the analysis.
   
   
6. Use financial resources fairly, efficiently, and effectively. This dimension relates indirectly to police work in that the public expects the police to do their work in an efficient and economically sustainable manner. Just as in any other public or private organization, waste of resources within police departments is usually met with public disapproval.
   
   The measurement of police performance in this dimension receives the least amount of attention because it is not seen as a primary output of police work. Further, the measurements are not straightforward because it can be complicated to set benchmarks for what constitutes efficient spending. For example, large police jurisdictions could be spending more in all areas of police work because they deal with larger or more complex problems.  One approach to measure police performance in this area is to look into the innovative approaches that police take to spending: new economical ways to train officers, use of technology, innovative scheduling methods that reduce the amount of overtime payments, etc.^Footnote4
   
   There are various police expenditure surveys as well as administrative surveys performed by police and external bodies that track police spending. Further, most, if not all, police forces collect data on budget spending and expenses on human resources. Different analyses could be performed with such data: money spent per case cleared, money spent per call for service, money spent per 100,000 population, etc.
   
   
7. Quality services/customer satisfaction. The last dimension of police work has to do with the courtesy with which the police perform their duties. On the one hand, it could be argued that it should not be expected of the police to be polite or courteous to the criminals they arrest. On the other hand, the public might not grant much legitimacy to the police if the latter are known to be abrasive. Further, it is expected of the police to be courteous in instances when they perform work other than battling crime and arresting criminals such as: helping a homeless person to a shelter; providing first aid to a victim; or assisting a confused senior citizen.
   
   Much like customer satisfaction with the performance of a private business or a corporation, the obvious approach to the measurement of this dimension of police work is through public opinion surveys. Surveys with those who have had contact with the police and asking them questions on their experiences is an important tool to understand whether police have been courteous and fair. These measures could further enrich the knowledge about the dimension of the use of authority and force. It is important to ensure, however, that questions are designed to find out about citizens’ experiences during the encounter with the police and not to provide a venue for public complaints against the police.

Other than the above-discussed approaches to the measurement of police performance in Moore and Braga’s (2003) seven dimensions of police work, it is possible, and often  advisable, to supplement them with indirect measures through public opinion surveys. These surveys would assess citizens’ favourable views, confidence, and/or trust in the work that police do.

Police Performance Measurement through Public Opinion Polling

The public’s level of satisfaction with the police is a complex concept that is often difficult to quantify and gathering data can be costly (Clarke, 2002; Castle, 2008; Ganjavi, 2000; Gallagher et al, 2001; Gomes, 2007; Marx, 1976; Moore and Braga, 2003). Still, it is extremely important to capture the public’s perceptions about the performance of the police through surveys and track it over time periods. These performance measures can be used on their own in some cases, or they can be combined with other direct measures, such as crime and clearance rates (Ganjavi, 2000).

According to the study described in Ganjavi (2000), the overall quality of police service was rated as most important to the public, over other specific measures, such as the level of crime and the visibility of police in the neighbourhood. Moreover, there is a symbiotic relationship between the public and the police with respect to how effective the police are, or are deemed to be: “police depend heavily on the public to provide vital information about criminal or suspicious activities and to serve as witnesses in trials, both of which are predicated upon positive police-community relations” (Chow, 2011: 649).

Chow’s study, which focussed on the perceptions of police by adolescents, recommended more research on a wider population beyond adolescents in one Canadian city. It was further recommended that more variables, such as “neighbourhood quality” and fear of crime, be included and a longitudinal design be adapted to ascertain causality between positive views of the police and citizen cooperation with the police.  Interviews or focus groups might also qualitatively explore factors affecting the public’s opinion of the police (Chow, 2011). These actions might help to determine which public opinion indicators would best measure police performance.

In a 2005 study, a series of variables was used to measure Chinese immigrants’ perceptions of the Toronto Police Service (Chu and Huey-Long Song, 2008). Dependent variables such as perceived police prejudice, police effectiveness and level of respect for police were measured against independent variables, such as fear of crime and previous contact with police, and control variables, such as age, gender and household income. The authors concluded that generalization about the relationship among these variables is limited, based on sampling size, and that more research would be required.

In O’Connor (2008), data from the Canadian 1999 General Social Survey were analysed with respect to victimization, personal safety, attitudes toward the criminal justice system, as well as how well the police do their work (e.g., enforcing the laws, promptly responding to calls, being approachable, and ensuring citizen safety). The study’s author found that people’s perception of the police was mediated by whether or not they themselves had been victims of crime and whether or not they felt safe. If they had been victimized, they gave less favourable assessments of police. Conversely, if they felt safe, they thought more highly of the police. Appropriate weighting techniques and data path analysis were used to understand the relationship between socio-demographic and non-socio-demographic variables and attitude, toward the police. The author noted the importance of developing “ratio level measures (or in the case of satisfaction with safety, better measures) of these concepts in order to better determine the mediating effects of these variables” (O’Connor, 2008: 592). Also, the author contended that there is a need for “a more reliable measure of police contact, that would be able to distinguish between positive and negative police contact to allow a better understanding of police contact’s role in determining citizen attitudes toward the police” (O’Connor, 2008: 592). Other findings from this study were that no demographic variables were predictive of overall satisfaction with the police.

A 1976 U.S. study complements these findings (Marx, 1976) by highlighting the importance of correlating citizen behaviour, perceptions of personal safety, levels of victimization and reported crime levels in order to achieve a true measure of police effectiveness. On a related note, Marx suggests that measures of the level of citizen cooperation could be developed from data such as information about serious crimes generated by the general public; damage to police property; the number of false alarms; the number of attacks on police; the number of instances where citizens assist officers; and how many people are out on the street at night. Data could be gathered by police supervisors and interviews with citizens.

These and numerous other studies point towards the importance of measuring the public perceptions of police. Knowing how the public feels about the police can help maintain public order, increase community-police cooperation, develop appropriate policing policies and practices, as well as concentrate police efforts where they are needed the most (Ashcroft et al, 2003).

General Questions on Police Performance

The most popular police performance measure that assesses the public’s overall satisfaction with the police comes from a pool of survey questions that ask respondents whether they hold “favourable views” of the police; “approve of” the police; “respect” the police; are “satisfied with” the police; have “confidence in” the police; or “trust” the local, community, municipal, provincial/state police.  In practice, asking these kinds of generalized questions is supposed to be the simplest and quickest way to measure the overall level of satisfaction of citizens with the police. These questions may be classified as general because they are not aimed at evaluating any of the specific aspects of a police services’ complex array of work.

Before discussing the shortfalls of these generalized questions, however, it is important to mention that they hold some added value for at least three reasons. First, the generalized measures provide a quick indicator for the overall support for police among the citizens. If standardized (i.e., asked in the same way with same categories for answers), these measures can be used to compare the overall rating of police across time periods as well as in different communities (Gallagher et al, 2001). Second, the overall assessment of police in a community carries implications for the support constituents will give to police work, such as organizing neighbourhood watch programs, aiding the police with information and various cooperative initiatives, or support of the public for the funding of the police services. A lower assessment of police in a community could further increase the chances of public complaints and lawsuits against the police, lack of cooperation or, in some extreme cases, a rebellion against the police, and negative reflections in the media (Gallagher et al, 2001). Lastly, past research has shown that a decrease in the perceived legitimacy of the police could potentially lead to non-compliance with the authority of the police and increased crime rates (LaFree, 1998; Tyler, 1990).

While questions on the overall assessment of police provide some value to measuring police performance, their major shortfall is that they are too general to tease out specific information on what it is that citizens favour about the police services. More importantly, it is impossible to tease out that about which citizens are unhappy. When prompted about their overall satisfaction with the police, the respondents, in their mind, may be referring to police courtesy, effectiveness with battling crime, use of force, or any aspect of police work (Gallagher et al, 2001; Jackson and Bradford, 2010a). Respondents may further be referring to issues that may only indirectly relate to police efforts, such as social disorder in their neighbourhoods or their general fear of crime. Evidently, none of these aspects could be captured by the general question on the level of satisfaction with police. This information is necessary for the police and their regulators| to understand so that they can make necessary adjustments to their activities or organizational structure.

There are several other methodological challenges associated with general questions on the levels of satisfaction with police that should be addressed. First, the wording of survey questions and their response options are extremely important and need to be selected very carefully. Past research in the U.S., for example, has showed that a slight change in the response options in two different surveys will produce significantly different results. In 1981, the CBS News and the New York Times asked American residents about the level of confidence in the police in their community. The response options were “a great deal,” “quite a lot,” “some,” “very little,” and “no opinion.” According to the results, 53% of respondents said they had “a great deal” or “quite a lot” of confidence in the police in their community. The survey was repeated eight months later in 1982 by the Gallup Corporation, but the response categories changed to “a great deal,” “quite a lot,” “not very much,” “none at all,” and “no opinion”. This time around, 76% of respondents indicated they had “a great deal” or “quite a lot” of confidence in the police in their community (Gallagher et al, 2001). It would be wrong to conclude that based on these results, confidence in local police in the U.S. increased by 23% in the period of eight months. The more likely reason for the shift is the change in response categories; the responses to “some” category in the first poll were probably shifted upwards in the second poll because “not very much” and “not at all” were perceived as being more than “some” or “very little.”

In the previous example, both polls featured a four-point response scale with two different response options, plus the “no opinion” response option at the end. Some questions enquiring about respondents’ favourable views of their police would feature a five-point response scale that includes a mid-point, usually considered to be a neutral category. In many cases, this category would be “neither favourable, nor unfavourable,” reflecting the true neutrality of the response option. However, some questions on confidence in police would feature “some confidence” as a mid-point response option. Worse yet, some survey questions could feature “some confidence” as a response option in a four-point response as an upper tier response option. The interpretation of the answers to the category “some confidence” as a mid-point or as an upper tier response option is somewhat cumbersome since it is not a true neutral response option.  Neither could it be considered as an upper tier response option. Does the fact that that a certain, often sizeable proportion of citizens have “some confidence” in their police indicates their pessimistic or optimistic opinion of the police? More importantly, would police leaders be encouraged to report that citizens have “some confidence” in their police?

Yet another methodological challenge is found in the wording of the questions that are being asked on surveys. Questions on favourable opinions or views of the police, for example, have been shown to produce consistently higher levels of satisfaction with police than questions on the levels of confidence in the police (Gallagher et al, 2001).  Much theoretical work still needs to be done to understand the nature of this disparity. Some possible explanations of the phenomenon may be found in the conceptual nature of the two notions. These two notions are undoubtedly interconnected, yet they are somewhat distinct. The notion of “favourableness” may be tied to persons’ subjective beliefs with regards to the police, which could include a broad scope of attributes of the police such as their intentions and efforts. On the other hand, the notion of “confidence” might be steering persons responding to the question towards assessing the more objective actual police actions. In a similar vein, a patient could have a favourable impression of a doctor because of their polite manners and plausible approach, but have little confidence that the doctor will be able to cure a complicated illness (Gallagher et al, 2001).

Another important notion that is frequently featured in the wording of public opinion questions on satisfaction with police is that of “trust” of, or in, one’s police. While there is no agreement among social scientists on a common definition of trust, some prominent thinkers, like Niklas Luhman, have attempted to operationalize the concept, making it practically applicable in social research. For Luhman, the concept of “trust” is a mode of interaction among various social systems, institutions, and individuals (Luhman, 1979; 1988); it is a hopeful acceptance of reality during some sort of a vulnerable situation with a perceived positive resolution of a situation.

Trust carries an important function of decreasing complexity in a society in that by placing trust in individuals or institutions, people’s decision-making processes and the ensuing actions become much simpler (Pearson et al, 2001). If someone, for example, has trust that the police ensure safety in public places, it would make it easier to take a stroll at night. An individual’s decision to trust something or someone is based on both past experience in a similar situation and perceived dangers of accepting the situation (Luhman, 1979). In other words, knowing their experiences from the past, trust allows individuals to act in a manner in which perceived dangers are precluded from occurring (Jackson, 2009).

Luhman further makes an important distinction between the concepts of “trust” and “confidence.” The two concepts are inter-related in that they both refer to beliefs in outcomes which might prove unsatisfactory or even upsetting. However, the mechanism of “trust” involves a consideration of dangers associated with a certain circumstance (Luhman, 1988). When an individual pursues an action, despite the perceived risks and despite the possible negative outcome and subsequent disappointment, it may be said that the individual trusted. “Confidence,” on the other hand, is a mechanism through which we may act based on our beliefs only and not engage with the perceived possible dangers (Luhman, 1988). In the case of “trust,” the blame for a disappointment is attributed internally since we should know better what or who to trust; in the case of “confidence,” this blame is placed on external circumstances (Meyer and Ward, 2009).

Comparing the concepts of “confidence” and “trust” in the field of measurement of levels of satisfaction with police, it may then be argued that “confidence” summarizes people’s sentiments or attitudes towards the police to perform the duties assigned to them. It is something people have towards the police as a societal institution.  Alternatively, people “trust” their police, and it is something they do rather than have (Bradford and Jackson, 2010; Jackson 2009; Stanko et al, 2012). In other words, whereas “confidence” is largely “a passive regard of the police by the public,” (Stanko et al, 2012: 321) trust may be understood as something that “is embedded in the relationship between [the police and the public], and is grounded in the interpersonal relationships people have (or imagine they have) with the police” (Stanko et al, 2012: 321).

In research on police performance measurement, “trust” is often mistakenly used as a substitute for the “favourable views of” and “confidence in” police. Often, these three concepts are used interchangeably, all referring to sentiments of satisfaction with police, police legitimacy, confidence and trust in police, etc. (Jackson, 2009).  While related to favourable views of and confidence in police, it must be underlined that trust is a much more powerful concept that measures somewhat different perceptions and attitudes. It is an active, not passive, mechanism through which people regard their police. To build upon the previous example of a patient-doctor relationship, a patient could have a favourable impression of a doctor because of their plausible approach. The patient may also have some level of confidence that the doctor will be able to look into and cure the patient’s illness. But if matters come to a surgical intervention, the patient may not trust their life to the same doctor because of past experiences they may have had and the risks associated with a complicated surgery.

It is possible to speculate, then, that due to the more active and powerful nature of the concept of “trust,” responses to questions featuring “trust” in the police would generate even lower ratings for satisfaction with police than responses to questions on “favourable views” of and “confidence” in the police. However, more research is needed to confirm such a hypothesis.

The final methodological challenge of asking the public about their level of favourable views, confidence or trust in the police is the idea of “forcing” the respondent to state their opinion on a matter with which they might have little familiarity, thorough understanding, or simply no opinion. Although this methodological limitation is a recurrent issue in the field of public opinion research, its effects might be amplified with the more serious nature of the topic of confidence or trust in the police. Respondents might provide opinions that are not a true reflection of their real opinion when they are “pressured” to respond. One of the ways suggested in the literature to remedy this limitation is to replace the standard “no opinion” response option with a softer, more inviting “…or haven’t you thought about this recently?” (Gallagher et al, 2001). Another option is to introduce a screening question enquiring how much, and to what extent, the respondent has thought lately about police performance. Asking such a question before inquiring about the respondent’s opinion on confidence or trust in the police will allow the researcher to better classify the opinions of respondents during the analysis of the  poll results (Gallagher et al, 2001).

Towards Specific Questions on Police Performance

It is clear that general questions on police performance will provide nothing more than vague ideas regarding how citizens feel about the police. The more general the question, the more positive responses tend to be (Gallagher et al, 2001), as shown in the example of “favourable views of” and “confidence in” the police. What is mainly lacking in the general questions, however, is the understanding of specifically why citizens are happy or unhappy when it comes to police performance.

There are many examples of specific performance questions that are being asked of police in contemporary public opinion surveys in Canada and the Western World (see the Appendix for examples of questions). All examples, to a certain degree, measure police performance in the seven dimension of police work. Unfortunately, research in the area of validity and reliability of public opinion questions on police performance is lacking (Maguire and Johnson, 2010). This perhaps should not come as a surprise given the fact that police performance is habitually measured using mostly traditional performance measures of crime and clearance rates, number of arrests and response time, occasionally supplementing them with generic questions on confidence in the police (Coleman, 2012). It is only in the last two decades that researchers have begun looking into more specific police performance measures through public opinion surveys. However, groundwork on the meaning of the questions, whether they measure what they are supposed to measure, and how well they measure it, is largely absent from the literature.

There are three notable studies that attempt to conceptualize and test specific measures of police performance measured through public opinion polling: Maguire and Johnson (2010) in the U.S. who derived questions from theoretical constructs and tested them against Mastrofski’s six dimensions of policing as a service industry;^Footnote5 Coleman (2012) in Canada who consulted policing stakeholders in order to derive and test the factors that may constitute police performance; and Jackson et al (2011) in the U.K. and Europe who derived, tested and implemented European indicators of trust in judicial systems.

Deriving Questions from Theoretical Constructs

According to Maguire and Johnson (2010), research on the measurement of citizen perception of police performance takes roots in four areas: 1) service quality; 2) justice; 3) the concept of the public’s overall satisfaction with their government, agencies, and public service; and 4) the legitimacy.

Researchers from the first area, service quality, developed standardised measurements that assess the quality of services across traditional service industries (Parasuraman et al., 1988). Some researchers applied these measures to the policing sector in an attempt to measure the performance of the police (Donnelly et al., 2006; Reisig and Chandek, 2001; Webb, 1998). While the police sector shares some characteristics with certain aspects of purely customer service industries, it is different in many ways. Unlike most customer service industries, police have: government-sanctioned permission to use force, including lethal force; a monopoly of the market in which their “business” is conducted; and deal with involuntary “customers.” Still, measuring the quality of service that citizens receive from the police could prove to be a helpful indicator of police performance.

The second area from which the research on perceptions of police performance originates is that of justice. Greenberg (1987), a notable justice theorist, divided the notion of justice into three measurable categories: 1) procedural justice – which describes whether the citizens perceive a just and fair process of justice leading to a particular outcome; 2) interactional justice – which describes whether the interaction between a citizen and a representative of justice is just and fair; and 3) distributive justice – which is whether the outcome of the process of justice is just and fair (Maguire and Johnson, 2010). Since police play a key role in the administration of justice in society, it would be important to measure their performance in this area. A few notable research studies have examined the effects that perceptions of justice have on obedience to the law and compliance with legal authorities (Reisig and Lloyd, 2009; Reisig et al, 2007; Sunshine and Tyler, 2003; Tyler and Huo, 2002).

The third source of origins for research on perceptions of police performance comes from the field of citizens’ satisfaction with government, its agencies, and public service. Research in this area is distinct from the two other areas in that it looks into the overall diffuse notion of citizens’ satisfaction, taking into account the fact that many, if not most, citizens do not have contact with the government or its agencies, including police services (Maguire and Johnson, 2010). It, therefore, assesses the generalised feelings that citizens have about the authorities or symbolic images associated with institutions. In comparison, the areas of service quality and justice usually study the feelings of citizens that have had prior direct contact with a particular service industry or agency of justice. Numerous questions, in different formats, regarding satisfaction with government and police are already being asked in various public opinion surveys (Gallagher et al, 2001).

Finally, the area of legitimacy of police is the fourth source of origin for research on perceptions on police performance. Legitimacy may be seen as a characteristic of a social institution that will either encourage or discourage the public to obey its authority (Tyler and Huo, 2002). When it comes to police, Skogan and Frydl (2004) point out that the legitimacy of the police may be constructed or damaged by the conduct of the police. When the police lose legitimacy in the eyes of the public, non-compliance with the law and opposition to police authority, and even hostility, are likely to occur.  If this is true, it would be very important to measure the public’s perception of the legitimacy of the police. Questions on citizens’ opinions regarding obeying the law, their scepticism about the moral authority of the government agencies and the police to act, and their overall support for these institutions are aimed at measuring the concept of legitimacy (Maguire and Johnson, 2010).

Maguire and Johnson (2010) attempted to combine the theory and research from the four backgrounds and create unique indicators of citizens’ perceptions of police. They combined the four backgrounds with a purpose to test Mastrofski’s (1999) six-dimensional model of the service aspects of policing, resulting in 27 indicators of service quality of, and satisfaction with, the police^Footnote6. The survey was then administered through mail to a small suburban community in Virginia, U.S. A total of 374 questionnaires was sent out, and 138 were returned completed. A method of confirmatory factor analysis was applied to the data with a goal of determining whether Mastrofski’s six dimensions were supported by unique factors; results indicated that they were not. Maguire and Johnson speculated that the unique factors were not supported for three reasons. First, the community where the questionnaires were administered was fairly well-off and homogenous, with low rates of crime and disorder. Second, the level of citizen-police contact that respondents indicated on surveys was not frequent and was of low intensity; this could potentially skew the depth of opinions that citizens have about the police. Third, the sampling strategy that was used in the study was of a hybrid nature, where the authors tried to capture recent contacts with the police; however, broad questions that surpass the recent contact with police were asked, which might have taken away from the specificity sought by the authors. Despite the lack of support for unique factors, however, this research is one of the very few examples of scholarly work attempting to empirically test indicators of citizens’ perceptions of police against a theoretical model.

Deriving Questions through Engagement with Policing Stakeholders

While the work of Maguire and Johnson (2010) focused on deriving police performance measures from a solid theoretical foundation, the approach taken by Coleman (2012) focused on exploring the views of policing stakeholders on the measurement of police performance. By policing stakeholders, Coleman is referring to “police leaders, elected persons of municipal governments, members of police governance authorities, senior public servants of municipal and provincial governments, and members of the public” (Coleman, 2012: 81). With an objective to contribute to a largely absent academic work on the design and implementation of contemporary police performance measurement, Coleman’s work intended to explore and build a unified model for subsequent studies that would look into police performance measurement.

Taking a deductive quantitative approach, whereby a few measurable factors are created  from multiple theoretical constructs, Coleman (2012) designed an exploratory survey consisting of 21 questions and multiple sub-questions that were derived from the literature. A total of 4,285 identified policing stakeholders were invited to complete the questionnaire online in Canada and 560 responses were received. The responses were analysed using a method of exploratory factor analysis and principal component analysis, as well as other methods that tend to be used when a researcher wants to create factors or composite indices, such as parallel analysis and reliability-internal consistency analysis.

The results of Coleman’s analysis revealed seven factors of police performance measurement (in order of relevance of factors):

1. Legitimacy of the Local Police Agency;
2. Leadership and Management of the Local Police Agency;
3. Enforcement by the Local Police Agency;
4. The Community’s Feeling of Safety;
5. Crime and Social Disorder in the Community;
6. Misconduct of Local Police Personnel; and
7. Mutually Beneficial Police/Community Relationships. 

Coleman’s (2012) work is instrumental in helping police leaders set up and improve the existing police performance frameworks in Canada and around the world. It was not Coleman’s intent to construct actual survey questions on police performance. Neither was the goal of the work to assess the views of the general public, but rather that of a group of law enforcement experts and representatives of various levels of government. Nevertheless, his work is an important contribution to the pool of knowledge of police performance measured through public opinion. The factors identified in the analysis could potentially form a solid basis for the construction of questions that could be posed to the general public on public opinion surveys.

Deriving Questions through Engagement of Multidisciplinary Panel of Experts

Taking a slightly different conceptual approach, Jonathan Jackson and his colleagues’ work in the U.K. and Europe was instrumental in the creation and application of round five of the European Social Survey (ESS) (Jackson et al, 2011). This round was constructed by the authors and implemented across Europe with a goal to collect standardized and comparable data on social indicators of trust in the judicial system, including the police, across Europe.

Jackson and colleagues’ conceptual model behind the creation of the indicators is rooted in the notions of trust and legitimacy. Trust in police and the wider justice system, the authors argued, is a concrete and measurable notion. They deconstructed the notion of trust from the general idea, such as the hopeful acceptance of some sort of a vulnerable circumstance with a perceived positive resolution of this circumstance, to a narrowed notion that the societal belief that the police have the right understanding and intentions towards the citizens they are policing, and that they “are competent to act in specific ways in specific situations” (Jackson et al, 2011: 270). Specifying further, “trust in the police means believing that officers have appropriate motives and are technically competent (in the roles assigned to them within social relationships and systems) to carry out their fiduciary obligations (that is, in certain situations place the interests of others above their own)” (Jackson et al, 2011: 270).

Jackson and colleagues continued deconstructing the notion of trust after the implementation of round five of the ESS. Inspired by the work of Tyler and Hou (2002) on motive-based trust, and by Stoutland’s (2001: 233) qualitative research on trust in police among residents of Boston, MA, Jackson and colleagues go further and dissect the concept of trust into four distinct dimensions: (1) Trust 1: priorities; (2) Trust 2: competence; (3) Trust 3: dependability; and (4) Trust 4: respectfulness (Stanko et al, 2012: 321-322):

• Trust 1: Do they (current or potential allies) share our priorities or motives?  Can we trust the police to share our priorities? Can we trust them to care about our concerns as they plan and implement policies to control crime in our neighbourhood?
• Trust 2: Are they competent? Do they have the knowledge and skills to play the roles required of them?  Do they understand the issues?  Can we trust the police to have the knowledge and skills to effectively and consistently enforce the law, control crime and maintain high levels of safety?
• Trust 3: Are they dependable in fulfilling their responsibilities? Even if they have the best of motives, do they have the necessary resources to follow through on their promises? Can we trust the police to have sufficient resources to maintain low levels of crime and high levels of safety?
• Trust 4: Are they respectful? Can we trust the police to be respectful, courteous and fair in their interactions with us?’

Referring back to their work on creating European indicators on trust and legitimacy, Jackson and colleagues deconstructed the notion of legitimacy as well. In general terms, they defined the notion of legitimacy as an institution’s “right to govern and the recognition by the gov­erned of that right” (Jackson et al, 2011: 271; see also Jackson and Bradford, 2010b). Narrowing down to the legitimacy of the judicial system and the police, legitimacy is rooted in citizens’ acknowledgement “of the criminal justice system’s right to exist, and in the justification of its authority in determining the law, governing through the use of coercive force, and punishing those who act illegally” (Jackson et al, 2011: 271; see also Jackson and Bradford, 2010b). It may be said that police are seen as legitimate in the eyes of the citizens when: (a) they feel obliged to follow police authority and the law; (b) they feel the police shares their morals; and (c) they feel that the police (and the judicial system) abides by its own rules.

Following much of the theoretical and conceptual effort, Jackson and colleagues (2011) defined trust as “the belief that the police and courts have the right intentions and are competent to do what citizens trust them to do” (Jackson et al, 2011: 273). They defined legitimacy as “(a) expressed consent, (b) normative justifiability of power and (c) legality of action” (Jackson et al, 2011: 273). The rather complex and cumbersome process of turning these conceptual definitions into actual questions on the ESS entailed thirteen stages. Some of the stages included: cognitive interviews; proposals for questions from experts in the relevant fields; a review of questions by a panel of multi-disciplinary panel of experts; assessments of questions using software called Survey Quality Predictor Program (SQP) and aimed at assessing the validity and reliability of the questions; and another revision of the questions by the designers. The stages were then repeated once again to ensure the iterative approach to designing new items for questionnaire design. Once the wording of the questions was settled, they were tested in two pilot surveys administered to a sample of respondents in the U.K. and Bulgaria. Results of the pilot tests were thoroughly analysed for non-response items, the extent of scalability, the structure of factors, correlations, etc. The results of the analysis were once again referred to the panel of experts for final revisions.

The process of development, testing, and implementation of questions by Jackson and colleagues is a good example of how to turn theoretical constructs into measurable indicators of the public’s perception of police performance, or, put simply, how to develop good survey questions. The items that ended up on the ESS are valid and reliable measures, assuring the researchers who are using the data gathered from the survey that it measures what it is supposed to measure (validity), and measures it well (reliability). Because the indicators were developed with a purpose to compare citizens’ perceptions of the judicial systems in European countries, with some of the Eastern European countries that may still have corrupted or unreliable judicial systems, not all the questions are applicable in the Canadian context. For example, it is probably not very useful to ask questions on police corruption in Canada, such as “The decisions and actions of the police are unduly influenced by pressure from political parties and politicians” (agreement scale) or “How often would you say that the police in [country] take bribes?” However, questions on police legitimacy such as “To what extent is it your duty to… (1) back the decisions made by the police even when you disagree with them? (2) do what the police tell you even if you don’t understand or agree with the reasons? (3) do what the police tell you to do, even if you don’t like how they treat you?” could be applied in Canada.^Footnote7

Public Opinion Polling on Police Performance in Canada

Statistics Canada conducts the GSS on an annual basis, with cycles repeated every five years. Six questions on satisfaction with police in Canada are already asked on the Victimization cycle of the GSS^Footnote8 and repeated every five years, making them comparable across time. While these six questions provide a valuable and much needed standardized measure of Canadians’ perception of the performance of their local police, they do not and cannot measure the performance of the police in all of the dimensions of police work. These six questions are currently the only comparable tool that provides an assessment of the public’s perception of police performance in Canada.

Other than the six questions on the GSS, the majority of municipal police services commission their own annual or bi-annual public opinion community surveys that include numerous general and specific police performance questions.  Questions asked on surveys on multiple municipal police forces in Canada were reviewed at the time of writing this paper and can be viewed in the Appendix.  Even a quick overview of the questions reveals that most of the questions that are supposed to measure similar concepts have different wording of the questions and the categorization, as well as different scales for responses (i.e., four- versus five-point scales).

For example, consider the general questions that are being asked on the overall satisfaction with police in seven Canadian cities, one provincial police service, and the RCMP.^Footnote9 These questions are measuring the same concept of general satisfaction with police. They all feature a word “satisfaction” or, in some rare cases, “confidence” in the question, but the wording is different for all of them. The wording of the categorization of responses and their scale are different as well, ranging from three to five response categories.  It is simply impossible to compare the answers to these questions because of different wording and categorizations. In other words, the level of general satisfaction with police cannot be compared in areas covered by these police services based on the responses to these questions.

This circumstance becomes even more complicated when it comes to specific questions on police performance because the concepts that need to be measured are numerous and more complex. Much like in the example on the general question of satisfaction with police, a lot of specific questions are similar to each other in that they all measure similar concepts. However, the wording of the questions and the response categories, as well as the scale of the response categories are different across all of the questions, making them incomparable to each other and to the national average. For example, in one case the response category “somewhat [well]” is treated as a mid-point, whereas in other cases “somewhat [satisfied] [confident]” is treated as an upper-tier response category. Further, while some odd-numbered scale questions have a true mid-point in the “neither…nor” style, a mid-point in some other cases is treated as “don’t know” category and is only recorded if such response is prompted by the respondent.

The word “confidence” rarely appears in the wording of the general police performance questions in Canadian surveys. Instead, the questions concentrate around the notion of “satisfaction” with police. When it comes to specific questions, the wording of most questions enquires about either the level of “satisfaction with,” or prompts a level of agreement or disagreement to specific statements describing the activities of police work. A few questions ask the respondents whether they think their police are doing a “good job” or “bad job” in certain areas of police work. The concept of “trust” only appears once in the question asked by the RCMP, where it is used interchangeably with “confidence.”^Footnote10

Practically all the police jurisdictions in Canada ask their own specific questions on perceptions of police performance. The questions could have been designed by the employees of the police services or by the polling firms responsible for the survey. There is a great deal of discrepancy among the specific police performance questions that are currently being asked; all the questions reviewed have different wording and categorizations. Interestingly, only two Canadian police services among those reviewed – Regina Police Service and Edmonton Police Service – asked the same block of questions on satisfaction with police service that are asked on a national survey administered every five years by Statistics Canada, the victimization cycle of the GSS. This allowed researchers responsible for the survey to compare the perceptions of residents of Regina and Edmonton of their local police to the national figures. No other comparisons can be made because of the disparity in the questions and the response categories.

Suggestions for Further Research

Over a decade ago, a thorough overview of the public image of police in the U.S. was conducted by a group of academics (Gallagher et al, 2001). In it, the authors called for the development of a national standardized research instrument called Uniform Public Opinion Poll on Policing (UPOPP). The process of development for such an instrument would cover the U.S. using standardized questions on public opinion of police that would be comparable across states and over time. Unfortunately, the project never materialized due to funding issues. However, if the project was to be implemented, the process of developing indicators to be used on the survey would have probably gone through a similar rigorous process of consultations and validations, such as the one undertaken by Jackson and colleagues in Europe (2011).

Since fielding public opinion surveys tends to be expensive, with each additional question adding to the overall cost of the survey, a good cost-efficient approach would be to use fewer but more precise measures. Single measures of public perceptions of police performance, such as the ones discussed by Maguire and Johnson (2010) or developed by Johnson and colleagues (2011), are good examples of such an approach. To derive these measures, confirmatory factor analysis could be applied to the existing data gathered by police services. The data would need to consist of multiple specific questions that, at least on the surface, measure similar concepts. Similar to the approach taken by Maguire and Johnson (2010), confirmatory factor analysis would then be used to group the specific questions into factors, which would aid the researcher in deriving the single measures that would be applied.

Another, more involved approach to developing and testing indicators of police performance measured through polling could be the one proposed by Gallagher et al. (2001) or the one undertaken by Jackson et al (2011). It might begin with focus groups that would ask citizens in a community what satisfaction, confidence, or trust in the police mean to them. Rich data from these focus groups could then be reviewed by a panel of experts from law enforcement agencies, academia, and/or government with a goal of constructing specific questions on police performance. It should be noted that this consultative practice exists among some police jurisdictions in Canada; some jurisdictions, in their attempts to set up or improve their existing performance measurement framework, consult their peers through ad hoc peer-to-peer committees within the policing community (Kiedrowski et al, 2013).

Alternatively, existing questions (examples of which are found in the Appendix) might be referred to the panel for revisions and recommendations. What could follow is a process similar to the one undertaken by Jackson et al (2011). Such a project could involve testing the questions with SQP software for validity and reliability and referring them back to a panel of experts for further revisions.  Once established, the questions could then be tested in a pilot study, the results of which could be thoroughly analysed for non-response items, the extent of scalability, the structure of factors, correlations, etc. The results of the analysis could then be referred back to a panel of experts for final revisions, following which the final questions could be constructed.

Following all the iterative steps, the indicators could become part of a national survey in Canada; one option of an existing survey where these questions could be asked is the General Social Survey (GSS) administered by Statistics Canada.^Footnote11 Alternatively, a polling research protocol could be established for policing partners, which could lead to subsequent syntheses or meta-analyses of findings to be undertaken. Individual police jurisdictions would also have a pool of valid and reliable questions that they could use on their own community surveys.  This would allow them to get better measures of how their constituents feel about their police. While this could be a lengthy and costly process, it would result in solid, standardized, and comparable police performance indicators measured through the surveys of Canadians.

Conclusion

Police are responsible for, and involved in, many activities, only a minority of which deal with serious crime (Marx, 1976; Statistics Canada, 2014). In fact, public calls to police show that the majority of calls do not relate to crime at all, but rather to order maintenance, information and service-related activities (Ericson, 1982). Police services have increasingly been asked to be more accountable and reliable, which in turn, has created the need for quantifiable performance measures (Nuchia, 1993). A good and reliable framework for the measurement of police performance could aid in the overall assessment of the quality and effectiveness of public policing as it would both assess the community’s needs and priorities, as well as the performance of the police in satisfying these needs. Such a performance measurement framework would enable: citizens to evaluate the police service; police managers to better able to guide their organizations to provide maximum value to citizens; and police officers to know what is expected of them (Moore, 2002).

This report focuses on one area of police performance measurement – public opinion polling. It is argued that measuring citizens’ perceptions of police performance is very important and is not a straightforward task. In as much as police work is complex and multi-dimensional, so is its performance measurement in the eyes of the public. Simply put, there is no one-size-fits-all, off-the-shelf performance measurement tool or framework (Castle, 2008; Coleman, 2012). While somewhat valuable, the general questions on “favourable views,” “confidence,” or “trust” provide nothing more than a general sense of the public’s satisfaction with the police. More specific questions need to be asked in order to understand what it is that the citizens are satisfied or dissatisfied with when it comes to the police service.

A large number of specific police performance questions are already being asked in numerous public opinion surveys. However, the validity and reliability of these questions, for the most part, have not been tested and are of questionable value. Moreover, the existing questions are not standardized, making them impossible to compare across time and place. In Canada, for example, only two police jurisdictions are asking the same questions that are asked on GSS, making them comparable to each other and to the national average, as well as across time periods. The rest of the public opinion surveys on satisfaction with police in Canadian municipalities, found online and included in this study, had varying questions that are not comparable between different police jurisdictions or to the national average.

More research into either testing the existing questions or creating new standardized questions is required to improve the measurement of public satisfaction with the services provided by the police in Canada. A project similar to the creation of European indicators of trust in justice (Jackson et al, 2011) or the UPOPP (Gallagher et al, 2001) is one way to provide solid data for police performance measurement. Such an initiative could provide much needed standardized indicators of perceptions of the police and justice. While this might represent a significant investment of resources, these indicators could finally be proven as valid and reliable, as well as comparable across the Canadian municipalities and provinces and across time periods (Lithopoulos and Rigakos, 2005).  In the absence of standardized indicators, police services and policy makers will continue to ask different questions that are not comparable across place and time.  Such an inconsistent measurement of police performance runs the risk of evaluating the performance or police and policing policies and practices rather poorly across Canada, leading to inefficient and ineffective policing and ultimately compromising the public safety of Canadians. 

Bibliography

ABC News (2001). “Testing police honesty: A PrimeTime investigation with lost wallets” PrimeTime Thursday, May 17th [television show].

Ashcroft, J., Daniels, D. J., and Hart, S. V. (2003) “Factors that Influence the Public Opinion of the Police” Research for Practice, June 2003. National Institute of Justice, U.S. Department of Justice.  Available at https://www.ncjrs.gov/pdffiles1/nij/197925.pdf (accessed on December 12, 2013)

Bradford, B. and Jackson, J. (2010) “Trust and Confidence in the Police: A Conceptual Review” Available at SSRN: http://ssrn.com/abstract=1684508 (accessed on November 8, 2013).

Brennan, R. J. (14 March, 2014) “Municipalities pushing back against OPP costs.” The Toronto Star. Retrieved from http://www.thestar.com/news/queenspark/2014/03/14/municipalities_pushing_back_against_opp_costs.html (accessed on April 14, 2014).

Castle, A. (2008). “Measuring the impact of law enforcement on organized crime” Trends in Organized Crime,11: 135-156.

CBC (June 24, 2013) “Eastern Ontario communities struggle with police costs.” CBC. Retrieved from http://www.cbc.ca/news/canada/ottawa/eastern-ontario-communities-struggle-with-police-costs-1.1308870 (accessed April 14, 2014).

Chow, H. P.H. (2011) “Adolescent attitudes toward the police in a western Canadian city” Policing: An International Journal of Police Strategies & Management, 34(4): 638-653.

Chu, D. C. and Huey-Long Song, J. (2008) “Chinese immigrants’ perceptions of the police in Toronto, Canada” Policing: An International Journal of Police Strategies & Management, 31(4): 610-630.

Clarke, C. A. (2002) “Between a rock and a hard place: RCMP organizational change” Policing: An International Journal of Police Strategies & Management, 25(1): 14-31.

Coleman, T. G. (2012) “A Model for Improving the Strategic Measurement and Management of Policing: The Police Organizational Performance Index (POPI)” PhD Dissertation, University of Regina: Saskatchewan, Canada.

Cordner, G. W. (1989) “Police agency size and investigative effectiveness” Journal of Criminal Justice, 17: 145-155.

Donnelly, M., Kerr, N.J., Rimmer, R. and Shiu, E.M. (2006) “Assessing the Quality of Police Services Using SERVQUAL” Policing: An International Journal of Police Strategies and Management, 29(1), 92-105.

Eck, J. (1992) “Criminal Investigation” in Cordner, G. and Hale, D. (eds.) What Works in Policing? (pp. 19-34).Cincinnati: Anderson.

Ericson, R. V. (1982) “Reproducing Order: A Study of Police Patrol Work” Toronto: University of Toronto Press.

Gallagher, C., Maguire, E. R., Mastrofski, S. D., and Reisig, M. D. (2001) “The Public Image of the Police” Final Report to The International Association of Chiefs of Police by the Administration of Justice Program, George Mason University.  Available at http://www.theiacp.org/PoliceServices/ProfessionalAssistance/ThePublicImageofthePolice/tabid/198/Default.aspx (accessed on October 15, 2013)

Ganjavi, O., LeBrasseur, R. and Whissell, R. (2000) “Night walking safety and overall satisfaction with police” Policing: An International Journal of Police Strategies & Management, 23(1): 22-37.

Gomes, P., Mendes, S. M., and Carvalho, J. B. (2007) “Use of Performance Measurement in the Public Sector: the case of the police service” 11th Biennial CIGAR Conference, June 14-15, 2007.

Greenberg, J. (1987) “A Taxonomy of Organizational Justice Theories” Academy of Management Review, 12 (1): 9-22.

Hatry, H.P. (1999) “Performance measurement” Washington, DC: Urban Institute

Jackson, J. (2009) “Trust and Confidence in Policing: A British Policing” Presentation given as part of BOCSAR's seminar series on 16 April 2009 at the Justice and Police Museum, Circular Quay, Sydney. Available at http://www.bocsar.nsw.gov.au/agdbasev7wr/bocsar/documents/pdf/jackson_jon_paper.pdf#xml=http://search.lawlink.nsw.gov.au/isysquery/5b700aa8-f65a-468b-a2fe-e2b2bdd3b7bc/7/hilite/ (accessed on November 8, 2013)

Jackson, J., Bradford, B., and Hough, M., Kouha, J., Stares, S., Widdop, S., Fitzgerald, R., Yordanova, M., and Galev, T. (2011) “Developing European Indicators of Trust in Justice” European Journal of Criminology, 8: 267-285.

Jackson, J., Bradford, B., and Hough, M., Kouha, J., Stares, S., Widdop, S., Fitzgerald, R., Yordanova, M., and Galev, T. (2010) “Trust in justice: Notes on the development of European social indicators” Available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1717924 (accessed on November 26, 2013)

Jackson, J., and Bradford, B. (2010a) “What is trust and confidence in the police?” Policing: a Journal of Policy and Practice, 4 (3): 241-248.

Jackson, J., and Bradford, B. (2010b) “Police Legitimacy: A Conceptual Review (September 29, 2010)” Available at SSRN: http://ssrn.com/abstract=1684507 or http://dx.doi.org/10.2139/ssrn.1684507 (accessed on November 26, 2013)

Kiedrowski, J., Petrunik, M., Macdonald, T., Melchers, R. (2013). “Canadian Police Board Views on the Use of Police Performance Metrics” Ottawa: Public safety Canada, # PS14-12/2013E. 53 pages.

LaFree, G. (1998) “Losing Legitimacy: Street Crime and the Decline of Institutions in America” Boulder, CO: Westview Perseus.

Lebowitz, M. (2010) “Socialism: the Goal, the Paths and the Compass.” Presentation of “El Socialismo no Cae del Cielo: Un Neuvo Comienzo” at the 2010 Havana Book Fair, February 2010, Socialist Project, E-Bulletin No. 135, February 20, 2010. Available at http://www.socialistproject.ca/bullet/315.pdf (accessed on February 3, 2014)

Leckie, R. (2012). “Academic Views of Police Performance Metrics: An Overview of the Literature” Ottawa: Public safety Canada. 40 pages (unpublished draft).

Lithopoulos, S. and  Rigakos., G. S. (2005) “Neo-liberalism, community, and police regionalization in Canada: A critical empirical analysis” Policing: An International Journal of Police Strategies & Management, 28(2): 337-352.

Luhmann, N. (1979) “Trust and Power: Two works by Niklas Luhmann” Brisbane: John Wiley and Sons.

Luhmann, N. (1988) “Trust: Making and Breaking Cooperative Relations” in Gambetta, D. (ed.) Familiarity, Confidence, Trust: Problems and Alternatives (pp. 94-107). New York: Basil Blackwell.

Maguire, E.R. (2003) “Measuring the Performance of Law Enforcement Agencies - Part 1 of a 2-Part article” CALEA Update Magazine, Issue 83. Available at http://www.calea.org/calea-update-magazine/issue-83/measuring-performance-law-enforcement-agencies-part-1of-2-oart-articl (accessed on October 2, 2013)

Maguire, E.R. (2004) “Measuring the Performance of Law Enforcement Agencies - Part 2 of a 2-Part article” CALEA Update Magazine, Issue 84. Available at http://www.calea.org/calea-update-magazine/issue-83/measuring-performance-law-enforcement-agencies-part-1of-2-oart-articl (accessed on October 2, 2013)

Maguire, E. R. and Johnson, D. (2010) “Measuring public perceptions of the police” Policing: An International Journal of Police Strategies & Management, 33(4): 703-730.

Maguire, E.R. and C.D. Uchida (2000) “Measurement and Explanation in the Comparative Study of American Police Organizations” in Duffee, D. (ed.) Criminal Justice 2000: Vol 4. Measurement and Analysis of Crime and Justice. Washington, DC: National Institute of Justice. Available at http://www.ncjrs.gov/criminal_justice2000/vol_4/04j.pdf (accessed on October 8, 2013)

Marx, G. T. (1976) “Alternative Measures of Police Performance” inn E. Viano, (ed.), Criminal Justice Research, United States: Lexington Books.  Available at http://web.mit.edu/gtmarx/www/alt.html (accessed on November 18, 2013).

Mastrofski, S. (1999) “Policing for people” Paper presented at the third Ideas in American Policing lecture, November 1998, Police Foundation, Washington, DC. Available at http://www.policefoundation.org/sites/g/files/g798246/f/Mastrofski%20%281999%29%20-%20Policing%20For%20People.pdf (accessed on November 6, 2013).

Mastrofski, S.D., Parks, R.B., Reiss, A.J. Jr., Worden, R.E., DeJong, C., Snipes, J.B., & Terrill, W. (1998) “Systematic Observation of Public Police: Applying Field Research Methods to Policy Issues” Washington, DC: National Institute of Justice (December).

Meyer, S. and Ward, P. R. (2009) “Reworking the sociology of trust: making a semantic distinction between trust and dependence” Paper presented at The Australian Sociological Association 2009 Annual Conference, December 1-4, 2009, The Australian National University, Canberra. Available at http://www.tasa.org.au/conferences/conferencepapers09/appliedsociology.htm  (accessed on November 18, 2013)

Moore, M. H. and Braga, A. (2003) “The “Bottom Line” of Policing: What Citizens should Value (and Measure!)” in Police Performance, Police Executive Research Forum. Available at http://www.policeforum.org/library/police-evaluation/BottomLineofPolicing.pdf (accessed on October 3, 2013).

Moore, M. H, Thacher, D., Dodge, A., and Moore, T. (2002) “Recognizing Value in Policing: the Challenge of Measuring Police Performance” Police Executive Research Forum.

Moore, M. H. and Trojanowicz, R. C. (1988) “Policing and the Fear of Crime” Perspectives on Policing, No. 3. Washington, D.C.: National Institute of Justice.

Neely, A., Gregory, M., and Platts, K. (1995) “Performance measurement system design: A literature review and research agenda" International Journal of Operations & Production Management, 15(4): 80-116.

Neyroud, P. (2008) “Past, Present and Future Performance: Lessons and Prospects for the Measurement of Police Performance” Policing, 2(3): 340-348.

Nuchia, S. (1993) “Evaluating Patrol Officer Performance under Community Policing: the Houston Experiment. Research in Brief Report.” National Institute of Justice, April 1993.

O’Connor, C. D. (2008) “Citizen Attitudes toward the Police in Canada” Policing: An International Journal of Police Strategies & Management, 31(4): 578-595.

Parasuraman, A., Zeithaml, V. and Berry, L. (1988) “SERVQUAL: A Multiple Item Scale for Measuring Consumer Perceptions of Service Quality” Journal of Retailing, 64(1): 12-40.

Paré, P. and Ouimet, M. (2003) “A Measure of Police Performance: Analyzing Police Clearance and Change Statistics.”  Revue Internationale de Criminologie et de Police Technique et Scientifique (2003): 23-42.  Available at http://crpr.icaap.org/index.php/crpr/article/view/15/14 (accessed on November 25, 2013).

Pearson, S., Crane, S. and Mont, M. C. (2005) “Persistent and Dynamic Trust: Analysis of Trust Properties and Related Impact of Trusted Platforms” in L.P. Hewlett-Packard Development Company (ed.) Trust Management (pp. 355-363). Bristol: Springer Berlin.

Riedel, M., & Jarvis, J. (1999) “The decline of arrest clearances for criminal homicide: Causes, correlates and third parties” Criminal Justice Policy Review, 9(3): 279-306.

Reisig, M.D. and Lloyd, C. (2009) “Procedural Justice, Police Legitimacy, and Helping the Police Fight Crime” Police Quarterly, 12(1): 42-62.

Reisig, M.D., Bratton, J. and Gertz, M.G. (2007) “The Construct Validity and Refinement of Process-Based Policing Measures” Criminal Justice and Behavior, 34(8): 1005-28.

Reisig, M.D. and Chandek, M. (2001) “The Effects of Expectancy Disconfirmation on Outcome Satisfaction in Police-Citizen Encounters” Policing: An International Journal of Police Strategies and Management, 24(1): 88-99.

Sampson, R.J., and Raudenbush, S. (1999) “Systematic social observation of public spaces: A new look at disorder in urban neighborhoods” American Journal of Sociology, 105: 603-651.

Sherman, L.W. (1980a) “Enforcement Workshop: Defining Arrest, Practical Consequences of Agency Differences (Part 1)” Criminal Law Bulletin, 16(4): 376-380.

Sherman, L.W. (1980b) “Enforcement Workshop: Defining Arrest, Practical Consequences of Agency Differences (Part 2)” Criminal Law Bulletin, 16(5): 468-471.

Skogan, W. and Frydl, K. (Eds) (2004) “Fairness and Effectiveness in Policing: The Evidence” National Academies Press: Washington, DC.

Skogan, W. and Hartnett. S. (1997) “Community Policing, Chicago Style” New York: Oxford University Press.

Stanko, B., Jackson, J., Bradford, B., and Hohl, K (2012) “A Golden Thread, a Presence amongst Uniforms, and a Good Deal of Data: Studying Public Confidence in the London Metropolitan Police” Policing and Society: An International Journal of Research and Policy, 22(3): 317-331.

Statistics Canada (2014) “Police Resources in Canada, 2013.”  Catalogue Number 85-002-X.  Available at http://www.statcan.gc.ca/pub/85-002-x/2014001/article/11914-eng.htm#a16 (accessed on April 14, 2014).

Statistics Canada (2011) “Neighbourhood Characteristics and the Distribution of Crime in Toronto: Additional Analysis on Youth Crime.”  Available at http://www.statcan.gc.ca/pub/85-561-m/2011022/part-partie1-eng.htm (accessed on November 25, 2013).

Statistics Canada (2009) “Measuring Crime in Canada: Introducing the Crime Severity Index and Improvements to the Uniform Crime Reporting Survey.”  Catalogue Number 85-004-X.  Available at http://www.statcan.gc.ca/pub/85-004-x/85-004-x2009001-eng.pdf (accessed on December 11, 2013).

Stoutland, S.E. (2001) “The multiple dimensions of trust in resident/police relations in Boston” Journal of Research in Crime and Delinquency, 38: 226-256.

Sunshine, J. and Tyler, T. (2003) “The Role of Procedural Justice and Legitimacy in Shaping Public Support for Policing” Law and Society Review, 37(3): 513-48.

Tyler, T. R. (1990) “Why People Obey the Law” New Haven, CT: Yale.

Tyler, T. and Huo, Y. (2002) “Trust in the Law, Encouraging Public Cooperation with the Police and Courts” Russell Sage Foundation, New York, NY.

Webb, D.A. (1998) “Segmenting Police Customers on the Basis of Their Service Quality Expectations” The Services Industries Journal, 18(1): 72-100.

Whitaker, G. P., Mastrofski, S., Ostrom, E., Parks, R. B., and Percy, S. L (1982) “Basic Issues in Police Performance” United States: Department of Justice, July, 1982.

Appendix: Selected Questions on Police Performance Asked on Public Opinion Polls

This table represents some of the questions on police performance that are asked on public opinion polls.  It is not meant to be exhaustive, but to provide the reader with the idea on wording and categorization used by the polls.  In some instances, the exact categorization of response items is not publically available.