Measuring Police Attitudes 1 POLICE DISCRETION Measuring Police Attitudes Towards Discretion

This paper describes the construction of two scales to measure police attitudes towards the selective enforcement of the law. The Service-Legalistic (S-L) scale measures police discretion along a flexible-inflexible continuum. Service-oriented police advocate the use of discretion to help solve social problems; legalistic police oppose discretion because it interferes with their duty to enforce the law equitably. The Watchman (WM) scale examines the use of discretion to maintain control. Watchman-oriented police simultaneously ignore minor offenses while calling for greater powers to deal with serious crime. Service-related discretion was found to negatively correlate with authoritarianism and the belief that crime is caused by the individual dispositions of offenders; watchman-related discretion positively correlated with authoritarianism, ethnocentrism, and a belief in individual crime causation. Measuring Police Attitudes 3 Measuring Police Attitudes Towards Discretion Of all the personnel in the criminal justice system, arguably police have the greatest opportunity to exercise discretionary judgement. Unlike judges, magistrates or parole boards, police have the ability to act as more-or-less autonomous agents. Every day, the average street-level police officer observes many offenses for which no arrest is made. Away from public scrutiny, unencumbered by due process and subject to no review an individual police officer can totally exonerate an offender by simply deciding to take some unofficial action such as issuing a caution or ignoring an offense entirely. By the selective enforcement of the penal code police have the power to pre-empt the entire course of a criminal prosecution. There are two views on the desirability of police discretion. On the one hand, discretion has been hailed as a flexible and enlightened way to deal with many social problems (De Lint, 1998; Gallagher, 1979; Kinsey & Young, 1982). It is argued that when assessing the culpability of an offender it is necessary to consider not only the illegality of the offense but also contextual and mitigating factors. Strict adherence to the letter of the law in many cases would be too harsh and justice may be better served by not introducing an offender into the criminal justice process. On the other hand, serious concern has been expressed about the dangers involved in leaving to police unchecked responsibility for deciding who is and who is not worthy of prosecution (Egger & Findlay, 1988; Goldsmith, 1990; Gottfredson & Gottfredson, 1988; Pike, 1985; Walker, 1983). Selective enforcement of the law allows police to redefine justice in terms of their own priorities, which might not correspond to the priorities of the wider community. When arrest decisions become based upon personal judgements, there is a real potential for arbitrary and discriminatory enforcement of the law. Measuring Police Attitudes 4 What all commentators agree upon is that discretion is an inevitable part of policing. Yet, despite the scope and importance of police discretion, there has been very little direct psychological examination of police decision making. This is not to say that psychologists have been uninterested in police behavior. On the contrary, a great deal of research has been devoted to examining the psychological attributes of police thought to relate to police performance. In particular, police officers have been studied in terms of a range of right-wing personality dimensions such as authoritarianism (Brown & Willis, 1985; Colman & Gorman, 1982; Perrott & Taylor, 1995; Wortley & Homel, 1995), dogmatism (Colman & Gorman, 1982; Henkel, Sheehan & Reichel, 1997; Regoli & Schrink, 1977), ethnocentrism (Teahan, 1975; Wortley & Homel, 1995), conservatism (Colman & Gorman, 1982; Cook, 1977; Dalley, 1975) and so forth. The direction this research has taken is based on the conviction that either the policing occupation attracts those who crave exercising authority over others, or that working as a police officer necessarily engenders this tendency. What ever its cause, the view that there is a definable police personality that is characterized by prejudice and intolerance of outsiders, rigid adherence to middleclass values, punitiveness towards those who violate conventional norms, and a preoccupation with power and toughness, is widespread in both popular and academic literature. The validity of the police personality has not gone unchallenged, however. For one thing, not all studies have found differences between the personalities of police and non-police (Fenster, Wiedemann & Locke, 1977; Rokeach, Miller & Snyder, 1971). In addition, in recent years many police forces have actively promoted the concepts of community policing and police professionalism, backed by higher entry standards and better training, in an attempt to change the police culture (Fielding, Measuring Police Attitudes 5 1988; Poole, 1988). To the extent that these measures have been successful, it would be expected that the classic police personality is becoming less predominant. Moreover, even where it can be shown that police do exhibit elements of a right-wing personality, the practical implications of these attributes for day-to-day police work are rarely demonstrated. It is simply assumed that the various personality scales say something meaningful about the way police exercise their authority. It is the supposed occupational relevance of the police personality that, after all, ultimately justifies the research. If the typical police officer is ethnocentric, authoritarian, dogmatic and so forth, then surely, it is reasoned, he/she also displays a discriminatory, draconian, intolerant, and otherwise undesirable approach to law enforcement. The implicit assumption in much of the writing about the police personality is that the exercise of discretion is an expression of liberalism, while the strict application of the law is a sign of authoritarianism. However, research on other areas of criminal justice decision making -notably jury deliberations and judicial sentencing -has revealed a complex relationship between the personality of the decision maker and decision preference. Certainly there is a tendency for authoritarian decision-makers to make more punitive judgements about offenders (Bray & Noble, 1978; Carroll et al., 1987; Ellison & Buckhout, 1981). But there are exceptions. Some studies have found no relationship between authoritarianism and punitiveness (Sue, Smith & Pedroza, 1975) or that the relationship holds only for some types of offenses (Wortley, 1990). Sometimes high-authoritarians have been found to be less punitive than low-authoritarians, such as when the defendant is exercising authority (Garcia & Griffitt, 1978; Mitchell in Kassin & Wrightsman, 1983), also displays an authoritarian personality (Mitchell & Byrne, 1973), or has committed a crime in the process of obeying an order (Hamilton, 1976). Measuring Police Attitudes 6 There are two implications of the inconsistent relationship between standard personality scales and decision making about offenders. First, the research challenges the utility of conceptualizing punitiveness towards offenders simply as a component of a global right-wing construct. While decisions made about offenders are related to broader personality structures they are also influenced by more fine-grained individual differences. Better results might be obtained with scales that more explicitly measure specific predispositions to criminal justice decision-making (Kaplan & Miller, 1978; Kassin & Wrightsman, 1983). Second, any such scale is itself likely to be multidimensional. Decision-making in one situation does not necessarily predict decision making in another. That is, punitiveness towards offenders does not seem to be a generalized attribute that allows decision-makers to be placed along a single continuum from strict to lenient. One strategy for investigating police decision making, then, is to examine directly police attitudes towards the utilization of discretion. A moments thought will reveal that the exercise of individual police discretion need not indicate a liberal concern by the police officer for the circumstances and needs of the suspected offender. As noted earlier when discussing the desirability of police discretion, the decision not to arrest a suspected offender might equally reflect an arbitrary and discriminatory approach to law enforcement by the individual officer. For example, a police officer may choose not to arrest a suspect because the officer cannot be bothered, does not want to cause trouble or is biased in favor of the perpetrator. A similar point about the multidimensionality of discretion is made in Wilson's (1968) classic description of policing styles. According to Wilson, police departments adopt one of three basic positions on law enforcement. In the service style, the principles of community policing are practiced. Offenses are typically not ignored, but there is a Measuring Police Attitudes 7 high reliance on informal, non-arrest alternatives to resolve minor matters. In exercising discretion, police consider both the nature of the offense and mitigating personal characteristics of the offender. In the watchman style of policing, the purpose of policing is defined as maintaining order rather than enforcing law. Non-enforcement of the law is common. Many minor offenses are simply ignored, as are offenses defined as private disputes. However, here the motivation for non-enforcement is ‘not rocking the boat’ rather than any coherent philosophical rationale. At the same time, where necessary police will get tough with offenders to keep control. Finally, in the legalistic style, arrest is the preferred mode of dealing with crime, even that which is of a minor nature. The seriousness of crime is defined in terms of what was done; little consideration is given to who the offender is. While Wilson was interested in discretion from an organizational perspective, his policing style models nevertheless provide some clues to the range of possible motives for the arrest decisions of individual officers. The aim of the current research is to construct scales for measuring police attitudes towards discretion. Three studies are described. The first study explores the dimensionality of police discretion, that is, it seeks to identify the various personal rationales that underpin the selective enforcement of the law. On the basis of these results, the second study undertakes the construction of relevant scales to measure attitudes to discretion. In the third study, correlations between these scales and existing personality scales are examined. Study 1 The purpose of the first study is develop an initial set of scale items relating to police attitudes towards discretion and to explore the relationships among these items. It is expected that a single strict-lenient continuum with respect to law enforcement will not emerge, but rather that multiple rationales for exercising discretion will be found. Measuring Police Attitudes 8 Method Participants A total of 260 New South Wales (Australia) police officers, all with one year of operational experience, participated in the study (210 males, 50 females; age M = 21.2 years, SD = 3.9). The nature of the sample was determined to a large extent by logistic considerations. Three relatively large and separate samples of police were required for the series of studies described in this paper. It would have been very difficult and time consuming to obtain those samples through conventional mail-out methods of data collection. However, all New South Wales police officers are required to return to the police academy after one year of service for two weeks of refresher courses. It was decided, therefore, to use these police officers as a convenient pool of participants. Materials An initial pool of 28 items relating to the exercise of police discretion was specifically written for the study. Half of the items were written in the forward direction (supporting discretion) and half in reverse direction (supporting no discretion). Items covered a range of perspectives on discretion. In this regard, item writing was in part guided by Wilson’s (1968) three styles of policing, with statements included that expressed service (e.g., ‘Often a caution by a police officer can do more good than an arrest’), watchman (e.g., ‘When it comes to making an arrest smart police know when to mind their own business and not rock the boat’) and legalistic (e.g., ‘The law is the law: police can make no exceptions’) orientations. A questionnaire was constructed using these 28 items. Items were randomly ordered so that forward and reverse items were dispersed. The first page of the questionnaire provided instructions of how to record responses. Each item required a response on a Measuring Police Attitudes 9 5-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = undecided, 4 = agree, 5 = strongly agree) on a separate answer sheet. Procedure The study was conducted with the approval and cooperation of the New South Wales Police Service. Data collection took place at the New South Wales Police Academy. Special classes were timetabled at the end of the two-week refresher course to allow the author to present the questionnaire to participants. A small number of police were not able to attend the classes because of other commitments (e.g., medical appointments) but these absences are not considered a significant source of bias. The author explained the purpose of the study to the class and read out the instructions. Participants were informed that their participation was voluntary. They were also assured that the results were confidential and for research purposes only, and that raw data would not be made available to the Police Service. Completed answer sheets were returned directly to the author in sealed envelopes. Participants were not paid for completing the questionnaire. Results Principal components extraction with varimax rotation was performed on the 28 items using SPSS9 for Windows. There were 9 factors with eigenvalues in excess of 1, accounting for 58.5% of the variance. However, many of these factors contained only a few items and so a more parsimonious solution was sought. An inspection of the scree plot indicated a change in slope after the second eigenvalue. Examination of all solutions involving 1 to 9 factors confirmed that a two-factor solution provided the best conceptual clustering of items (Table 1). Factor 1 had an eigenvalue of 4.24 (14.1% of the variance) and factor 2 had an eigenvalue of 2.66 (10.6% of the variance). Measuring Police Attitudes 10