1 Introduction: understanding police innovation
David Weisburd and Anthony A. Braga
Over the last three decades American policing has gone through a period of significant change and innovation. In what is a relatively short historical time frame the police began to reconsider their fundamental mission, the nature of the core strategies of policing, and the character of their relationships with the communities that they serve. Innovations in policing in this period were not insular and restricted to police professionals and scholars, but were often seen on the front pages of America’s newspapers and magazines, and spoken about in the electronic media. Some approaches, like broken windows policing – termed by some as zero tolerance policing – became the subject of heated political debate. Community policing, one of the most important police programs that emerged in this period, was even to give its name to a large federal agency – The Office of Community Oriented Policing Services – created by the Violent Crime Control and Law Enforcement Act of 1994.
Some have described this period of change as the most dramatic in the history of policing (e.g., see Bayley 1994). This claim does not perhaps do justice to the radical reforms that led to the creation of modern police forces in the nineteenth century, or even the wide-scale innovations in tactics or approaches to policing that emerged after the Second World War. However, observers of the police today are inevitably struck by the pace and variety of innovation in the last few decades. Whether this period of change is greater than those of previous generations is difficult to know since systematic observation of police practices is a relatively modern phenomenon. But there is broad consensus among police scholars that the last three decades have “witnessed a remarkable degree of innovation in policing” (Committee to Review Research 2004: 82).
In this volume we bring together leading police scholars to examine the major innovations in policing that emerged during the last decades of the twentieth century. We focus on eight innovations that are concerned with change in police strategies and practices: community policing, broken windows policing, problem-oriented policing, pulling levers policing, third-party policing, hot spots policing, Compstat and evidence-based policing. This is of course not an exhaustive list of innovation in policing during this period. For example, we do not examine technological innovations such as advances in computerized crime mapping or the use of DNA in criminal investigations. We also do not examine innovations in tactics and strategies that affected only specialized units or were applied to very specific types of crimes. Our approach was to identify innovations that had influence on the broad array of police tasks and on the practices and strategies that broadly affected the policing of American communities.
We title the chapters examining each police innovation reviewed in this volume under the heading “Advocate” and “Critic.” In this context our book seeks to clarify police innovation in the context of chapters written by those who have played important roles in developing innovation, and those who have stood as critics of such innovation. Nonetheless, we do not take a debate format in our book. Authors did not respond to each other’s papers, but rather sought to present a perspective that would clarify the benefits of the innovation examined, or the potential problems that the innovation raises for policing. The critics often identify promising elements of innovation while pointing out the difficulties that have been encountered in the applications of innovation in the field. The advocates often note the drawbacks of particular strategies, while arguing that they should be widely adopted. Accordingly, our chapters represent serious scholarly examination of innovations in policing, recognizing that established scholars may disagree about the directions that policing should take while drawing from the same empirical evidence.
By design the essays in this volume take a “micro” approach to the problem of police innovation, focusing on the specific components, goals, and outcomes associated with a specific program or practice. In this introductory chapter, we would like to take a “macro” approach to the problem of police innovation that allows us to see how innovation more generally emerged and developed during this period. We do not think that the dramatic surge in police innovation of the last few decades occurred as a matter of chance. Our approach is to see the development of innovation in policing as a response to a common set of problems and dilemmas. This approach can also help us to understand the broad trends of police innovation that we observe.
Understanding innovation and policing
Many scholars seem to take for granted what strikes us as a central problem in understanding the broader phenomenon that our volume examines. Why did we observe a period of significant innovation in policing in the last decades of the twentieth century? One simple answer to this question would be to note that institutions change, and that when faced with new ideas that have potential to improve their functioning they will naturally choose what is innovative. However, those who have studied the diffusion of innovation have been led to a very different view of the processes that underlie the adoption of new products, programs, or practices. Everett Rogers, who pioneered the scientific study of diffusion of innovation, argues, for example, that “more than just a beneficial innovation is necessary” to explain its widespread diffusion and adoption (1995: 8). Indeed, there are many examples of innovations that represent clear improvements over prior practice, yet fail to be widely adopted.
Rogers brings the example of the “Dvorak Keyboard” named after a University of Washington researcher, who sought to improve on the “Qwerty” keyboard in use since the late nineteenth century. The “Qwerty” keyboard was engineered to slow down typists in the nineteenth century in order to prevent jamming of keys that was common in the manual typewriters in that period. However, as the engineering of typewriters improved in the twentieth century, there was no longer a need for a keyboard engineered to slow typists down. Indeed, it seemed natural that a better arrangement of the keyboard would be developed that would allow for quicker typing that would cause less fatigue. Dvorak developed such a keyboard in 1932 basing his arrangement of the keys on time and motion studies. Dvorak’s keyboard was clearly an improvement on the Qwerty keyboard. It allowed for more efficient and faster typing, and led to less fatigue on the part of typists. But today more than seventy years after Dvorak’s development of a better and more efficient keyboard, the Qwerty keyboard remains the dominant method. Indeed, Dvorak’s keyboard is merely an interesting historical curiosity.
The diffusion of innovation requires that there be a “perceived need” for change in the social system in which an innovation emerges (Rogers 1995: 11). That need can be created by industries or interest groups, for example through advertisements that lead consumers to believe that they must have a particular new product or service. Often in social systems, the recognition that something must change is brought about by a period of crisis or challenge to existing programs or practices (see e.g., Rogers 1995; Altschuler and Behn 1997). In this context, we think that the key to understanding the emergence of a period of rapid innovation in policing in the last decades of the twentieth century lies in a crisis in policing that emerged in the late 1960s. Identifying that period of crisis can help us to understand not only why we observe so much police innovation in recent decades, but also why that innovation follows particular patterns of change.
The crisis of confidence in American policing
The decade of the 1970s began with a host of challenges to the police as well as the criminal justice system more generally (LaFree 1998). This was the case in part because of the tremendous social unrest that characterized the end of the previous decade. Race riots in American cities, and growing opposition, especially among younger Americans, to the Vietnam War, often placed the police in conflict with the young and with minorities. But American fears of a failing criminal justice system were also to play a role in a growing sense of crisis for American policing. In 1967, a presidential commission report on the Challenge of Crime in a Free Society reinforced doubts about the effectiveness of criminal justice in combating crime in the United States:
In sum, America’s system of criminal justice is overcrowded and overworked, undermanned, underfinanced, and very often misunderstood. It needs more information and more knowledge. It needs more technical resources. It needs more coordination among its many parts. It needs more public support. It needs the help of community programs and institutions in dealing with offenders and potential offenders. It needs, above all, the willingness to reexamine old ways of doing things, to reform itself, to experiment, to run risks, to dare. It needs vision. (President’s Commission 1967, 80–81)
Shortly after the presidential report on the Challenge of Crime in a Free Society, the Kerner Commission on Civil Disorders published a report which was also to raise significant questions about the nature of criminal justice in the United States, and the organization of American policing. However, in this case it was the question of race and the relationship between police and minority communities that was to have center stage. The challenges to patterns of American discrimination against African Americans were not focused primarily on the police, but the police, in addition to other criminal justice agencies, were seen as “part of the problem” and not necessarily working to help in producing a solution to difficult social issues:
In Newark, Detroit, Watts and Harlem, in practically every city that has experienced racial disruption since the summer of 1964, abrasive relationships between police and Negroes and other minority groups have been a major source of grievance, tension and ultimately disorder. (Kerner Commission 1968: 157)
The concerns of the commission reports in the 1960s and the sense of growing alienation between the police and the public in the latter half of that decade led policymakers, the police, and scholars to question the nature of American policing, and in particular the strategies that were dominant in policing since World War Ⅱ. A recent National Research Council Committee to Review Research on Police Policy and Practices has termed these approaches as the “standard model of policing” (Committee to Review Research 2004; see also Weisburd and Eck 2004):
This model relies generally on a “one size fits all” application of reactive strategies to suppress crime, and continues to be the dominant form of police practices in the United States. The standard model is based on the assumption that generic strategies for crime reduction can be applied throughout a jurisdiction regardless of the level of crime, the nature of crime, or other variations. Such strategies as increasing the size of police agencies, random patrol across all parts of the community, rapid response to calls for service, generally applied follow-up investigations, and generally applied intensive enforcement and arrest policies are all examples of this standard model of policing. (Weisburd and Eck 2004: 44)
A number of important questions about the standard model of policing had been raised in the 1960s. Nonetheless, there was little serious academic inquiry into the impact of policing strategies on crime or on public attitudes. The need for such research was apparent, and in the 1970s serious research attention was to begin. One important impetus for such studies of the police came from the federal government. With the Omnibus Crime Control and Safe Streets Act of 1968 a research arm of the Department of Justice was established, eventually to become the National Institute of Justice, which was to invest significant resources in research on police and other components of the criminal justice system. But important funding for research on policing was also to come from private foundations. Perhaps the most important contribution to policing was made by the Ford Foundation in 1970 when it established the Police Development Fund. The Fund and the Police Foundation which it established, was to foster a series of large-scale studies on American policing. McGeorge Bundy, then president of the Ford Foundation, argued in announcing the establishment of a Police Development Fund in 1970:
The need for reinforcement and change in police work has become more urgent than ever in the last decade because of rising rates of crime, increased resort to violence and rising tension, in many communities, between disaffected or angry groups and the police. (Bundy 1970)
With the establishment of the Police Foundation and the newly established federal support for research on the criminal justice system, the activities of the police began to come under systematic scrutiny by researchers. Until this time, there had been a general assumption that policing in the post-World War Ⅱ era represented an important advance over previous decades and was effective in controlling crime.
For example, perhaps the dominant policing strategy in the post-World War Ⅱ period was routine preventive patrol in police cars. It was drawn from a long history of faith in the idea of “police patrol” that had become a standard dogma of policing for generations. As George Kelling and his colleagues wrote in their introduction to the Kansas City Preventive Patrol Experiment, a study conducted by the Police Foundation:
Ever since the creation of a patrolling force in 13th century Hangchow, preventive patrol by uniformed personnel has been a primary function of policing. In 20th century America, about $2 billion is spent each year for the maintenance and operation of uniformed and often superbly equipped patrol forces. Police themselves, the general public, and elected officials have always believed that the presence or potential presence of police offices on patrol severely inhibits criminal activity. (Kelling, Pate, Dieckman, and Brown 1974: 1)
Preventive patrol in police cars was the main staple of police crime prevention efforts at the beginning of the decade of the 1970s. As Kelling and colleagues noted in the Police Foundation report on the Kansas City study, “(t)oday’s police recruits, like virtually all those before them, learn from both teacher and textbook that patrol is the ‘backbone’ of police work” (Kelling et al. 1974: 1). The Police Foundation study sought to establish whether empirical evidence actually supported the broadly accepted assumptions regarding preventive patrol. The fact that questions were raised about routine preventive patrol suggests that the concerns about the police voiced in the decade before had begun to impact the confidence of police managers. As Kansas City Police Chief Clarence M. Kelley, later to become director of the Federal Bureau of Investigation (FBI), said in explaining the need for the Kansas City experiment: “Many of us in the department had the feeling we were training, equipping, and deploying men to do a job neither we, nor anyone else, knew much about” (Murphy 1974: ⅴ).
To understand the impact of the Kansas City study on police managers and researchers, it is important to recognize not only that the study examined a core police practice but that its methodological approach represented a radical departure from the small-scale evaluations of police practices that had come earlier. The Kansas City Preventive Patrol Experiment was a social experiment in policing on a grand scale, and it was conducted in a new Foundation that had significant resources and was backed by the well-established and respected Ford Foundation. Patrick Murphy, the distinguished police manager, and president of the Police Foundation at the time, suggests just how much the Foundation itself saw the experiment as a radical and important change in the quality of police research.1
This is a summary report of the findings of an experiment in policing that ranks among the few major social experiments ever to be completed. The experiment was unique in that never before had there been an attempt to determine through such extensive scientific evaluation the value of visible police patrol. (Murphy 1974: ⅴ)
This context, both in terms of the centrality of the strategy examined, the scale of the research, and the prestige of the institutions that supported the study, including the Kansas City Police Department and its chief, Clarence Kelley, were to give the findings of the study an impact that is in retrospect out of proportion to the actual findings. One study in one jurisdiction, no matter how systematic, cannot provide a comprehensive portrait of the effects of a strategy as broad as routine preventive patrol. Moreover, the study design was to come under significant academic criticism in later years (Minneapolis Medical Research Foundation 1976; Larson and Cahn 1985; Sherman and Weisburd 1995). Nonetheless, in the context of the decade in which it was conducted, this study was to have a critical impact upon the police and police researchers. This was especially the case since the research findings were to be consistent with a series of other studies of core police practices.
Kelling and his colleagues, in cooperation with the Kansas City Police Department, took fifteen police beats and divided them up into three groups. In five of these, called “reactive” beats, “routine preventive patrol was eliminated and officers were instructed to respond only to calls for service” (Kelling et al. 1974: 3). In five others, defined as “control” beats, “routine preventive patrol was maintained at its usual level of one car per beat” (ibid.: 3). In the remaining five beats, termed “proactive” beats, “routine preventive patrol was intensified by two to three times its usual level through the assignment of additional patrol cars” (Kelling et al. 1974: 3). When Kelling and his colleagues published the results of their study in 1974 it shattered one of the bedrock assumptions of police practitioners – that preventive patrol was an effective way to prevent crime and increase citizens’ feelings of safety. They concluded simply that increasing or decreasing the intensity of routine preventive patrol in police cars did not affect either crime, service delivery to citizens, or citizens’ feelings of security.
Another large-scale study conducted by William Spelman and Dale Brown and published in 1984 was also to challenge a core police assumption of that period – that improvement in rapid response to calls for service would lead to improvements in crime fighting. This study was developed in good part because of the findings of a prior investigation in Kansas City that found little support for the crime control effectiveness of rapid response to calls for service (Kansas City Police Department 1977). With support from the National Institute of Justice, Spelman and Brown interviewed 4000 victims, witnesses, and bystanders in some 3300 serious crimes in four American cities. This was another major study in terms of the resources brought to bear and the methods used. Again it examined a strategy that was aided by technological advances in the twentieth century and that was a central dogma of police administrators – that police must get to the scene of a crime quickly if they are to apprehend criminal offenders. Spelman and Brown explained:
For at least half a century, police have considered it important to cut to a minimum their response times to crime calls. The faster the response, they have reasoned, the better the chances of catching a criminal at or near the scene of the crime. (Spelman and Brown 1984: ⅹⅺ)
Based on the data they collected, however, Spelman and Brown provided a very different portrait of the crime control effectiveness of rapid response to calls for service:
Rapid police response may be unnecessary for three out of every four serious crimes reported to police. The traditional practice of immediate response to all reports of serious crimes currently leads to on-scene arrests in only 29 of every 1,000 cases. By implementing innovative programs, police may be able to increase this response-related arrest rate to 50 or even 60 per 1000, but there is little hope that further increases can be generated. (Spelman and Brown 1984: ⅺⅹ)
These findings based on a host of systematic data sources from multiple jurisdictions provided little support for the strategy of rapid response as a police practice to do something about crime. Indeed, Spelman and Brown found that citizen reporting time, not police response time, most influenced the possibility of on-scene arrest. Marginal improvement in police response times was predicted to have no real impact on the apprehension or arrest of offenders.
The Kansas City Preventive Patrol Experiment and the National Institute of Justice study of police response time were not the only studies to “debunk” existing police practices. James Levine, for example, analyzed national crime data on the effectiveness of increasing the number of police in an article published in 1975. His title sums up his findings: “The Ineffectiveness of Adding Police to Prevent Crime.” Despite the fact that this effort and many others that challenged conventional police practices did not represent the kind of systematic data collection or analysis of the Police Foundation and National Institute of Justice studies, they followed a similar “narrative” which became increasingly common as the 1990s approached. Levine, for example, begins by noting the broad consensus for the principle that adding more police will make cities safer. He then goes on to note that “(s)ensible as intensified policing may sound on the surface, its effectiveness in combating crime has yet to be demonstrated” (Levine 1975: 523). Finally, drawing upon simple tabular data on police strength and crime rates over time, he concludes:
It is tempting for politicians and government leaders to add more police: it is an intuitively sensible and symbolically satisfying solution to the unrelenting problem of criminal violence . . . The sad fact is, however, that they receive a false sense of security; in most situations they are just as vulnerable with these extra police as without them. (Levine 1975: 544)
Follow-up investigations were also the subject of critical empirical research during this period. The standard model of policing had assumed that general improvements in methods of police investigations would lead to crime control gains both because more active offenders would be imprisoned and thus unable to commit crime, and because potential offenders would be deterred by the prospect of discovery and arrest (Committee to Review Research 2004). But a series of studies in the 1970s and early 1980s suggested that investigations had little impact upon crime (Greenwood et al. 1975; Greenwood, Petersilia, and Chaiken 1977; Skogan and Antunes 1979; Eck 1983). This was the case in good part because many crimes, especially property crimes, were found to be unlikely to be solved by police investigations. These studies consistently showed that if citizens did not provide information about suspects to first responding officers, follow-up investigations were unlikely to lead to successful outcomes.
In retrospect, many of these studies overstated what could be learned about standard police practices from the findings gained (Weisburd and Eck 2004). And, in practice, there were evaluations in this period that produced more promising findings regarding standard police practices such as routine preventive patrol (e.g., see Press 1971; Schnelle, Kirchner et al. 1977; Chaiken 1978). Nonetheless, as the United States entered the decade of the 1990s there appeared to be a general consensus that traditional police practices did not work in preventing or controlling crime. As Michael Gottfredson and Travis Hirschi wrote in their classic book A General Theory of Crime in 1990: “No evidence exists that augmentation of patrol forces or equipment, differential patrol strategies, or differential intensities of surveillance have an effect on crime rates” (Gottfredson and Hirschi 1990: 270).
David Bayley wrote even more strongly in 1994:
The police do not prevent crime. This is one of the best kept secrets of modern life. Experts know it, the police know it, but the public does not know it. Yet the police pretend that they are society’s best defense against crime . . . This is a myth. First, repeated analysis has consistently failed to find any connection between the number of police officers and crime rates. Secondly, the primary strategies adopted by modern police have been shown to have little or no effect on crime. (Bayley 1994: 3)
|Image not available in HTML version|
Figure 1.1 Total violent crime (trends in violent crime rates by city size)
Source: Reiss and Roth, 1993
This view of the ineffectiveness of policing strategies was reinforced by official crime statistics. These statistics, widely available to the public, suggested that the police were losing the “war on crime.” In particular, in America’s largest cities, with their well-established professional police forces, crime rates and especially violent crime rates were rising at alarming rates. Between 1973 and 1990 violent crime doubled (Reiss and Roth 1993). It did not take a statistician to understand that the trends were dramatic. For example, in Figure 1.1 we report the trends in violent crime rates by city size per 100,000 population. Clearly crime was on the rise, and the trend had been fairly consistent over a long period. Thus, not only were scholars showing that police strategies did little to impact upon crime, but the overall crime statistics commonly used by the government and community to define police effectiveness were providing a similar message.