Problem-oriented policing has its roots in public administration and political science. It starts from an intellectual interest in how to get the police to be more effective in carrying out their functions in democratic societies. Problem-oriented policing as a distinct model of police reform evolved out of Herman Gold steins early involvement in the American Bar Foundation Survey of Criminal Justice of the 1950's. Thus, in one sense problem-oriented policing is only 20 years old, but its intellectual heritage is several decades older.

The findings and conclusions that emerged from the survey provided much of the intellectual foundation for problem-oriented policing (Goldstein 1991). Problem-oriented policing has at times been criticized for lacking a criminological theory for its foundation. This criticism presumes that a theory for improving police service must first set forth a theory for preventing crime. This, however, is a far more ambitious, and perhaps unrealistic, goal to which problem-oriented policing never aspired. Moreover, any proposal to improve the quality of policing must address the full range of police tasks and responsibilities, and not merely the control of serious crime. Problem-oriented policing is best understood as a framework for organizing the police and their activities so that the police are better positioned to learn how to prevent crime and disorder, and to apply that knowledge.

It has no explicit preference for one criminological theory over others. It seeks to leave the police open to understanding various criminological theories, and experimenting with practical applications of those theories to determine what works best under what circumstances. This is not to say that problem-oriented policing proponents do not have favored criminological theories. Indeed, among the reasons there has been so much cross-fertilization of ideas between problem-oriented policing and situational crime prevention is that problem-oriented proponents have found merit in the theories underlying situational crime prevention, and police practitioners find the situational crime prevention studies relevant to their own work.

But if those theories were ultimately proven wrong, it is unlikely that problem-oriented policing advocates would similarly conclude that the problem-oriented approach was also wrong. It would merely add to the knowledge base from which police practitioners could draw to guide their strategic decisions. Many police agencies have systematically analyzed reported crime data for a number of reasons to identify potential suspects in specific crimes, to spot geographic and temporal crime trends, and to generally report crime and account for police responses to it. However, crime analysis, as it has conventionally been practiced, is quite different from problem analysis in several respects.

Crime analysis was used principally to identify offenders or predict the next crime in a pattern. Problem analysis is used to bring about more permanent reductions in the levels or severity of problems. Crime analysis concentrated on police activities to address crime. Problem analysis explores the whole community response to the problem. Some agencies now have their crime analysts engaged in broader problem analysis, though mainly by providing, on request, statistical reports and analyses to those line officers leading problem solving initiatives. Problem-oriented policing specifically calls for a broad inquiry into many types of community problems demanding police attention.

It also calls for analyzing multiple sources of information to develop a fuller understanding of each problem. Where a Compstat-style method results in commanders selecting from among a limited and conventional set of responses to address problems, such as extra patrol or increased enforcement, it also departs radically from a problem-oriented methodology. Problem-oriented policing calls for a broad and uninhibited search for responses to particular problems, placing special emphasis on responses that minimize the need for the police to use force and large-scale arrest campaigns. A Compstat-style method can foster a hostile atmosphere, more like an inquisition than an inquiry; in this sense, it also differs from problem-oriented policing. Problem-oriented policing, while stressing accountability also places a high priority on the free exchange of ideas, an exchange that is difficult to achieve in a tension-filled and rigidly hierarchical setting. Finally, a Compstat-style method relies exclusively on police analysis of data and results in decisions made exclusively by the police; in this sense, it also does not resemble problem-oriented policing.

Problem- oriented policing puts a high premium on communication, consultation and collaboration with entities outside the police department at all stages of the planning process. Ideally, a Compstat-style method would be entirely consistent with problem-oriented policing. As one way to identify specific problems, a computer-generated pattern of crimes would be only the beginning of a more in-depth and broader analysis of the nature of the problems, their underlying conditions and the limits of current responses. For example, if computerized systems recognized a sudden spate of incidents classified as robberies in a police precinct, this information would not be used merely to mobilize conventional police responses like stakeouts and extra patrol, but instead might launch a closer analysis of the incidents that could reveal several discrete forms of problems, all related to the crime of robbery, each calling for a different set of responses. This should not be understood as an attack on the Compstat method. For many police agencies, this method is a significant advancement in the use of crime data to inform operational decisions.

Problem-oriented policing, however, is a considerably more sophisticated and involved approach to handling police business than a Compstat method simplistically practiced. Problem-oriented policing specifically calls for, among other things, an analysis of police incidents in terms of location as a potentially useful way to aggregate incidents into clusters. A spatial incident pattern can help stimulate a better understanding of the underlying causes of certain community problems. Crime mapping also fits well with situational crime prevention, in which understanding crime in the specific context of its location is critical.

For example, crime maps might reveal a pattern of storage-facility burglaries, and that revelation might then prompt a closer analysis of those facilities physical layout and management.