The Decision to Arrest

  • Michael R. Gottfredson
  • Don M. Gottfredson
Part of the Law, Society, and Policy book series (LSPO, volume 3)


Police decisions have not been a neglected area of study and commentary by social scientists and legal scholars.1 On the contrary, the literature about police work is vast, diverse, and impossible to summarize briefly without violating its complexity.2 Our purposes, however, do not require a thorough review. Rather, we seek to investigate the information requirements, goals, and alternatives of the decision to arrest; to review studies bearing on the ways in which arrest decisions are made routinely; and to offer some comments on directions for the enhancement of rationality. This itself is no small task, for no single aspect of police work has received more scholarly attention recently than has the decision to arrest. Our review must therefore be selective, including only the most informative studies and commentaries.


Criminal Justice Police Officer Criminal Justice System Police Work Patrol Officer 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    One bibliography of works only about police decisions listed 138 entries. See Neith-ercutt and Moseley (1974).Google Scholar
  2. 2.
    Of the many scholarly discussions of police work, the three that should be consulted first are Goldstein (1977), Bittner (1975), and LaFave (1965).Google Scholar
  3. 3.
    The law of arrest is complex and varies, to a limited extent, by jurisdiction. For a discussion of the legal requisites for arrest and the problems inherent in defining arrest for research or analytical purposes, see LaFave (1965: Chapter 1). Our definition here is purposefully restrictive (e.g., excluding citizens’ arrests and arrests of witnesses).Google Scholar
  4. 4.
    It clearly would be easy to minimize the impact of the criminal law function of police work on the daily activities and roles of the police. Such an impression, even if generated on the basis of the studies cited above, would be profoundly erroneous. Even though the amount of time police actually spend engaged in criminal law matters such as arrest is small, the impact of this activity on the role of the police is pervasive. See Goldstein (1977).Google Scholar
  5. 5.
    Because of these relations, Reiss (1971) characterizes citizens as “enforcers of the law.”Google Scholar
  6. 6.
    Quite obviously, this is something of an overstatement. It applies most convincingly to crimes of common theft and assault and much less so to the so-called victimless crimes—e.g., prostitution and drug offenses—although some evidence indicates that citizen complaints are equally important in bringing drunkenness offenses to the attention of the police. See R. Lundman (1974). There are other types of criminal behavior for which the police role is extremely limited—e.g., corporate fraud. Here, as in the rest of this book, our discussion pertains principally to the routine common law crimes.Google Scholar
  7. 7.
    The significant exception, of course, is the use of deadly force. The decision to use deadly force is beyond the intended scope of this chapter. We believe that this decision could, however, be analyzed according to the requirements of rationality employed in this book, particularly with respect to agency policy about the use of deadly force. For extended commentary about deadly force policies and their consequences, see Sherman (1980b).Google Scholar
  8. 8.
    This position has striking analogues to other decisions we discuss in this book. At the sentencing decision, for example (see Chapter 6), recent reform proposals that seek more determinacy and greater consistency would have the legislature define explicitly the penalty that must be given for a certain offense. The motive for determinate sentencing is the same as discussed here for arrests—eliminate discretion by legislation in order to control its abuses. We see the same problems in both solutions. The immense variability inherent in criminal acts, persons, and circumstances precludes the elimination of discretion at both decisions. The inequities that such “discretionless” systems would foster would compel resolution by decision makers somewhere in the system. And the problems of low visibility, reviewability, and control would again surface. As we describe in detail in the final chapter, we believe it possible to design systems of decision making that at once both acknowledge the legitimate role of individualized decision making (discretion) and adequately control it.Google Scholar
  9. 9.
    Some may argue that individualized decision making is indeed a value to be pursued by criminal law but that others (prosecutors or judges) who are elected representatives should be empowered to exercise such discretion. Visibility (and hence reviewability) may be enhanced by requiring these latter decision makers to exercise such judgments. Such a position has merit but must be balanced against the impact that arrest decisions per se can involve. Thus, if an arrest would aggravate the situation, unfairly stigmatize the suspect, or otherwise be detrimental to important values, the police may be in the best position to employ discretion. Police discretion involves a good many activities other than deciding, in a given case, whether to arrest. Whether to enforce a law at all, whether to use force, where to patrol, how to patrol, and so forth are all areas of considerable discretion. Although our focus is exclusively on arrest decisions (as defined earlier), these aspects of police discretion are important. See Goldstein (1977:Chapter 5).Google Scholar
  10. 10.
    Decision problems such as these are far from uncommon. See the examples and analysis in Parnas (1967, 1971).Google Scholar
  11. 11.
    Similar positions are advanced by others. Wilson (1965:21) argues, for example, “Discretion exists both because many of the relevant laws are necessarily ambiguous and because under the law of many states governing arrests for certain forms of disorder, the ‘victim’ must cooperate with the patrolman if the law is to be invoked at all.” Most contemporary standard-setting bodies argue for the acknowledgment of arrest discretion. See American Bar Association (1973), National Advisory Commission on Criminal Justice Standards and Goals (1973), and President’s Commission on Law Enforcement and Administration of Justice, Task Force Report: The Police (1967).Google Scholar
  12. 12.
    We describe these aims fully in Chapter 6, where sentencing decisions are discussed. Most of the contemporary discussion about goals of the criminal law has taken place in the context of sentencing. One theme of this book is that these goals transcend a narrow focus on sentencing; they are critically important to the evaluation of victim, police, bail, correction, and parole decisions as well. As we argue in the final chapter however, any one of these aims is best pursued within the context of specific decision points.Google Scholar
  13. 13.
    The general prevention function of police work has received some evaluation attention (see, e.g., Kelling, Pate, Dieckman, and Brown, 1974). Although a review of this growing literature is beyond the scope of our inquiry here, it is a critical contribution to the information requirements (as we use that term) for rational police decisions. For a review of this and similar issues, see Goldstein (1977) and Sherman (1983).Google Scholar
  14. 14.
    An extensive discussion of reasons for “noninvocation” in arrest decisions is found in LaFave (1965). On the basis of observations in three states, LaFave discovered that some common reasons were: The legislature may not desire full enforcement; there are limited enforcement resources; the victim refuses to prosecute; the victim is involved in the misconduct; the intent is to benefit other aspects of law enforcement (e.g., protect the informant system); the harm caused to the offender or to the victim would outweigh the risk of inaction.Google Scholar
  15. 15.
    It may strike some as strange to consider doing nothing a major alternative in arrest decisions. But we agree with Goldstein that “in trying to develop a rational scheme for handling police business, the option of doing nothing should be recognized as an appropriate alternative in some situations” (1977:40).Google Scholar
  16. 16.
    Sherman (1980a) has provided a classification of research studies concerning police behavior and has reviewed many of their results.Google Scholar
  17. 17.
    See Wilson (1965) and Smith (1984) for studies using this analytical approach. For a study that compares the relative influence of community characteristics and type of police organization, see Swanson (1979) and Talarico and Swanson (1979). For an interesting approach that considers individual rather than organizational policy styles to be influential, see Muir (1977). There are considerable difficulties involved in interagency studies of policy correlates of arrest rates, not the least of which involves the definition and subsequent control of extraneous variance. Although such aggregate studies are widely reported, particularly in the sociological literature, we do not find them convincing. For review of many of these studies, and discussion of potential limitations, see Sherman (1980a).Google Scholar
  18. 18.
    The proportion of cases falling into these categories is not given. It should be noted, however, that these data indicate that the juvenile officers in this study may not be characteristic of police on patrol, given the widely reported finding that calls to the police, rather than direct observation, is by far the principal way in which police become involved in such encounters.Google Scholar
  19. 19.
    Piliavin and Briar do not present tabulations of arrest by specific offense type, and therefore the extent of the relation between offense behavior and the arrest decision cannot be determined. They do report that the importance of demeanor appeared to be much less significant for offenders with known prior records.Google Scholar
  20. 20.
    Their data have been analyzed also by Friedrich (1977).Google Scholar
  21. 21.
    Friedrich also examined how well the characteristics of the officers in the study (e.g., race, length of service, and job satisfaction) predicted their arrest decisions in multiple regression analysis. Virtually none of the variance in arrest (R = .014) was explained by these factors.Google Scholar
  22. 22.
    A number of papers have been generated by these data: Sykes, Fox, and Clarke (1976); Sykes and Clarke (1975); Lundman, Sykes, and Clarke (1978); and Lundman (1974).Google Scholar
  23. 23.
    At least one body of sociological theory, however, posits no relation between decisions made in the criminal justice system and behavioral differences among the clients of the system (e.g., Black, 1976). For a data-oriented critique of these theories, see Gottfredson and Hindelang (1979b, 1980).Google Scholar
  24. 24.
    A now common practice in the sociological literature is to classify variables into legal and extralegal categories and to assess their relative contributions to explain processing decisions. Some such variables are more easily sorted thus than others (e.g., “amount of evidence” is legal and suspect’s race is extralegal). Others are considerably more difficult to classify. The preference of the complainant certainly has implications for evidence, although it is often considered to be an extralegal variable. The demeanor of the suspect (e.g., antagonistic) may be perceived (either rightly or wrongly) by the police to be indicative of the need for custody pending adjudication. It also may be a proxy for racial prejudice and hence, an extralegal consideration.Google Scholar
  25. 25.
    We argue in the final chapter that data systems permitting such studies should be in place and studied routinely by police departments as one step toward the enhancement of rationality.Google Scholar
  26. 26.
    Several excellent reviews of the police referral decision for juveniles are available. See Cohen and Kluegel (1978, 1979), Hirschi (1975), and Nettler (1978). The general findings from this research are compatible with the correlates discussed in the text. In deciding whether to refer arrested juveniles to court, the police seem to rely principally on the seriousness of the alleged behavior and any record of prior illegal conduct. When seriousness of offense is controlled, suspect characteristics (e.g., race and social class) appear to play a role in some places at some times, but not as significant a role as does offense. See Chapter 10 for some implications of these findings, along with the correlates of the other major decisions discussed in this book, for conceptualization about the criminal justice process. Although most police referral studies deal with youth, Pope (1978a) studied police referral decisions for a group of adult burglary arrestees in California. The criterion variable was the postarrest decision (by police) to release suspects or to detain them prior to trial. Although the results are restricted to one type of crime (and hence do not permit inferences about whether the crime itself is related to the decision) the results are informative. The single strongest correlate, among those Pope studied of the release decision, was whether the suspect had a prior record; those without prior records were more likely to be released. Among those with prior records, the young (under eighteen) and those without a history of drug use were less likely to be released. For those without prior records, blacks were more likely to be detained than were whites.Google Scholar
  27. 27.
    Here we are using the term goal in a different context than we do usually in this book. Conviction may be better thought of as an objective—that is, as one method of achieving the aims of desert, deterrence, and the like that are associated typically with decision goals in this book.Google Scholar
  28. 28.
    A consideration of errors in predictive judgments is important here, as with most criminal justice decisions. We take up this topic in detail in Chapter 4, where pretrial detention decisions are discussed. Much of what we say there is indeed applicable to predictive custody decisions at the arrest stage. And the omnipresence of predictive judgments throughout the system weighs heavily in our model designed to enhance rationality, presented in Chapter 10.Google Scholar
  29. 29.
    Although not usually discussed in this context, these errors are false positive errors. One important aspect of false positive errors is that they are hidden errors, in the absence of an experiment showing for certain how many exist. Because predictive judgments are made throughout the criminal justice system, these errors are an issue at every major decision. Predictive judgments are discussed more fully in our consideration of bail decision making (Chapter 4), where they have received more scholarly attention. To some extent the correlates of arrest decisions, identified in the previous section, are indicative of the presumed need for custody. For example, it may be thought that the more serious the offense and the greater the penalties attendant upon conviction, the greater is the custody need at the arrest decision. Such presumptions are in need of empirical testing.Google Scholar
  30. 30.
    Several comprehensive reviews are available: Carter and Klein (1976); Dunford (1977); Klein, Teilman, Styles, Lincoln, and Labin-Rosensweig (1976). For a good study of British practices, see Mott (1983).Google Scholar
  31. 31.
    One review of this literature concludes: “While there seems to be wide-spread agreement about the desirability of diverting youth from the juvenile justice system and a sizeable mobilization of federal, state and local resources for the development of community diversion projects, little has been done to examine how diversion programs have been operationalized or how effectively they function” Dunford (1977:336).Google Scholar
  32. 32.
    For excellent discussions of the problems with and prospects for these various methods of discretion control, see H. Goldstein (1967) and Davis (1975).Google Scholar

Copyright information

© Springer Science+Business Media New York 1988

Authors and Affiliations

  • Michael R. Gottfredson
    • 1
  • Don M. Gottfredson
    • 2
  1. 1.Department of Management and PolicyUniversity of ArizonaTucsonUSA
  2. 2.School of Criminal JusticeRutgers UniversityNewarkUSA

Personalised recommendations