Northwestern University,



University of California, Riverside

Current theoretical and research formulations in the sociology of deviance are cast within the general framework of social and cultural differentiation, deviance, and social control. In contrast to the earlier moralistic conceptions of the "pathologies," the focus of description and analysis has shifted from the vagaries of morbid behavior to the patterning effects of the social cultural environment on forms of deviant conduct. These forms of deviation are conceived as social products of the organization of groups, social structures, and institutions.

Three major lines of inquiry have developed within this general framework. One development has been the problem of explaining the rates of various fortes of deviation among various segments of the population. The research devoted to this problem has produced a large body of literature in which individual, group, and areal (e.g., census tracts, regions, states, etc. ) characteristics are correlated with rates of deviation. Durkheim's pioneer study of suicide is a classic example of this sociological interest. Merton's more general theory of social structure and anomies may be cited as the most widely circulated statement of this problem.

The second line of investigation has been directed to the question of how individuals come to engage in various types of deviant behavior. From the theoretical standpoint, this question has been posed by the fact that although an aggregate of individuals may be exposed to the 'same' sociogenic factors associated with deviant behavior some individuals become deviant while others do not. Research into this problem has led some sociologists into the field of actuarial statistics and others to social and depth psychology to investigate differences in individual "adaptation" to the social-cultural environment. The search for the etiology of deviant behavior in individual differences has re-introduced the notion of "pathology," in the garb of "emotionally disturbed" "psychopathic personality" "weak ego-structure" and other psychological concepts which has created an hiatus between sociological and social psychological approaches. Sutherland's differential association theory3 represents formulation which at tempts to account for the etiology of deviant behavior within the general framework of "normal'' learning processes.

A third line of inquiry has been concerned with the developmental processes of "behavior systems." The- and research on this aspect of behavior focuses on the relation between the social differentiation of .the deviant, the organization of deviant activity, and the individual's conception of himself as deviant. Studies of the professional thief, convicts, prostitutes- alcoholics, hobos, drug addicts, carnival men, and others describe and analyze the deviant subculture and its patterning effects on the interaction between deviant and others. The work of Lemert presents a systematic theoretical and empirical integration of this interest in the sociology of deviance.

Although the three lines of investigation share a common interest in the organizational "sources" of deviant behavior, a theoretical integration between them has not been achieved. This is particularly apparent in the theoretical and methodological difficulties posed by the problem of relating the rates of deviant behavior to the distribution of "sociogenic" factors within the social structure. These difficulties may be stated in the form of two questions: (1) How is "deviant behavior" to be defined sociologically, and (2) what are the relevant rates of deviant behavior which constitute the "facts to be explained"? We shall propose that these difficulties arise as a consequence of the failure to distinguish between the social conduct which produces a unit of behavior (the behavior- producing processes) and organizational activity which produces a unit in the rate of deviant behavior (the rate- producing processes. )5 The failure to make this distinction has led sociologists to direct their theoretical and empirical investigations to the behavior-producing processes on the implicit assumption that the rates of deviant behavior may be explained by them. We shall discuss some of the consequences of this distinction for theory and research in the sociology of deviance by examining the problems of the "appropriateness'' and "reliability" of official statistics.6

The following statement by Merton is a pertinent and instructive point of departure for a discussion of the questions raised above:

Our primary aim is to discover how some social structures exerts a definite pressure upon certain persons in the society to engage in non-conforming rather than conforming conduct. If we can locate groups peculiarly subject to such pressures, we would expect to find fairly high rates of deviant behavior in those groups, not because the human beings comprising them are compounded of distinctive biological tendencies but because they are responding normally to the social situation in which they find themselves. Our perspective examines sociological variations in the rates of deviant behavior, not in its incidence."7

The central hypothesis that Merton derives from his theory is that "aberrant behavior may be regarded as a symptom of dissociation between culturally prescribed aspirations and socially structured avenues for realizing these aspirations."8 The test of this general hypothesis, he suggests, would be to compare the variations in the rates of aberrant behavior among populations occupying different positions within the social structure. The question arises: What are the units of behavior which are to be tabulated to compile these rates of aberrant behavior?

Merton answers this question by discussing the kinds of rates which are "inappropriate," but he is less explicit about what may be considered "appropriate" data for sociological research. Discussing the relevance of his theory for research on juvenile delinquency, Marion presents two arguments against the use of "official" rates of deviant behavior. He asks:

...to what extent and for which purposes is it feasible to make use of existing data in the study of deviant behavior? By existing data I mean the data which the machinery of society makes available census data, delinquency rates as recorded in official or unofficial sources, data on the income distribution of an area, on the state of housing in an area, and the like...
"There is little in the history of how statistical series on the incidence of juvenile delinquency came to be collected that shows them to be the result of efforts to identify either the sources or the contexts of juvenile delinquency. These are social bookkeeping data. And it would he a happy coincidence if some of them turned out to be in a form relevant for research.
"From the sociological standpoint, juvenile delinquency and what it encompasses is a form of deviant behavior for which the epidemiological data, as it were, may not be at hand. You may have to go out and collect your own appropriately organized data rather than to take those which are ready-made by governmental agencies.9

Our interpretation of this statement is that for the purposes of sociological research, official statistics may use categories which are unsuitable for the classification of deviant behavior. At best such statistics classify the "same'' forms of deviant behavior in different categories and "different" forms in the same categories. Thus, the "sources or the contexts" of the behavior are obscured.

Merton also argues against the use of official statistics on quite different grounds. He states that such data are "unreliable" because "successive layers of error intervene between the actual event and the recorded event, between the actual rates of deviant behavior and the records of deviant behavior."10 In this statement, the argument is that the statistics are unreliable because some individuals who manifest deviant behavior are apprehended, classified and duly recorded while others are not. It is assumed that if the acts of all such individuals were called to the attention of the official agencies they would be defined as deviant and so classified and recorded. In referring to the 'unreliability' of the statistics in this sense, however, Merton appears to suspend his "sociologically relevant' definition of deviant behavior and implicitly invokes the definitions applied by the agencies which have compiled the statistics. That is, the "unreliability" is viewed as a technical and organizational problem, not a matter of differences concerning the definition of deviant behavior.

Thus, Merton argues against the use of official statistics on two separate grounds. On the one hand, official statistics are not appropriately organized for sociological research because they are nor collected by the application of a "sociologically relevant" definition of deviant behavior. On the other hand, he implies, that official statistics could .be used if "successive layers of error" did' not make them "unreliable." But if the statistics are inappropriate for sociological research on the first ground, would they not be inappropriate regardless of their "unreliability''? It is evident, however, that inappropriate or not, sociologists, including Merton himself, do make use of the official statistics after a few conventional words of caution concerning the "unreliability" of such statistics. The "social bookkeeping data" are, after all, considered to bear some, if unknown, relation to the "actual" rates of deviant behavior that interest sociologists. But granted that there are practical reasons for the use of official statistics are there any theoretical grounds which justify their use, or is this large body of data useless for research in the sociology of deviance? This question directs us to examine more closely the theoretical and methodological bases of the two arguments against their use.


The objection to the official statistics because they are "inappropriate" is, as indicated above, on definitional grounds. The argument is that insofar as the definitions of deviant behavior incorporated in the official statistics are not "sociologically relevant," such statistics are in principle "inappropriate" for sociological research. What then is a sociologically relevant definition of deviant behavior and what are to be considered appropriately organized data for sociological research?12

We suggest that the question of the theoretical significance of the official statistics can be rephrased by shifting the focus of investigation from the processes by which certain forms of behavior are socially and culturally generated to the processes by which rates of deviant are produced. Merton states that his primary aim is to explain the former processes, and he proposes to look at variations in the rates of deviant behavior as indices of the processes. Implicit in this proposal is the assumption that an explanation of the behavior-producing processes is also an explanation of the rate-producing processes. This assumption leads Merton to consider the correspondence between the forms of behavior which his theory is designed to explain and their distribution in the social structure as reflected in some set of statistics, including those commonly used "official statistics" which are ready-made by governmental agencies.

Let us propose, however, the following: Our primary aim is to explain the rates of deviant behavior. So stated, the question which orients the investigation is not how individuals are motivated to engage in behavior defined by the sociologist as "deviant." Rather the definition and content of deviant behavior are viewed as problematic, and the focus of inquiry shifts from the forms of behavior (modes of individual adaptation in Merton's terminology) to the "societal reactions" which define various forms of behavior as deviant.13 In contrast to Merton's formulation which focuses on forms of behavior as dependent variables (with structural pressures conceived to be the independent variables), we propose here to view the rates of deviant behavior as dependent variables. Thus, the explanation of rates of deviant behavior would be concerned specifically with the processes of rate construction.

The problem of the definition of ''deviant behavior" is directly related to the shift in focus proposed here. The theoretical conception which guides us is that the rates of deviant behavior are produced by the actions taken by persons in the social system which define, classify and record certain behaviors as deviant.14 If a given form of behavior is not interpreted as deviant by such persons it would not appear as a unit in whatever set of rates we may attempt to explain (e.g., the statistics of local social welfare agencies, "crimes known to the police," Uniform Crime Reports, court records etc.). The persons who define and activate the rate-producing processes may range from the neighborhood "busybody" to officials of law enforcement agencies15. From this point of view, deviant behavior is behavior which is organizationally defined, processed, and treated as "strange," "abnormal," "theft," "delinquent," etc., by the personnel in the social system which has produced the rate. By these definitions, a sociological theory of deviance would focus on three interrelated problems of explanation: (1) How different forms of behavior come to be defined as deviant by various groups or organizations in the society (2) how individuals manifesting such behaviors are organizationally processed to produce rates of deviant behavior among various segments of the population, and (3) how acts which are officially or unofficially defined as deviant are generated by such conditions as family organization, role inconsistencies or situational "pressures."

What are the consequences of these definitions for the question regarding the relevance of official statistics for sociological research? First, the focus on the processes by which rates are produced allows us to consider any set of statistics, "official" as well as "unofficial," to be relevant. The question of whether or not the statistics are "appropriately organized" is not one which is determined by reference, to the correspondence between the sociologist's definition of deviant behavior and the organizational criteria used to compile the statistics. Rather the categories which organize a given set of statistics are taken as given the "cultural definitions" to use Merton's term, of deviant behavior are par excellence the relevant definitions for research. The specification of the definitions explicitly or implicitly stated in the statistical categories is viewed as an empirical problem. Thus, the question to be asked is not about the "appropriateness" of the statistics but about the definitions incorporated in the categories applied by the personnel of the rate-producing social system to identify, classify, and record behavior as deviant.

Second, a unit in a given rate of deviant behavior is not defined in terms of a given form of behavior or a "syndrome" of behavior. The behaviors which result in the classification of individuals in a given deviant category are not necessarily similar, i.e., the objective manifestation of the "same" forms of behavior may result in the classification of some individuals as deviant but not others. For example, with reference to the rates of delinquency reported by the police department we would ask: What are the criteria that the police personnel use to identify and process a youth as "incorrigible," "sex offender," "vandal" etc.? The criteria of such categories are vague enough to include a wide range of behaviors which in turn may be produced by various sources and contexts" within the social structure.''16

Third, the definition of deviant behavior as behavior which is organizationally processed as deviant provides a different perspective on the problem of the "unreliability" of the official statistics. Insofar as we are primarily concerned with explaining rates rather than the forms of deviant behavior, such statistics may be accepted as a record of the number of those who have been differentiated as variously deviant at different levels of social control and treatment. The "successive layers of error" which may result from the failure of control agencies to record all instances of certain forms of behavior or from the exclusion of cases from one set of statistics that are included in another, do not render such statistics "unreliable," unless they are assigned self-evident status. By the definition of deviance proposed here, such cases are not among those processed as deviant by the organizations which have produced the statistics and thus are nor officially deviant. To reject these statistics as "unreliable" because they fail to record the "actual" rate of deviant behavior assumes that certain behavior is always deviant independent of social actions which define it as deviant.

Forth, the conception of rates of deviant behavior as the product of the socially organized activities of social structures provides a method of specifying the "relevant structure" to be investigated. The rates are constructed from the statistics compiled by specifiable organizations, and those rates must be explained in terms of the deviant processing activities of those organizations. Thus, rates can be viewed as indices of organizational processes rather than as indices of the incidence of certain forms of behavior. For example, variations in the rates of deviant behavior among a given group (e.g., Negroes) as reflected in the statistics of different organizations may be a product of the differing definitions of deviant behavior used by those organizations, differences in the processing of deviant behavior, differences in the ideological, political, and other organizational conditions which effect the rate-making processes.


We wish now to discuss briefly some recent work17 concerning adult and juvenile criminal acts which lends support to the thesis presented above. Let us assume that an ideal system of law-enforcement would lead to the apprehension of all persons who have committed criminal acts as defined by the statutes, and adjudicated in the manner prescribed by those statutes. In the ideal case, there would be little room for administrative interpretation and discretion. The adjudication process would proceed on the basis of evidence deemed legally admissible and the use of the adversary system to convict those who are guilty and exonerate those against whom there is insufficient evidence18. Criminologists have long recognized that the practiced and enforced system of criminal law, at all levels of the process, does not fulfill this ideal conception of criminal justice strictly governed by the definitions and prescriptions of statutes. Therefore, criminal statistics clearly cannot be assumed to reflect a system of criminal justice functioning as ideally conceived and "labels assigned convicted defendants" are not to be viewed as "the statutory equivalents of their actual conduct.19

What such statistics do reflect, however, are the specifically organizational contingencies which condition the application of specific statutes to actual conduct through the interpretations, decisions and actions of law enforcement personnel. The decisions and discretionary actions of persons who administer criminal justice have been documented by the American Bar Foundation study cited above. That study and other research20 indicates the following:

  1. There is considerable ambiguity in defining the nature of criminal conduct within the limits defined by the statutes. Categories of criminal conduct are the product of actual practices within these limits and the decisions which must be made to provide the basis for choosing the laws which will receive the greatest attention.
  2. The discretion allowed within the administration of criminal justice means that admissible evidence may give way to the prosecutor's power to determine whether or not to proceed even in cases where there is adequate evidence to prosecute. The judge, as well as the police or the victim also has discretion (e.g., sentencing) and some discretion is also extended to correctional institutions.
  3. Most persons charged with criminal conduct plead guilty ( from 80 to 90 per cent. according to the references cited by Newman ) and jury trials are rare. Thus, the adversary aspect of the law is not always practiced because many of the lower income offenders cannot afford lawyers and often distrust public defender. Criminal justice depends upon a large number of guilty pleas. Many of these cases would be acquitted if there were more trials.
  4. Statistics are affected by such "accommodations in the conviction process." Some offenders are excluded because they are not processed even though known to be guilty (e.g., drug addicts, prostitutes and gamblers are often hired by the police or coerced by then to help apprehend other offenders ), and the practice of re-labeling offenses and reducing sentences because of insufficient evidence, "deals," and tricks (e.g., telling the defendant or his lawyer that because the offender "seems like a decent person" the charge will be reduced from a felony to a misdemeanor, when in fact the prosecution finds there is insufficient evidence for either charge.) These accommodations may occur at the time of arrest, or during prior or subsequent investigation of crimes, filing of complaints, adjudication, sentencing and post-sentencing relations with authorities, and so on.

The significance of the American Bar Foundation study goes beyond the documentation of the usual complaints about inadequate recording, inflated recording, and the like. More importantly, it underlines the way criminal statistics fail to reflect the decisions made and discretion used by law-enforcement personnel and administrators, and the general accommodations that can and do occur. An offender's record, then, may never reflect the ambiguous decisions, administrative discretions, or accommodations of law enforcement personnel; a statistical account may thus seriously distort an offender's past activities. The administration of justice vis-a-vis juveniles is even more discretionary than for adults due to the philosophy of the juvenile court law. The juvenile offender is not officially viewed as a criminal, but rather as an adolescent who is "misdirected," "disturbed," from a "poor environment," and the like. The legal concept of an adversary system is notably absent. The philosophy, however, is differentially interpreted with police more likely to view juveniles as adult criminals, while probation officers and some judges view the offender within the intended meaning of the law. The early work of Paul Tappan on juvenile court practices21 shows how a juvenile court judge, on the counsel of a social worker or other "treatment oriented" personnel, may dispose of a case in a manner which negates all previous characterizations of the offender by police, probation officer, school officials, and the like. The report of the more recent California Special Study Commission on Juvenile Justice22 alludes vaguely to and in some passages flatly states that many variations of organizational procedures and interpretations by personnel differentially influence the administration of juvenile justice in California. The use of existing stereotypes and imputations of social characteristics to juvenile defendants by law enforcement personnel routinely introduce non-legal criteria and actions into the organizational procedures of the legal process and significantly influences the realization of judicial objectives.23We wish to state explicitly that the interpretation of official statistics proposed here does not imply that the forms of behavior which the sociologist define and categorize as deviant (e.g., Merton's modes of adaptation ) have no factual basis or theoretical importance. Nor do we wish to imply .that the question of how behaviors so defined are produced by the social structure is not a sociologically relevant question. The implication of our interpretation is rather that with respect to the problem of rates of behavior the theoretical question is: what forms of behavior are organizationally defined as deviant; and how are they classified, recorded and treated by persons in the society?

In our discussion, we have taken the view that official statistics, reflecting as the, do the variety of organizational contingencies in the process by which deviants are differentiated from non-deviants, are sociologically relevant data. An individual who is processed as "convicted," for example, is sociologically differentiable from one who is "known to the police" as criminal-the former may legally be incarcerated, incapacitated and socially ostracized, while the latter remains "free." The fact that both may have "objectively" committed the same crime is of theoretical and empirical significance, but it does nor alter the sociological difference between them. The pattern of such "errors" is among the facts that a sociological theory of deviance must explain, for they are indications of the organizationally defined processes by which individuals are differentiated as deviant. Indeed, in modern societies where bureaucratically organized agencies are increasingly invested with social. control functions, the activities of such agencies are centrally important "sources and contexts" which generate as well as maintain definitions of deviance and produce populations of deviants. Thus, rates of deviance constructed by the use of statistics routinely issued by these agencies are social facts par excellence. A further implication of this view is that if the sociologist is interested in how forms of defiant behavior are produced by social structures, the forms that must be explained are those which not only are defined as deviant by members of such structures but those which also activate the unofficial and/or "official" processes of social control. By directing attention to such processes, the behavior-producing and rate-producing processes may be investigated and compared within a single framework.


1. We wish to acknowledge the support of the Youth Development Program of the Ford Foundation in facilitating the preparation of this paper.

2 Robert K. Merton, Social Theory and Social Structure revised, Glencoe: The Free Press, 1957, Chapter 4.

3 Edwin H. Sutherland and Donald R. Cressey Principle of Criminology fifth edition, New York: Macmillan I 1956 Chapter 4.

4 Edwin M Lemert Social Pathology New York: McGraw-Hill 1951 esp. Chapters 1 4. See also, Sutherland and Cressey op. cit., Chapters 12.

5 The conception of the "rate-producing" processes as socially organized activities is taken from work by Harold Garfinkel, and is primarily an application of w hat he ha!; termed the "praxeological rule." See Harold Garfinkel, 'Some Sociological Concepts and Methods for Psychiatrists," Psychiatry Research Reports. 6 (October, 1956), pp 181- 195; Harold Garfinkel and Harry Brickman, "A Study of the (compositions of the Clinic Patient Population of the Out patient Department of the U.C.L.A Neuropsychiatric Institute unpublished manuscript.

6. For a discussion of these problems, see Sophia M. Robison Can Delinquency Be Measured?, New York Columbia University Press, 1936 See also Sutherland and Cressey, op. cit., Chapter 2.

7 Robert K. Merton op. Cit. p. 147 Merton 's comments on the theory of social structure and anomie may be found in Chapter 5 of that volume, and in "Social Conformity Deviation, and Opportunity Structures A Comment on the Contributions of Dubin and Cloward American Sociological Review 24 (April, 1959 pp.177-189 See also his remarks in New Perspectives for Research on Juvenile Delinquency H. Witmeer and R. Kotinsky, editors, U. S. Government Printing Office. 1956.

8 Social Theory and Social Structure op. Cit. Pp. 1 34.

9 New Perspectives for Research, Juvenile Delinquency. op. Cit. p 32

10 Ibid., p 31.

11. For example, ''. . crude (and not necessarily reliable) crime statistics suggest . . .etc. Social Theory and Social Structure op. cit. p. 147. In a more extensive comment on the limitations imposed on research by the use of official statistics Merton states: "Its decisive limitations derives from a circumstance which regularly confronts sociologists seeking to devise measures of theoretical concepts by drawing upon an array of social data which happen to be recorded in the statistical series established by agencies of the society -namely the circumstance that these data of social bookkeeping which happen to be on hand are not necessarily the data which best measure the concept. Pragmatic considerations of this sort are of course no suitable alternative to theoretically derived indicators of the concept." p. 165.

12. Merton proposes to define deviant behavior in terms of the "acceptance" or rejection" of cultural goals and/or institutionalized means. Interpreting the two terms literally, a given form of behavior (adaptation ) is to be considered deviant if it is oriented by some cultural goals (to be specified by the sociologists) and/or the institutionalized means (also to he specified) which govern conduct with respect to those goals. By this definition appropriately organized data would require that behaviors be classified in the typology of "modes of individual adaptation." But what are the operational criteria by which acceptance or ''rejection" of cultural goals and institutionalized means are to be inferred from observed behavior? How for example. is the sociologist to distinguish, between behavior which indicates 'conformity'' from ''over-conformity" (which presumably would be classified as ''ritualism ) or retreatism from innovation? Unless a set of rules for the classification of behavior as deviant can be derived from the theory, rates of deviant behavior cannot be constructed to test its validity.

13. For a discussion of the concept of "societal reaction" see Edwin M. Lemert op. en., Chapter 4.

14. For a preliminary research application of this formulation, see John 1. Kitsuse, "Societal Reaction to Deviant Behavior: Problems of Theory and Method,'' Social Problems 9 Winter 1962), pp. 247-56.

15. We recognize, of course, that many individuals may be labeled ''strange," 'crooks," ''crazy," etc., and ostracized by members of a community, yet be unknown to the police or any other official agency. Insofar as such individuals are labeled and treated as deviants, they constitute a population which must be explained in any theory of deviance. In this paper, however, we are primarily concerned with the theoretical relevance of official statistics for the study of deviance.

16. In any empirical investigation of such criteria it is necessary to distinguish between the formal (official) interpretive rules (as defined by a manual of procedures, constitution, and the like) which are to be employed by the personnel of the organizations in question, and the unofficial rules used by the personnel in their deviant-processing activities e.g., differential treatment of the basis of social class, race, ethnicity, or varying conceptions of "deviant" behavior.

17. The material in this section is taken from an unpublished paper by Cicourel entitled "Social Class, Family Structure and the Administration of Juvenile Justice," and is based on a study of the social organization of juvenile justice in two Southern California communities with populations of approximately 100,000 each.

18. See Donald J. Newman, "The Effects of Accommodations in Justice Administration on Criminal Statistics, Sociology and Social Research 46 (Jan., 1962), pp. 144-155; Administration of Criminal Justice, Chicago: American Bar Foundation, 1955, unpublished.

19 Newman, "The Effects of Accommodations. . . ," op. cit., pp. 145-146.

20 See ibid., pp. 146-151, and the references cited.

21 Juvenile Delinquency, New York: McGraw-Hill, 1949.22 Report of the Governor's Special Study Commission on Juvenile Justice, Parts I and II, Sacramento: California State Printing Office, 1960.23 To illustrate how organizational procedures and imputations can affect official statistics, we refer to a preliminary finding by (cited in footnote 17) which shows that one of two communities studied (Community A) has both a slightly larger population and a higher adult crime rate. Yet this community had (as of November, 1962) 3200 current cases of juveniles suspected or confirmed to be offenders. Community B. on the other hand, had approximately 8000 current suspected or confirmed juvenile cases. Community A has two juvenile officers on its staff, while Community B has five juvenile officers.