1964 Book DelinquencyCrimeAndDifferentia

DELINQUENCY, CRIME
AND
DIFFERENTIAL ASSOCIATION
DELINQUENCY, CRIME
AND
by
DONALD R. CRESSEY, PH. D .
Professor of Sociology
University of California, Santa Barbara
II
THE HAGUE
MARTINUS NI]HOFF
1964
ISBN 978-94-011-8336-9 ISBN 978-94-011-9015-2 (eBook)
DOl 10.1007/978-94-011-9015-2
Copyright 1964 by Martinus Nijhoff, The Hague, Netherlands

All rights reserved, including the right to translate or to
reproduce this book or parts thereof in any form
PREFACE
This is a book about Edwin H. Sutherland's theory of differ-

ential association. I received my Ph.D. from Indiana University,
where I worked with Sutherland, and the volume is made up
principally of my writings on differential association during the
years 1952-1963. However, the volume is neither a festschrift nor
a book of reprints. The original materials have in most cases been
quite severely edited in order to give the volume coherence and in
order to minimize repetition and redundancy. For example,
portions of one journal article appear in Chapters I, IV and V;
parts of a chapter published in a recent book appear in Chapters
I, II and III; and Chapter IX is composed of two inter-related
articles, published eight years apart. Chapter I has not appeared
elsewhere in its present form, but most of it consists of snippets
culled from several of my articles and books and woven together
in new form.
The book is intended primarily for non-American readers,
who on the whole are not as familiar with Sutherland's theory
(or with other sociological and social psychological theories about
delinquency, crime and corrections) as are Americans. Yet at
least a nodding acquaintance with Sutherland's work is becoming
increasingly necessary to an intelligent reading of the American
literature in criminology. For instance, American authors are
beginning to refer to "the theory of differential association", or
just "differential association", without citing Sutherland or any-
one else, apparently on the assumption that readers are familiar
with the theory and its origins. 1 It was with this idea in mind that
Professor Leon Radzinowicz, Director of the Institute of Crimi-
nology at the University of Cambridge, encouraged me to as-
1 See, for example, AlbertJ. Reiss, Jr. and Albert Lewis Rhodes, "Delinquency and
Social Class Structure", American Sociological Review, 26 (October, 1961), pp. 720-732.
VI PREFACE
semble the volume during my 1961-1962 year as Visiting Fellow

at the Institute. His recent survey of criminological research and
teaching in Europe and the United States revealed an appalling
lack of consensus among criminologists regarding even such
fundamental problems as definition of subject matter. 2 A nuclear
physicist working in a laboratory in any nation in the world
could quite easily (barring political considerations) transfer his
work to a laboratory in another nation, quickly integrating his
research with that of the men in the new laboratory. This is not
true of criminologists. The research being conducted in one
centre or institute may be completely unrelated to the research
being conducted in another, principally because there are wide
variations in the basic definitions of the subject matter of crimi-
nology, in conceptions of criminology's objectives, and in the
"basic" theory about the etiology of crime and delinquency and
the reformation of criminals. To some degree, this lack of con-
sensus is due to lack of familiarity with the work of men in other
nations, and to some degree this lack of familiarity, in turn, is
due to the fact that criminological works tend to be published,
especially by Americans, in a wide range of legal, psychological,
sociological and correctional journals, some of them quite ob-
scure.
We do not expect this volume to correct these deficiencies in
criminology, but we hope it will increase criminological con-
sensus by bringing together in one place some comments on and
criticisms of one of the most important ideas put forth by an
American criminologist. Sutherland's theory of differential as-
sociation anticipated much of the current sociological and social
psychological thought in criminology and I believe it is fair to say
that he, more than any other single person, was responsible for
such theoretical and methodological consensus as now exists
among the sociologists and social psychologists who give their
professional attention to problems in the field of criminology.
This volume will at least make his theory more accessible than
it has been. We hope, also, that our comments on the theory and
our criticisms of it will lead to a more widespread concern for its
mi plications than has been the case in the past, especially among
Europeans.
2 Leon Radzinowicz, In Search o/Criminology, London: Heinemann, 1961.
PREFACE VII
We wish to express our thanks to the following publishers,

for permission to reprint. To the J.B. Lippincott Company for
the material from Edwin H. Sutherland, Principles of Criminology,
Third Edition, 1939, and from Edwin H. Sutherland and
Donald R. Cressey, Principles of Criminology, Sixth Edition, 1960,
appearing here in Chapter I, pp. 4-5, 6-11 and 12-19. To
the editors of Social Problems for the material from "The Theory of
Differential Association: An Introduction", 8 (Summer, 1960),
pp. 2-6, appearing here in Chapter I, pp. 3-4, 6 and 11-12. To
Harcourt Brace and World for material from Robert K. Merton
and Robert A. Nisbet, Editors, Contemporary Social Problems, 1961,
appearing here in Chapter I, pp. 21-33; Chapter II, pp. 32-34
and 36-41; and Chapter III, pp. 50-66. To the editors of the
Pacific Sociological Review for the material from "Epidemiology and
Individual Conduct: A Case from Criminology", 3 (Fall, 1960),
pp. 47-48, appearing here in Chapter I, pp. 24-29; Chapter IV,
pp. 67-78; and Chapter V, pp. 81-89. To the editors of Crime
and Delinquency (formerly National Probation and Parole Association
Journal) for the material from "The State of Criminal Statistics",
National Probation and Parole Association Journal, 3 (July, 1957),
pp. 230-241, appearing here in Chapter II, pp. 34-36 and
41-49. To the editors of the Journal of Criminal Law, Criminology
and Police Science for the material from "The Differential As-
sociation Theory and Compulsive Crimes", 45 (June, 1954),
pp. 29-40, appearing here as Chapter VI, pp. 90-107; and
for the material from "Application and Verification of the
Differential Association Theory", 43 (May-J une, 1952), pp.
43-52, appearing here as Chapter VII, pp. 108-118. These two
articles were copyrighted © in 1954 and 1952 by the North-
western University School of Law. To the editors of Federal
Probation for the material from "Contradictory Directives in
Correctional Group Therapy Programs", 23 (June, 1954), pp.
20-26, appearing here in Chapter VIII, pp. 127-138. To the
University of Chicago Press for the material from "Changing
Criminals: The Application of the Theory of Differential As-
sociation", American Journal of Sociology, 61 (September, 1955),
pp. 116-120, appearing here in Chapter VIII, pp. 123-126
and in Chapter IX, pp. 139-143; and for the material from
"Differential Association and the Rehabilitation ofDrug Addicts",
VIII PREFACE
American Journal oj Sociology, 69 (September, 1963), pp. 129-142

(with Rita Volkman), appearing here in Chapter IX, pp. 143-164.
Santa Barbara, California DONALD R. CRESSEY
January, 1964
TABLE OF CONTENTS
Preface V
AN INTRODUCTION TO THE THEORY OF DIFFERENTIAL ASSOCIATION
I. A Statement of the Theory. . . . . . . . . 3

II. Some Obstacles to Generalizing in Criminology 30
III. Some Facts a Theory Must Fit . . . . . . . 50
IV. Differential Association, Differential Social Organi-
zation, and the Epidemiology of Crime. . . . . . 67
II
SOME CRITICISMS OF THE THEORY
V. Some Popular Criticism of Differential Association. 81

VI. Differential Association and Compulsive Crimes 90
VII. Differential Association and Trust Violation. . . . 108
III
USE OF THE THEORY IN CORRECTIONS
VIII. Differential Association and Correctional Group

Therapy. . . . . . . . . . . . . . . 123
IX. Differential Association and Rehabilitation . . . . 139
Index 165
I
AN INTRODUCTION TO THE THEORY

OF DIFFERENTIAL ASSOCIATION
CHAPTER I
A STATEMENT OF THE THEORY
The first formal statement of Edwin H. Sutherland's theory

of differential association appeared in the third edition of his
Principles if Criminology, in 1939. Sutherland later pointed out
that the idea of differential association was included in an earlier
edition of the text, where it was stated that any person can be
trained to adopt and follow any pattern of behaviour which he is
able to execute, that failure to follow a prescribed pattern of
behaviour is due to inconsistencies in the culture, and that "culture
conflict" is therefore the fundamental condition to be considered
in any explanation of crime. 1 He confessed that he was unaware
that this statement was a general theory of criminal behaviour. 2
He had written the first edition of the book because the head of
his department had asked him to do so, thus launching a dis-
tinguished career in criminology. However, his principal theo-
retical interest at the time was one he shared with other crimi-
nologists-the controversy between heredity and environment.
In preparing his textbook, he reviewed much of the writing on
crime and criminals, especially the research studies, and organ-
ized the results topically-economic factors, political factors,
physical factors, etc.-rather than according to a theoretical
system. At the time, he made no effort to generalize about the
various facts that he and other scholars had discovered about
crime. Instead, he took pride in his "broadminded" lack of
concern for any of the contradictory theories then in vogue.
He also took the view that criminology was at most a synthetic
science, an organization of the research findings of disciplines
1 Edwin H. Sutherland, Principles of Criminology, Second Edition, Philadelphia:
Lippincott, 1934, pp. 51-52.
2 Edwin H. Sutherland, "Development of the Theory", in Albert Cohen, Alfred
Lindesmith and Karl Schuessler, Editors, The Sutherland Papers, Bloomington, Indiana
University Press, 1956, p. 16.
4 INTRODUCTION TO THE THEORY
such as economics, psychology and biology. It was his efforts to

apply sociological concepts to criminal behaviour that led him,
unwittingly, to the statement in the second edition of his book.
Especially important were Thomas's writings on "attitude-value",
"four wishes", "isolation" and "culture-conflict", and Park and
Burgess's writings on the processes of assimilation, accommo-
dation, competition and cooperation.
At the insistence of his colleagues, Sutherland drew up a formal
set of propositions based on the 1934 statement and appended it
to the third edition of his textbook. Michael and Adler had
published their critical appraisal of criminological research, 3
and this had turned Sutherland's attention toward a concern for
making generalizations about the facts regarding crime and
criminals. The "culture conflict" concept was coming into vogue,
and Sutherland worked with Sellin on the Social Science Re-
search Council Bulletin which organized some of the data of
criminology around that concept. 4 Henry McKay, Hans Riemer,
Harvey Locke, Lowell Sweetser and Alfred R. Lindesmith were
among the colleagues who prodded Sutherland into stating his
sociological ideas about criminal behaviour in a formal manner.
Lindesmith was especially influential, for in his research on
drug addiction he had insisted on a methodology aimed at the
production of universal generalizations, rather than at the pro-
duction of statistical summaries. 5 The method involves case
studies, directed by explicit hypotheses, of rigorously defined
categories of behaviour. The procedure has essentially the follow-
ing steps. First, a rough definition of the behaviour to be ex-
plained is formulated. Second, a hypothetical explanation of the
behaviour is formulated. Third, one case is studied in the light
of the hypothesis with the object of determining whether the
hypothesis fits the facts in that case. Fourth, if the hypothesis
does not fit the facts, either the hypothesis is reformulated or the
3 J. Michael and M. J. Adler, Crime, Law and Social Science, New York: Harcourt,
Brace, 1933.
• Thorsten Sellin, Culture Coriflict and Crime, New York: Social Science Research
Council, 1938. In the preface to this volume, Mark A. May says "Professor Sellin
wishes to record here his appreciation to his colleague on the subcommittee, Professor
Sutherland, who during the entire period that this monograph has been in preparation
has assisted with his wise counsel".
5 Alfred R. Lindesmith, Opiate Addiction, Bloomington: Principia Press, 1947.
Publication of this book was delayed about ten years by the war.
STATEMENT 5
behaviour to be explained is redefined, so that the case is ex-

cluded. This definition must be more precise than the first one,
and it may not be formulated solely to exclude a negative case.
The negative case is viewed as a sign that something is wrong
with the hypothesis, and redefinition takes place so that the cases
of behaviour being explained will be homogeneous. Fifth, practi-
cal certainty may be attained after a small number of cases have
been examined in this way, but the location by the investigator,
or anyone else, of a negative case disproves the explanation and
requires a reformulation. Sixth, this procedure of examining
cases, redefining the behaviour and reformulating the hypothesis,
is continued until a universal relationship is established, each
negative case calling for a redefinition or a reformulation. The
negative case-that is, the one which does not fit the hypothesis-
is the important point in the procedure for it calls for redefinition
or reformulation. Seventh, for purposes of proof, cases outside the
area circumscribed by the definition are examined to make
certain that the final hypothesis does not apply to them. This
step is in keeping with the observation that scientific generali-
zations consist of descriptions of conditions which are always
present when the phenomenon being explained is present but
which are never present when the phenomenon is absent. 6
This notion of proper procedures in social science research
insists, when applied to the study ofdelinquents, that the objective
should be to separate the behaviour of delinquents from the
behaviour of non-delinquents on the basis of causal process. It
was this objective that Sutherland sought when he published the
1939 version of his theory: "It was my conception that a general
theory should take account of all the factual information regard-
ing crime causation. It does this either by organizing the multiple
factors in relation to each other or by abstracting them from
certain common elements. It does not, or should not, neglect or
eliminate any factors that are included in the multiple factor
theory."7
In one sense, the formal statement appearing in the third
6 This summary has been taken from Edwin H. Sutherland and Donald R. Cressey,
Principles of Criminology, Sixth Edition, Philadelphia: Lippincott, 1960, pp. 68-69.
See also Donald R. Cressey, Other People's Money: A Stutfy of the Social Psyclwlogy of
Emblezzlement, Glencoe: The Free Press, 1953.
7 "Development of the Theory", Ope cit., p. 18.
edition of Principles of Criminology was short lived. For reasons

which have never been very clear, the statement of the theory
was qualified so that it pertained to only "systematic criminal
behaviour", rather than to the more general category, "criminal
behaviour". The theory was stated in the form of the following
seven propositions. 8
"First, the processes which result in systematic criminal-behaviour are

fundamentally the same in form as the processes which result in systematic
lawful behaviour. If criminality were specifically determined by
inheritance, the laws and principles of inheritance would be the
same for criminal behaviour and for lawful behaviour. The same
is true of imitation or any other genetic process in the development
of behaviour. Criminal behaviour differs from lawful behaviour
in the standards by which it is judged but not in the principles
of the genetic processes.
"Second, systematic criminal behaviour is determined in a process of

association with those who commit crimes, just as systematic lawful
behaviour is determined in a process of association with those who are
law-abiding. Any person can learn any pattern of behaviour which
he is able to execute. He inevitably assimilates such behaviour
from the surrounding culture. The pattern of behaviour may
cause him to suffer death, physical injury, loss of friendship, or
loss of money, but it may nevertheless be followed with joy
provided he has learned that it is the thing to do. Since criminal
behaviour is thus developed in association with criminals it
means that crime is the cause of crime. In the same manner war
is the cause of war, and the Southern practice of dropping the
"r" is the cause of the Southern practice of dropping the "r".
This proposition, stated negatively, is that a person does not
participate in systematic criminal behaviour by inheritance. No
individual inherits tendencies which inevitably make him crimi-
nal or inevitably make him law abiding. Also, the person who is
not already trained in crime does not invent systematic criminal
behaviour. While personality certainly includes an element of
inventiveness, a person does not invent a system of criminal
8 Edwin H. Sutherland, Principles of Criminology, Third Edition, Philadelphia:
Lippincott, 1939, pp. 4-9.
STATEMENT 7
behaviour unless he has had training in that kind of behaviour,

just as a person does not make systematic mechanical inventions
unless he has had training in mechanics.
"Third, dijferential association is the specific causal process in the

development oj systematic criminal behaviour. The principles of the
process of association by which criminal behaviour develops are
the same as the principles of the process by which lawful be-
haviour develops, but the contents of the patterns presented in
association differ. For that reason it is called differential associ-
ation. The association which is of primary importance in criminal
behaviour is association with persons who engage in systematic
criminal behaviour. A person who has never heard of professional
shoplifting may meet a professional shoplifter in his hotel, may
become acquainted with and like him, learn from him the
techniques, values, and codes of shoplifting, and under this
tutelage may become a professional shoplifter. He could not
become a professional shoplifter by reading newspapers, maga-
zines or books. The impersonal agencies of communication exert
some influence but are important principally in determining
receptivity to the patterns of criminal behaviour when they are
presented in personal association, and in producing incidental
offences. These patterns are presented through the impersonal
agencies of communication to everyone in our culture. Every
child capable of learning inevitably assimilates knowledge re-
garding property rights and thefts in the simpler situations. It is
probably for this reason that everyone is somewhat criminal.
College students, with a few exceptions doubtless due to poor
memories, report an average of eight thefts or series of thefts
during their lifetimes; a series of thefts in this case may include
scores of incidents, such as stealing fruit from neighbours' trees
from the age of seven to twelve. These thefts were reported
equally for males and females, and continued in most cases to
the age at which the reports were made. In the later years they
generally took the form of theft of books from the library, of
equipment from the gymnasium or laboratory, or of souvenirs
from hotels and restaurants. Students do not regard such thefts as
especially reprehensible; they regard them as amusing. Similarly,
boys in the delinquent areas of cities do not regard thefts of
automobiles or the burglary of stores as reprehensible, and

business or professional men do not regard their frauds and tricky
manipulations as reprehensible. A person engages in those
criIninal acts which are prevalent in his own groups, and he
assimilates them in association with the members of the groups.
"Fourth, the chance that a person will participate in systematic criminal

behaviour is determined roughly by the frequency and consistency of his
contacts with the patterns of criminal behaviour. If a person could come
into contact only with lawful behaviour he would inevitably be
completely law-abiding. If he could come into contact only with
criminal behaviour (which is impossible, since no group could
exist if all of its behaviour were criminal) he would inevitably be
completely criminal. The actual condition is between these
extremes. The ratio of criminal acts to lawful acts by a person is
roughly the same as the ratio of the contacts with the criminal
and with the lawful behaviour of others. It is true, of course, that
a single critical experience may be the turning point in a career.
But these critical experiences are generally based on a long series
of former experiences and they produce their effects generally
because they change the person's associations. One of these
critical experiences that is most important in determining crimi-
nal careers is the first public appearance as a criminal. A boy
who is arrested and convicted is thereby publicly defined as a
criminal. Thereafter his associations with lawful people are
restricted and he is thrown into association with other delinquents.
On the other hand a person who is consistently criminal is not
defined as law-abiding by a single lawful act. Every person is
expected to be law-abiding, and lawful behaviour is taken for
granted because the lawful culture is dominant, more extensive
and more pervasive than the criminal culture.
"Fifth, individual differences among people in respect to personal

characteristics or social situations cause crime only as they affect differential
association or frequency and consistency of contacts with criminal patterns.
Poverty in the home may force a family to reside in a low-rent
area where delinquency rates are high and thereby facilitate
association with delinquents. Parents who insist that their boy
return home immediately after school and who are able to enforce
STATEMENT 9
this regulation may prevent the boy from coming into frequent
contact with delinquents even though the family resides in a high
delinquency area. A child who is not wanted at home may be
emotionally upset, but the significant thing is that this condition
may drive him away from the home and he may therefore come
into contact with delinquents. A boy who is timid may be kept
from association with rough delinquents. It is not necessary to
assume a generic difference between persons by reason of which
some are generally receptive to criminality and others not
receptive. Such an assumption would be far-fetched and un-
justified. There may be receptivity at a particular moment to a
particular stimulation, but the elements are so complex that no
generalization regarding such receptivity is possible. The closest
approach to a generalization is to say that this specific receptivity
is determined principally by the frequency and consistency of
previous contacts with patterns of delinquency and that beyond
this the delinquent behaviour is adventitious.
"Sixth, cultural conflict is the underlying cause of dijJerential association

and therefore of systematic criminal behaviour. Differential association
is possible because society is composed of various groups with
varied cultures. These differences in culture are found in respect
to many values and are generally regarded as desirable. They
exist, also, with reference to the values which the laws are de-
signed to protect, and in that form are generally regarded as un-
desirable. This criminal culture is as real as lawful culture and is
much more prevalent than is usually believed. It is not confined
to the hoodlums in slums or to professional criminals. Prisoners
frequently state and undoubtedly believe that they are no worse
than the majority of people on the outside. The more intricate
manipulations of business and professional men may be kept
within the letter of the law as interpreted but be identical in logic
and effects with the criminal behaviour which results in imprison-
ment. These practices, even if they do not result in public condem-
nation as crimes, are a part of the criminal culture. The more the
cultural patterns conflict, the more inpredictable is the behaviour
of a particular person. It was possible to predict with almost
complete certainty how a person reared in a Chinese village fifty
years ago would behave because there was only one way for him
to behave. The attempts to explain the behaviour of a particular

person in a modern city have been rather unproductive because
the influences are in conflict and any particular influence may be
very evanescent.
"Seventh, social disorganization is the basic cause qf systematic criminal

behaviour. The origin and the persistence of culture conflicts
relating to the values expressed in the law and of differential
association which is based on the cultural conflicts are due to
social disorganization. Cultural conflict is a specific aspect of
social disorganization and in that sense the two concepts are
names for smaller and larger aspects of the same thing. But social
disorganizations is important in another sense. Since the law-
abiding culture is dominant and more extensive, it could over-
come systematic crime if organized for that purpose. But society
is organized around individual and small group interests on most
points. A law-abiding person is more interested in his own
immediate personal projects than in abstract social welfare or
justice. In this sense society permits crime to persist in systematic
form. Consequently systematic crime persists not only because of
differential association but also because of the reaction of the
general society toward such crime. When a society or a smaller
group develops a unified interest in crimes which touch its
fundamental and common values, it generally succeeds in
eliminating or at least greatly reducing crime. This occurred,
for instance, when baseball players in the world series took bribes
for throwing away a game they could have won. This affected so
many people in a manner which they regarded as vital, and they
reacted in such evident opposition, that the crime, so far as is
known, has never been repeated. Also, when many wealthy
people were kidnapped and held for ransom at the end of the
prohibition period, our society reorganized the legal and adminis-
trative system in violation of the slogans and myth of state
sovereignty and such kidnappings practically ceased. However,
in previous times when poor and helpless people were the victims
of kidnappings, as in the slave trade, impressment of sailors,
shanghaiing of sailors by crimps, and unjustifiable arrests, it took
generations and in some cases centuries for society to become
sufficiently aware and interested to stop kidnappings in those
STATEMENT 11
forms. When a gang starts in a disorganized district of a city it

keeps growing and other gangs develop. But when a delinquent
gang started on a business street adjacent to Hyde Park, a good
residential district in Chicago, the residents became concerned,
formed an organization, and decided that the best way to protect
themselves was by providing a club house and recreational
facilities for the delinquents. This practically eliminated the
gangs. Therefore, whether systematic delinquency does or does
not develop is determined not only be associations that people
make with the criminals, but also by the reactions of the rest of
society toward systematic criminal behaviour. If the society is
organized with reference to the values expressed in the law, the
crime is eliminated; if it is not organized, crime persists and
develops. The opposition of the society may take the form of
punishment, of reformation, or of prevention.
"The general theory which has been presented may be summa-
rized in the following statements: Systematic criminal behaviour
is due immediately to differential association in a situation in
which cultural conflicts exist, and ultimately to the social dis-
organization in that situation. A specific or incidental crime of a
particular person is due generally to the same process, but it is
not possible to include all cases because of the adventitious
character of delinquency when regarded as specific or incidental
acts."
It should also be noted that the above statement was some-

what redundant, for it proposed generally that criminal be-
haviour is learned in a process of differential association with
those who commit crime and those who are law-abiding, but it
then went on in propositions four and five to use "consistency"
of associations as one of the conditions affecting the impact of
differential association on individuals. Thus, "consistency" of
behaviour patterns presented was used as a general explanation
of systematic criminality, but "consistency" also was used to
describe the process by which differential association takes place.
Sutherland recognized these errors almost immediately, but
not soon enough to omit them from the first published statement
of his theory. He deleted the word "systematic" from the second
version, and he later explained that it was his belief that all but
"the very trivial criminal acts" were "systematic," but that he

deleted the word because some research workers were unable to
identify "systematic criminals," and other workers considered
only an insignificant proportion of prisoners to be "systematic
criminals."9 Limitation to "systematic criminality" was made
for what seemed to be practical rather than logical reasons,
and it was abandoned when it did not seem to have practical
utility.
The word "consistency" also was deleted from Sutherland's
second statement, which first appeared in the Fourth Edition of
his textbook (1947), was carried in the Fifth Edition, which I
brought out in 1955, and was continued in the Sixth Edition,
published in 1960. The following 71 / 2 pages constitute a repro-
duction of Chapter IV, "A Sociological Theory of Criminal
Behaviour," from the Sixth Edition of Principles if Criminology.
While I am the author of some of the introductory material and
some of the closing material, the section entitled "Genetic Expla-
nation of Criminal Behaviour" is carried in the form in which it
appeared in the 1947 edition.
The Problem for Criminological Theory. A scientific explanation
consists of a description of the conditions which are always present
when a phenomenon occurs and which are never present when
the phenomenon does not occur. Although a multitude of con-
ditions may be associated in greater or less degree with the
phenomenon in question, this information is relatively useless
for understanding or for control if the factors are left as a hodge-
podge of unorganized factors. Scientists strive to organize their
knowledge in interrelated general propositions, to which no
exceptions can be found. If criminology is to be scientific, the
heterogeneous collection of "multiple factors" known to be
associated with crime and criminality should be organized and
integrated by means of an explanatory theory which has the
same characteristics as the scientific explanations in other fields
of study. That is, the conditions which are said to cause crime
should always be present when crime is present, and they should
always be absent when crime is absent. Such a theory would
stimulate, simplify, and give direction to criminological re-
search and it would provide a framework for understanding the
9 "Development of the Theory", op. cit., p. 2 J.
STATEMENT 13
significance of much of the knowledge acquired about crime and

criminality in the past. Furthermore, it would be useful in control
of crime, providing it could be "applied" in much the same way
that the engineer "applies" the scientific theories of the physicist.
There are two complementary procedures which may be used
to put order into criminological knowledge, to develop a causal
theory of criminal behavior. The first is logical abstraction.
Negroes, urban-dwellers and young adults all have comparatively
high crime rates. What do they have in common that results in
these high crime rates? Research studies have shown that criminal
behaviour is associated in greater or less degree with the social
and personal pathologies, such as poverty, bad housing, slum-
residence, lack of recreational facilities, inadequate and demoral-
ized families, feeble-mindedness, emotional instability, and other
traits and conditions. What do these conditions have is common
which apparently produces excessive criminality? Research
studies have also demonstrated that many persons with those
pathological traits and conditions do not commit crimes and that
persons in the upper socio-economic class frequently violate the
law, although they are not in poverty, do not lack recreational
facilities and are not feeble-minded or emotionally instable.
Obviously, it is not the conditions or traits themselves which
cause crime, for the conditions are sometimes present when
criminality does not occur, and they also are sometimes absent
when criminality does occur. A causal explanation of criminal
behaviour can be reached by abstracting, logically, the mecha-
nisms and processes which are common to the rich and the poor,
Negroes and whites, urban- and the rural-dwellers, young adults
and old adults, and the emotionally stable and the emotionally
unstable who commit crimes.
In arriving at these abstract mechanisms and processes, crimi-
nal behaviour must be precisely defined and carefully dis-
tinguished from non-criminal behaviour. The problem in crimi-
nology is to explain the criminality of behaviour, not behaviour
as such. The abstract mechanisms and processes common to the
classes of criminals indicated above should not also be common to
non-criminals. Criminal behaviour is human behaviour, has
much in common with non-criminal behaviour and must be
explained with the same general framework used to explain other
human behaviour. However, an explanation of criminal be-

haviour should be a specific part of a general theory of be-
haviour. Its specific task should be to differentiate criminal from
non-criminal behaviour. Many things which are necessary for
behaviour are not for that reason important to the criminality of
behaviour. Respiration, for instance, is necessary for any be-
haviour, but the respiratory process cannot be used in an expla-
nation of criminal behaviour, for it does not differentiate criminal
behaviour from non-criminal behaviour.
The second procedure for putting order into criminological
knowledge is differentiation oflevels of analysis. This means that
the problem is limited to a particular part of the whole situation,
largely in terms of chronology. The causal analysis must be held
at a particular level. For example, when physicists stated the law
offalling bodies they were not concerned with the reasons why a
body began to fall except as this might affect the initial mo-
mentum. It made no difference to the physicist whether a body
began to fall because it was dropped from the hand of an experi-
mental physicist or rolled off the edge of a bridge because of
vibration caused by a passing vehicle. Also, a round object would
have rolled off the bridge more readily than a square object, but
this fact was not significant for the law of falling bodies. Such
facts were considered as existing on a different level ofexplanation
and were irrelevant to the problem with which the physicists
were concerned. Much of the confusion regarding criminal be-
haviour is due to a failure to define and hold constant the level
of explanation. By analogy, many criminologists would attribute
some degree of causal power to the "roundness" of the object in
the illustration above. However, consideration of time sequences
among the factors associated with crime and criminality may
lead to simplicity of statement. In the heterogeneous collection
of factors associated with criminal behaviour, one factor often
occurs prior to another factor (in much the way that "roundness"
occurs prior to "vibration," and "vibration" occurs prior to
"rolling off a bridge)," but a theoretical statement about criminal
behaviour can be made without referring to those early factors.
By holding the analysis at one level, the early factors are combined
with or differentiated from later factors or conditions, thus reducing
the number of variables which must be considered in a theory.
STATEMENT 15
A motion picture several years ago showed two boys engaged

in a minor theft; they ran when they were discovered; one boy
had longer legs, escaped, and became a priest, the other had
shorter legs, was caught, committed to a reformatory, and be-
came a gangster. In this comparison, the boy who became a
criminal was differentiated from the one who did not become a
criminal by the length of his legs. But "length oflegs" need not be
considered in a criminological theory for, in general, no signifi-
cant relationship has been found between criminality and length
oflegs and certainly many persons with short legs are law-abiding
and many persons with long legs are criminals. The length of the
legs does not determine criminality and has no necessary relation
to criminality. In the illustration, the differential in the length of
the boys' legs may be observed to be significant to subsequent
criminality or non-criminality only to the degree that it de-
termined the subsequent experiences and associations of the two
boys. It is in these experiences and associations, then, that the
mechanisms and processes which are important to criminality or
non-criminality are to be found. A "one-level" theoretical expla-
nation of crime would be concerned solely with these mechanisms
and processes, not with the earlier factor "length of legs".
Two Types of Explanations qf Criminal Behaviour. Scientific

explanations of criminal behaviour may be stated either in terms
of the processes which are operating at the moment of the
occurrence of crime or in terms of the processes operating in the
earlier history of the criminal. In the first case the explanation
may be called "mechanistic," "situational," or "dynamic"; in the
second, "historical" or "genetic." Both types of explanation are
desirable. The mechanistic type of explanation has been favoured
by physical and biological scientists and it probably could be the
more efficient type of explanation of criminal behaviour. How-
ever, criminological explanations of the mechanistic type have
thus far been notably unsuccessful, perhaps largely because they
have been formulated in connection with the attempt to isolate
personal and social pathologies among criminals. Work from this
point of view has, at least, resulted in the conclusion that the
immediate determinants of criminal behaviour lie in the person-
situation complex.
The objective situation is important to criminality largely to

the extent that it provides an opportunity for a criminal act.
A thief may steal from a fruit stand when the owner is not in sight
but refrain when the owner is in sight; a bank burglar may attack
a bank which is poorly protected but refrain from attacking a
bank protected by watchmen and burglar alarms. A corporation
which manufactures automobiles seldom or never violates the
Pure Food and Drug Law, but a meat-packing corporation might
violate this law with great frequency. But in another sense, a
psychological or sociological sense, the situation is not exclusive
of the person, for the situation which is important is the situation
as defined by the person who is involved. That is, some persons
define a situation in which a fruit-stand owner is out of sight as a
"crime-committing" situation, while others do not so define it.
Furthermore, the events in the person-situation complex at the
time a crime occurs cannot be separated from the prior life
experiences of the criminal. This means that the situation is
defined by the person in terms of the inclinations and abilities
which the person has acquired up to date. For example, while a
person could define a situation in such a manner that criminal
behaviour would be the inevitable result, his past experiences
would for the most part determine the way in which he defined
the situation. An explanation of criminal behaviour made in
terms of these past experiences is an historical or genetic expla-
nation.
The following paragraphs state such a genetic theory of
criminal behaviour on the assumption that a criminal act occurs
when a situation appropriate for it, as defined by the person, is
present. The theory should be regarded as tentative, and it should
be tested by the factual information presented in the later chap-
ters and by all other factual information and theories which are
applicable.
Genetic Explanation of Criminal Behaviour. The following state-

ment refers to the process by which a particular person comes to
engage in criminal behaviour.
( 1) Criminal behaviour is learned. Negatively, this means that
criminal behaviour is not inherited, as such; also, the person who
is not already trained in crime does not invent criminal be-
STATEMENT 17
haviour, just as a person does not make mechanical inventions

unless he has had training in mechanics.
(2) Criminal behaviour is learned in interaction with other persons in a
process of communication. This communication is verbal in many
respects but includes also "the communication of gestures."
(3) The principal part of the learning of criminal behaviour occurs within
intimate personal groups. Negatively, this means that the impersonal
agencies of communication, such as movies and newspapers, play
a relatively unimportant part in the genesis of criminal behaviour.
(4) When criminal behaviour is learned, the learning includes (a) tech-
niques of committing the crime, which are sometimes very complicated,
sometimes very simple; (b) the specific direction of motives, drives,
rationalizations, and attitudes.
(5) The specific direction of motives and drives is learned from definition
of the legal codes as favourable or unfavourable. In some societies an
individual is surrounded by persons who invariably define the
legal codes as rules to be observed, while in others he is sur-
rounded by persons whose definitions are favourable to the
violation of the legal codes. In our American society these
definitions are almost always mixed, with the consequence that
we have culture conflict in relation to the legal codes.
(6) A person becomes delinquent because of an excess of definitions favour-
able to violation of law over definitions unfavourable to violation of law.
This is the principle of differential association. It refers to both
criminal and anti-criminal associations and has to do with
counteracting forces. When persons become criminal, they do so
because of contacts with criminal patterns and also because of
isolation from anti-criminal patterns. Any person inevitably
assimilates the surrounding culture unless other patterns are in
conflict; a Southerner does not pronounce "r" because other
Southerners do not pronounce "r." Negatively, this proposition
of differential association means that associations which are
neutral so far as crime is concerned have little or no effect on the
genesis of criminal behaviour. Much of the experience of a person
is neutral in this sense, e.g. learning to brush one's teeth. This
behaviour has no negative or positive effect on criminal be-
haviour except as it may be related to associations which are
concerned with the legal codes. This neutral behaviour is im-
portant especially as an occupier of the time of a child so that he
is not in contact with criminal behaviour during the time he is so

engaged in the neutral behaviour.
(7) Differential associations may vary infrequency, duration, priority and
intensity. This means that association with criminal behaviour
and also associations with anti-criminal behaviour vary in those
respects. "Frequency" and "duration" as modalities of associ-
ations are obvious and need no explanation. "Priority" is assumed
to be important in the sense that lawful behaviour developed in
early childhood may persist throughout life, and also that de-
linquent behaviour developed in early childhood may persist
throughout life. This tendency, however, has not been adequately
demonstrated and priority seems to be important principally
through its selective influence. "Intensity" is not precisely defined
but it has to do with such things as the prestige of the source of a
criminal or anti-criminal pattern and with emotional reactions
related to the associations. In a precise description of the criminal
behaviour of a person these modalities would be stated in quanti-
tative form and a mathematical ratio be reached. A formula in
this sense has not been developed and the development of such a
formula would be extremely difficult.
(8) The process qf learning criminal behaviour by association with
criminal and anti-criminal patterns involves all if the mechanisms that are
involved in any other learning. Negatively, this means that the learn-
ing of criminal behaviour is not restricted to the process of
imitation. A person who is seduced, for instance, learns criminal
behaviour by association, but this process would not ordinarily be
described as imitation.
(9) While criminal behaviour is an expression of general needs and values,
it is not explained by those general needs and values since non-criminal
behaviour is an expression of the same needs and values. Thieves generally
steal in order to secure money, but likewise honest labourers work
in order to secure money. The attempts by many scholars to
explain criminal behaviour by general drives and values, such as
the happiness principle, striving for social status, the money
motive, or frustation, have been and must continue to be futile
since they explain lawful behaviour as completely as they explain
criminal behaviour. They are similar to respiration, which is
necessary for any behaviour but which does not differentiate
criminal from non-criminal behaviour.
STATEMENT 19
It is not necessary, at this level of explanation, to explain why

a person has the associations which he has; this certainly involves
a complex of many things. In an area where the delinquency rate
is high, a boy who is sociable, gregarious, active and athletic is
very likely to come in contact with the other boys in the neigh-
bourhood, learn delinquent behaviour from them and become a
gangster; in the same neighbourhood the psychopathic boy who is
isolated, introverted and inert may remain at home, not become
acquainted with the other boys in the neighbourhood and not
become delinquent. In another situation, the sociable, athletic,
aggressive boy may become a member of a scout troop and not
become involved in delinquent behaviour. The person's associ-
ations are determined in a general context of social organization.
A child is ordinarily reared in a family; the place of residence of
the family is determined largely by family income; and the
delinquency rate is in many respects related to the rental value
of the houses. Many other aspects of social organization affect the
kinds of associations a person has.
The preceding explanation of criminal behaviour purports to
explain the criminal and non-criminal behaviour of individual
persons. As indicated earlier, it is possible to state sociological
theories of criminal behaviour which explain the criminality of a
community, nation or other group. The problem, when thus
stated, is to account for variations in crime rates and involves a
comparison of the crime rates of various groups or the crime rates
of a particular group at different times. The explanation of a
crime rate must be consistent with the explanation of the criminal
behaviour of the person, since the crime rate is a summary state-
ment of the number of persons in the group who commit crimes
and the frequency with which they commit crimes. One of the
best explanations of crime rates from this point of view is that a
high crime rate is due to social disorganization. The term "social
disorganization" is not entirely satisfactory and it seems prefer-
able to substitute for it the term "differential social organization."
The postulate on which this theory is based, regardless of the
name, is that crime is rooted in the social organization and is an
expression of that social organization. A group may be organized
for criminal behaviour or organized against criminal behaviour
Most communities are organized both for criminal and anti-
criminal behaviour and in that sense the crime rate is an ex-

pression of the differential group organization. Differential group
organization as an explanation of variations in crime rates is
consistent with the differential association theory of the processes
by which persons become criminals.
Differential Association and Social Processes

A number of points in the above statement need to be empha-
sized. First, the statement holds, in essence, that delinquent or
criminal behaviour is learned in interaction with persons in a
pattern of communication, and that the specific direction of
motives, drives, rationalizations and attitudes-whether in the
direction of anti-criminality or criminality-is learned from
persons who define the codes as rules to be observed and from
persons whose attitudes are favourable to violation oflegal codes.
"A person becomes delinquent because of an excess of definitions
favourable to violation of law over definitions unfavourable to
violations oflaw."
Second, in any society, the two kinds of definitions of what is

desirable in reference to legal codes exist side by side, and a
person might present contradictory definitions to another person
at different times and in different situations. Sutherland called
the process of receiving these definitions "differential association"
because the content of what is learned in association with criminal
and delinquent behaviour patterns differs from the content ofwhat
is learned in association with anti-criminal and anti-delinquent
behaviour patterns. "When persons become criminals they do so
because of contacts with criminal behaviour patterns and also be-
cause of isolation from anti-criminal patterns." These contacts,
however, "may vary infrequency, duration, priority and intensity."
Third, when this idea is applied to a nation, a city or a group,
it becomes a sociological theory, rather than a social psychological
theory, for it deals with differential rates of crime and delinquency.
For example, a high crime rate in urban areas, as compared to
rural areas, can be considered an end product of a situation in
which a relatively large proportion of persons are presented with
an excess of criminal behaviour patterns. Similarly, the fact that
the rate for all delinquencies is not higher in some urban areas
STATEMENT 21
than it is in some rural areas can be attributed to differences in

probabilities of exposure to delinquency behaviour patterns. The
important general point is that in a multi-group type of social
organization, alternative and inconsistent standards of conduct
are possessed by various groups so that individuals who are
members of one group have a higher probability of learning to
use legal means for achieving success, or of learning to deny the
importance of success, while individuals in other groups learn to
accept the importance of success and to achieve it by illegal
means. Stated in another way, there are alternative educational
processes in operation, varying with groups, so that a person may
be educated in either conventional or criminal means of achieving
success. 10 Sutherland called this situation "differential social
organization" or "differential group organization," and he
proposed that "differential group organization should explain the
crime rate, while differential association should explain the
criminal behaviour or a person. The two explanations must be
consistent with each other,"ll
Fourth, it should be noted that the theory is consistent with
the process by which acts come to be defined as crimes. In an
earlier period criminal laws were enacted when almost all the
members of a society believed strongly that a particular kind of
conduct was so repulsive that formal machinery for dealing with
it should be set up. Thus, taboos against murder and theft were
considered so important that they became institutionalized; a
definite penalty for each offence was stipulated in advance,
certain officials (police) were designed to apprehend offenders,
other officials (judges, prosecutors) were designated to try
offenders and assign penalties to the guilty and still other officials
(jailers, wardens, executioners) were asked to carry out the
sentences. The action of such officials in reference to the offender
is "disinterested," in the sense that they act in the name of the
state rather than on the basis of any personal interest they might
have in the case. They are the persons whom the members of a
society have appointed to enforce the mores in a regular way, in
the name of all the people.
10 See Henry D. McKay, "The Neighbourhood and Child Conduct", Annals of the
American Academy of Political and Social Science, 261 (January, 1949), pp. 32-42.
11 "Development of the Theory", op. cit., p. 21.
The criminal laws established in more recent times, however,

do not have such a broad base in the society, for these are the
laws that institutionalize folkways.12 The same kinds of officials
are given authority to enforce the laws once they are established,
but the process of establishing them is different. Rather than a
codification of a widespread consensus that certain conduct is
morally wrong, these laws are likely to be codifications of the
consensus of only a rather small segment of the society. Laws
against operating a saloon on Sunday, discriminating against
Negroes in employment practices, practising certain professions
without having membership in the professional association,
selling merchandise above or below certain prices, building a
home without a certain kind of equipment, and a host of similar
laws, are examples of criminal laws promulgated by special
interest groups rather than by society as a whole. Even laws
specifically in the "public interest" are not necessarily initiated by
"the public." Laws against spitting in the floors of buses, for
example, are enacted because of pressure on legislators by non-
spitters, not the spitters themselves. Generally speaking, this
latter type of law is promoted by the middle class for the regu-
lation of working-class persons.
Even in reference to the most ancient of criminal laws, the
process of declaring an act to be crime involves a number of
persons who are opposed to this kind of behaviour, a struggle for
legislative or judicial dominance between the two groups, and
victory in this struggle by the persons opposing the behaviour.
For example, killing other persons became a crime only because
some persons persisted in the conduct, despite mores to the
contrary. Other members of society through their political
institution centralized and formalized the mechanisms for social
control of the conduct, and taboo against killing became the
"law of the land." Generally speaking, "breach is the mother of
law,"13 and a conflict between norms is thus reflected in the
criminal law. In complex societies such as ours one group has one
set of folkways, while another group has another set. Even the
same group may hold conflicting norms. Norms of one group may
12 See Richard C. Fuller, "Morals and the Criminal Law", Journal of Criminal Law
and Criminology, 32 (March-April, 1942), pp. 624-630.
13 William Seagle, "Primitive Law and Professor Malinowski", American Anthropo-
logist, 39 (April-June, 1937), p. 284.
STATEMENT 23
not be important to another, or a group may believe that uni-

formity in behaviour among persons is desirable while at the
same time holding that individualism, inventiveness and freedom
are also desirable. It is in this kind of society that crimes becomes
a social problem, for a considerable proportion of the population
persists in behaviour which has been officially prohibited and
threatened with punishment. And it is this kind of fact that gets
into Sutherland's theory of differential association and differential
social organization.
In some societies, informal and unofficial methods of social
control may keep nonconformity down to tolerable levels, and in
such societies it is not necessary to designate some acts as "crimes"
or "delinquencies" and to appoint officials to deal with them in an
impartial manner. If all persons behaved uniformly, so that no
one violated unofficial rules prohibiting behaviour which society
members held to be quite reprehensible, then there would be no
need for the criminal law. Thus, our complex society has criminal
laws precisely because all persons would not or could not refrain
from violating the unofficial rules, even if they agreed on an ab-
stract level that the rules are desirable. The informal methods for
inspiring conformity became inadequate, and we have attempted
to fill the breach with the more formal system of criminal law.
Yet the substitution of this formal and official system has necessa-
rily been incomplete-because of our respect for the rights of
individuals, we restrict the activities of our law-enforcement
officials and, in effect, provide numerous opportunities for per-
sons to commit crimes. Crime is a social problem because we do
not want to prevent it at all costs.
This social process must determine the point at which a theory
of delinquent and criminal behaviour will begin. Since the
process of declaring an act to be a crime is by definition social,
the act of perpetrating that crime must be by definition social.
Negatively, it is inconceivable that one could inherit a pre-
disposition to behave in a manner that is criminalistic only be-
cause it has arbitrarily been so labelled by a social group. Thus,
any general theory in criminology must recognize, as Suther-
land's theory does, that the process of declaring an act to be crime
is of a nature such that persistence in the outlawed behaviour is
highly probable. First, a value which is appreciated by a group
or a part of a group is involved. Second, a second group, or an-

other part of a single group, is isolated or in conflict with the first
unit, so that they do not appreciate the value or regard it less
highly, thus tending to endanger it. Third, the first group, or a
section of the group, resorts to coercion, applied in a disinterested
manner, in the name of the entire society, to those who disregard
the value.
Sutherland's theory has had an important effect on sociological
thought about criminality and crime, if only because it has be-
come the centre of controversy. Strangely, it seems to have
received more discussion, comment and research attention in the
last five years than in the first fifteen years of its existence. Also,
there is rapidly developing a situation in which probation, parole
and prison workers have at least heard of the theory, even if they
are barely beginning to try to use it for prevention of crime and
rehabilitation of criminals. A social worker has recently written,
"The hall mark of this new departure [in delinquency prevention]
is the recognition that delinquency is not primarily a psycho-
logical problem of neuroses but a social problem of differential
values. Essentially most delinquent behaviour arises from the
fact that core concepts of what is right and wrong, what is worth
striving for and what is attainable, are not transmitted with equal
force and clarity throughout the community."14
Some Literary Errors

The statement of the theory of differential assoCIatlOn is
neither precise nor clear. In the nine short propositions, Suther-
land presented some ideas, with little elaboration, that purport
to explain both the presence of crime and delinquency in indi-
vidual cases and the distribution of crime and delinquency rates
(epidemiology) .
It therefore is not surprising that his words do not always con-
vey the meaning he seemed to intend. Most significantly, as we
shall see later, the statement gives the impression that there is
little concern for explaining variations in crime and delinquency
rates. This is a serious error in communication on Sutherland's
part. In reference to the delinquent and criminal behaviour of
1< Bertram M. Beck, "The Young in Conflict: Blueprint for the Future", California
Touth Authoriry QuarterlY, 13:3-10, Summer, 1960.
STATEMENT 25
individuals, however, the difficulty in communication seems to

arise as much from readers' failure to study the words presented
as from the words themselves. Four principal errors, and a number
of minor ones, have arisen because readers do not always under-
stand what Sutherland seemed to be trying to say.
First, it is common to believe, or (perhaps necessarily) to as-
sume momentarily, if only for purposes of research and dis-
cussion, that the theory is concerned only with contacts or
associations with criminal and delinquent behaviour patternsY;
VoId, for example, says "One of the persistent problems that
always has bedeviled the theory of differential association is the
obvious fact that not everyone in contact with criminality adopts
or follows the criminal pattern."16 At first glance, at least, such
statements seem to overlook or ignore the words "differential"
and "excess" in Sutherland's presentation. After stating that a
person becomes delinquent because of an excess of definitions
favourable to violation of law over definitions unfavourable to
violation of law, Sutherland continues by saying "This js the
principle of differential association. It refers to both criminal and
anti-criminal associations and has to do with counter-acting
forces." Thus, he does not say that persons become criminals
because of associations with criminal behaviour patterns; he says
that they become criminals because of an overabundance of such
16 Robert G. Caldwell, Criminology, New York: Ronald Press, 1956, p. 182; Ruth
S. Cavan, Criminology, Second Edition, New York: Crowell, 1955, p. 701; Marshall
B. Clinard, The Process of Urbanization and Criminal Behaviour", American Journal
rifSociology, 48 (September, 1942), pp. 202-213; "Rural Criminal Offenders", American
Journal of Sociology, 50 (July, 1944), pp. 38--45; "Criminological Theories of Violations
of Wartime Regulations", American Sociological Review, II (June, 1946), pp. 258-270;
"The Sociology of Delinquency and Crime", in Joseph Gittler, Editor, Review of
Sociology, New York: Wiley, 1957, p. 477; and Sociology of Deviant Behaviour, New York:
Rinehart, 1957, p. 240; H. Warren Dunham and Mary Knauer, "The Juvenile Court
in Its Relationship to Adult Criminality", Social Forces, 32 (March, 1954), pp. 290-296 ;
Mabel A. Elliott, Crime in Modern Society, New York: Harper & Bros., 1952, pp. 347-
348; Sheldon Glueck, "Theory and Fact in Criminology", British Journal ofDelinquency,
7 (October, 1956), pp. 92-109; Robert E. Lane, "Why Businessmen Violate the
Law", Journal rifCriminal Law and Criminology, 44 (July-August, 1953); pp. 151-165;
Walter C. Reckless, The Etiology rif Delinquent and Criminal Behaviour, New York: Social
Science Council, 1943, p. 60; James F. Short, Jr., "Differential Association and
Delinquency", Social Problems, 4 (January, 1957), pp. 233-239; and "Differential
Association with Delinquent Friends and Delinquent Behaviour", Pacific Sociological
Review, 1 (Spring, 1958), pp. 20-25; Harrison M. Trice, "Sociological Factors in
Association with A. A.", Journal rifCriminal Law and Criminology, 48 (November-Decem-
ber, 1957), pp. 374-386; George B. VoId, Theoretical Criminology, New York: Oxford
University Press, 1958, pp. 194-195.
18 Op. cit., p. 194. But see below, p. 87, note 36.
associations, in comparison with associations with anti-criminal

behaviour patterns. Accordingly, it is erroneous to state or imply
that the theory is invalid because a category of persons-such as
policemen, prison workers or criminologists-have had extensive
association with criminal behaviour patterns but yet are not
criminals.
Second, it is commonly believed that Sutherland says persons
become criminals because of an excess of associations with
criminalsY Because of the manner in which the theory is stated,
and because of the popularity of the "bad companions" theory of
criminality in our society, this error is easy to make. Sutherland's
proposal is concerned with ratios of associations with patterns of
behaviour, no matter what the character of the person presenting
them. Throughout his formal statement, Sutherland uses terms
such as "definitions oflegal codes as favourable or unfavourable,"
"definitions favourable to violation of law over definitions un-
favourable to violation oflaw" and "association with criminal and
anti-criminal patterns." Thus, if a mother teaches her son that
"Honesty is the best policy" but also teaches him, perhaps inad-
vertently, that "It is all right to steal a loaf of bread when you are
starving," she is presenting him with an anti-criminal behaviour
pattern and a criminal behaviour pattern, even if she herself is
honest, non-criminal and even anti-criminal. One can learn
delinquency behaviour patterns from persons who are not crimi-
nals, and one can learn anti-delinquency behaviour patterns from
hoods, professional crooks, habitual offenders and gangsters.
Third, in periods of time ranging from five to twelve years after
publication of the formal statement with the word "systematic"
omitted, at least five authors have erroneously believed that the
theory pertains to "systematic" criminal behaviour only.Is Cald-
17 Harry Elmer Barnes and Negley K. Teeters, New Horizons in Criminology, Third
Edition, Englewood Cliffs, New Jersey: Prentice-Hall, 1959, p. 159; Caldwell, op. cit.,
pp. 182-183; Cavan, op. cit., p. 701; Clinard, "The Process of Urbanization and
Criminal Behaviour", op. cit.; "Rural Criminal Offenders", op. cit., and "Crimino-
logical Theories of Violations of Wartime Regulations", op. cit.; Elliott, op. cit., p. 274;
Daniel Glaser, "The Sociological Approach to Crime and Correction", Law and
Contemporary Problems, 23 (Autumn, 1958), pp. 683-702; and Differential Association
and Criminological Prediction: "Problems of Measurement", Social Problems, 8
(Summer, 1960) pp. 6-14; Glueck, op. cit.; Lane, op. cit.; Reckless, op. cit., p. 60;
Harry M. Shulman, "The Family and Juvenile Delinquency", Annals of the American
Academy rif Political and Social Science, 261 (January, 1949), pp. 21-31; Donald R. Taft,
Criminology, New York: Macmillan, 1956, p. 338.
18 Caldwell, op. cit., pp. 182-184; Cavan, op. cit., p. 701; Elliott, op. cit., p. 274;
STATEMENT 27
well has recently been as critical of the word "systematic" as was

an early article that attacked the original statement containing
the word "systematic."19
Fourth, it is commonplace to say that the theory is defective
because it does not explain why persons have the associations
they have. 20 Although such expressions are valuable statements of
what is needed in criminological research, they are erroneous
when applied to differential association. Sutherland recognized
that determining why persons have the associations they have is a
desirable problem for research, and we shall later see that when
his theory is viewed as a principle that attempts to account for
variations in crime rates it does deal in a general way with
differential opportunities for association with an excess of crimi-
nal behaviour patterns. Nevertheless, the fact that the "indivi-
dual conduct" part of the theory does not pretend to account for
a person's associations cannot be considered a defect in it. 21
Fifth, other authors have erroneously taken "theory" to be
synonymous with "bias" or "prejudice" and have condemned
Sutherland's statement on this ground. For example, in connection
with criticizing Sutherland for deleting "systematic" from the
1947 version of his theory, Caldwell has written that by 1947
"we had not acquired enough additional facts to enable [Suther-
land] to explain all criminal behaviour."22 This statement does
not clearly recognize that facts themselves do not explain any-
thing and that theory tries to account for the relationships be-
tween known facts, among other things. Confusion about the role
of theory also is apparent in Clinard's statement that Sutherland's
theory is "arbitrary," Glueck's statement that "social processes
are dogmatically shaped to fit into the prejudices of the pre-
Richard R. Korn and Lloyd W. McCorkle, Criminology and Penology, New York: Holt,
1959, pp. 297-298; VoId, op. cit., pp. 197-198.
19 Arthur L. Leader, "A Differential Theory of Criminality, Sociology and Social
Research, 26 (September, 1941), pp. 45-53.
20 Glueck, op. cit.; Clarence R. Jeffery, "An Integrated Theory of Crime and
Criminal Behaviour", Journal of Criminal Law and Criminology, 49 (March-April, 1959),
pp. 533-552; Leader, op. cit.; Martin H. Neumeyer, Juvenile Delinquency in Modern
Society, Second Edition, New York: Van Nostrand, 1955, p. 152 ; James F. Short, Jr.,
"Differential Association as a Hypothesis: Problems of Empirical Testing", Social
Problems,8 (Summer, 1960), pp. 14-25; Trice, op. cit.; S. Kirson Weinberg, "Theories
of Criminality and Problems of Prediction", Journal ofCriminal Law and Criminology, 45
(November-December, 1954), pp. 412--429.
21 See the statement on p. 19, above.
22 Op. cit., p. 182.
existing theory of 'differential association,'" and Jeffery's state-

ment, that "the theory does not differentiate between criminal
and non-criminal behaviour, since both types of behaviour can be
learned."23 Such statements are not so much errors in interpre-
tation of the differential association statement as they are errors
regarding the role of theory, hypotheses and facts in scientific
research. Later, we will show that Sutherland's whole theory does
organize and integrate known facts about crime. Here, we need
only indicate that Merton, like many others, has dispelled the
nation that sociological theory is arbitrarily imposed on the facts
it seeks to explain. 24
Additional errors stemIning from the form of Sutherland's
formal statement, from lack of careful reading of the statement,
or from assumptions necessary to conducting research, have been
made, but not with the frequency of the five listed above. Among
these are confusion of the concept "definition of the situation"
with the word "situation,"25 confusion of the notion that persons
associate with criminal and anti-criminal behaviour patterns
with the notion that it is groups that associate on a differential
basis,26 belief that the theory is concerned principally with learn-
ing the techniques for committing crimes,27 belief that the theory
refers to learning of behaviour patterns that are neither criminal
nor anti-criminal in nature,28 belief that "differential associ-
ation," when used in reference to professional thieves, means
maintaining "a certain necessary aloofness from ordinary
people," 29 failure to recognize that the shorthand phrase "differ-
23 Clinard, Sociology of Deviant Behaviour, op. cit., p. 204; Glueck, op. cit., p. 99;
Jeffery, op. cit., p. 537.
24 Robert K. Merton, Social Theory and Social Structure, Revised Edition, Glencoe:
The Free Press, 1957, pp. 85-117.
25 Milton L. Barron, The Juvenile in Delinquent Society, New York: Knopf, 1954,
p. 1OJ.
26 Elliott, op. cit., p. 274.
27 Clinard, "Criminological Theories of Violations of Wartime Regulations",
op. cit.
28 Taft writes of differential association "with others who have become relative
failures or criminals", but Sutherland's theory has nothing to say about association
with "failures", unless "failures" and "persons presenting criminal behaviour patterns"
are used synonymously. Op. cit.; p. 338.
29 Walter C. Reckless, The Crime Problem, Second Edition, New York: Appleton-
Century-Crofts, 1955, p. 169. This kind of error may stem from Sutherland himself,
for in his work on the professional thief he used the term "differential association" to
characterize the members of the behaviour system, rather than to describe the process
presented in the first statement of his theory, two years later. See Edwin H. Sutherland,
The Professional Thief, Chicago: University of Chicago Press, 1937, pp. 206-207.
STATEMENT 29
ential association" is equivalent to "differential association with

criminal and anti-criminal behaviour patterns," with the conse-
quent assumption that the theory attempts to explain all be-
haviour, not just criminal behaviour,30 and belief that the theory
is concerned only with a raw ratio of associations between the
two kinds of behaviour patterns and does not contain the state-
ment, explicitly made, that "differential association may vary in
frequency, duration, priority and intensity."31
80 Howard B. Gill, "An Operational View of Criminology", Archives of Criminal

Psychodynamics, October, 1957, p. 284; Jeffery, op. cit.
81 Clinard, "Criminological Theories of Violations of Wartime Regulations",
op. cit. If these "modalities", as Sutherland called them, are ignored, then the theory
would equate the impact of a behaviour pattern presented once in a radio show with
the impact of a pattern presented numerous times to a child who deeply loved and
resprected the donor. It does not so equate the patterns.
CHAPTER II
SOME OBSTACLES TO GENERALIZING IN

CRIMINOLOGY
In the years since criminology began, and especially since the

decline of the "Biological School" near the beginning of the
present century, the quantity of criminological research con-
ducted in the United States has increased enormously, but
adequate knowledge about crime causation has not increased
proportionately. Particularly prior to the last quarter of the nine-
teenth century, most men who advanced theories about crime
causation did so in an attempt to find a panacea for criminality.
Few efforts were made to verify the many theological and
moralistic assertions by actual investigations offactual situations;
writers usually selected a general "cause" of all criminality and
then sought to convince their readers that elimination of that
cause would eradicate crime both by reforming criminals and by
preventing future criminality. There was little or no attempt to
"make sense," in a logical generalization, of the known facts about
variations in crime rates, despite the fact that Europeans such as
Quetelet, Guerry, Joly, Liszt, Prins, Hamel, Fointsky, Tarde and
others had compiled such facts and had attempted to generalize,
in a preliminary way, about them.
Near the end of the nineteenth century, sociology began
making its way into the curricula of American universities and
colleges, and a survey conducted in 1901 indicated that crimi-
nology and penology were among the first courses offered under
the general title of "Sociology." 1 From that time to this the main
academic stronghold of criminology has been in departments of
sociology. As early as 1914 a sociologist wrote, "The longer the
study of crime has continued in this country, the greater has
1 Frank L. Tolman, "Study of Sociology in Institutions of Learning in the United

States", American Journal of Sociology, 7 (May, 1902), pp. 797-838; 8 (July, 1901),
pp. 85-121; 8 (September, 1902), pp. 251-272; 8 (January, 1903), pp. 531-558.
OBSTACLES TO GENERALIZING 31
grown the number of causes which can be described as social.

This is the aspect in the development of American criminology
which has given to that study in this country the title of 'The
American School."'2 This trend has continued down to the
present. With the passage of time, more and more concentrated
efforts have been made to integrate the findings of investigators
in the general fields of sociology and social psychology and to
develop a unified body of knowledge which would give an
adequate scientific explanation of crime and criminality. The
most noteworthy attempts to generalize about the data, or at
least about many of the data, on crime and criminality are those by
Sutherland, Sellin, Taft and Merton. 3
There have been two distinct handicaps under which crimi-
nological theoreticians like Sutherland have worked, and it is
these two handicaps which are the subject-matter of this chapter.
First, attempts to show logical and theoretical relationships be-
tween variations in crime rates and variations in social organi-
zation have suffered from the methodological difficulties stemIning
from the manner in which statistics on crime rates, one kind of
basic data, are compiled. There is considerable hazard involved
in making generalizations about crime as measured by con-
ventional criIninal statistics. Second, attempts by sociologists and
social psychologists to define and identifY processes-like differ-
ential association-by which persons become delinquents and
criIninals have been blocked by widespread subscription to a
philosophy of scientific despair, the so-called "multiple factor
theory." Scholars who are convinced that criminality and de-
linquency can be carefully studied but never explained cannot be
expected to follow up and, possibly, modifY and improve any
given theoretical generalization.
• John L. Gillin, "Social Factors Affecting the Volume of Crime", in Physical Basis
qfCrime: A Symposium, Easton, Pennsylvania: American Academy of Medicine, 1914,
pp.53-67.
8 Thorsten Sellin, Culture Conflict and Crime, New York: Social Science Research
Council, 1938, pp. 20-40; Robert K. Merton, "Social Structure and Anomie",
American Sociological Review, 3 (October, 1938), pp. 677-682 (This article has been
revised and enlarged and it now appears in Robert K. Merton, Social Theory and
Social Structure, Revised Edition, Glencoe: The Free Press, 1957, pp. 161-194); Donald
R. Taft, Crimonology, Third Edition, New York: Macmillan, 1956, pp. 336-349,
754-755.
Measuring Crime Rates

The principal point made in most reports on the evaluation of
the criminal statistics published by either local or central agencies
is a negative one-the statistics are not reliable indexes of crime
rates. After at least a quarter-century of articles of this sort, we
are aware that the general statistics on crime seem to be among
the most unsatisfactory of all social statistics. It is important to
note, however, that we do not know that the statistics are unreli-
able; problems ofinterpretation arise precisely because we do not
in fact know whether the statistics are reliable or not. It is possible
that any given set of general statistics on crime is a very accurate
indicator of the state of crime generally; problems arise because
we do not have any means for determining this with a measurable
degree of certainty. Our awareness of the difficulties in using
criminal statistics has not served to correct the presumed deficien-
cies, but it has certainly introduced an extreme note of caution in
the interpretation of any set of statistics dealing with crime. We
shall review some of the reasons why crime statistics must be
interpreted with caution and in the next chapter we shall ask
what facts of significance to criminological theory can be learned
from available statistics. Six principal criticisms of the general
statistics on crime have been made.
First, statistical data on the true crime rate cannot be compiled,
for the simple reason that it is impossible to determine the amount
of crime in any given locality at any particular time. Many
crimes are not discovered. Others are discovered but not re-
ported. Still others are reported but not recorded in any official
way, so they do not appear in sets of crime statistics. Consequent-
ly, any record of crimes is at best an "index" of the total number
of crimes committed.
Second, such "indexes" do not maintain a constant ratio with
the true crime rate, whatever it may be. Any statistical index,
such as the "cost of living index" compiled to show fluctuations
in the total cost ofliving, is based on a sample of items taken from
the whole. The relationship to the whole is known and a report
on the variations in the index items is thus a report on approxi-
mate variations in the whole. But in crime statistics the variations
in any set of figures cannot be considered a real index; the "index"
items cannot be a sample of the whole because the whole cannot
be specified. Both the true rate and the relationship between that
rate and any "index" vary with changes in public opinion, police
policies and other conditions.
Third, variation in the conditions affecting published records
of crime makes it foolhardy to compare crime rates in different
jurisdictions and it is hazardous even to compare the rates of the
same jurisdiction (such as a city, county, state or nation) in two
different years. Variations in definitions of crimes make inter-
national comparisons even more difficult.
Fourth, crime statistics are compiled primarily for administra-
tive purposes and for that reason they cannot be relied upon in
scientific research. A scholar who develops a theory that accounts
for variations in crime rates, whatever they may be, does so at
great risk, for the extent of statistical error in any observed
variation is unknown. A theory which accounts, in sociological
terms, for the fact that city A has much higher crime rate than
city B can become embarrassing if it is discovered that the fact of
variation is a consequence of different practices for reporting and
recording crimes, rather than of real differences in criminality.
Fifth, statistics on the variations in the recorded rates for some
crimes, such as white-collar crimes, are not routinely compiled.4
Many crimes committed by persons of upper socio-economic
status in the course of their business are handled by quasi-judicial
bodies, such as the Federal Trade Commission, in much the way
that children's cases are heard in juvenile courts rather than the
criminal courts used for adults.
Sixth, the statistics onjuvenile delinquency are subject to all of
above criticisms and, in addition, are inadequate because de-
linquency is not precisely defined in some jurisdictions. Even
when a child is committed to an institution for delinquents there
is no positive assurance that he has committed an act which,
except for his age, would be crime.
"Crimes known to the police" is the set of statistics most
generally accepted as the most adequate set of crime figures
available in the United States. Each year the Federal Bureau of
Investigation publishes a summary statement of the serious crimes
known to the police departments who report to it. However, as
indicated in the first three points above, this set of statistics is
• See Edwin H. Sutherland, White Collar Crime, New York: Dryden, 1949.
affected by conditions which cannot readily be taken into account

in statistical reporting, conditions such as police practices, politics,
laws and public opinion. Consequently, even "crimes known to
the police" may be an inadequate index of true rates. Yet the
decision to use this rate is probably the best way out of a bad
situation for, as Professor Sellin, the nation's foremost expert on
crime statistics, has repeatedly pointed out, "the value of criminal
statistics as a basis for measurement in geographic areas decreases
as the procedure takes us farther away from the offence itself."5
Despite the assumed efficiency of "crimes known to the police,"
however, there are five principal reasons for doubting that this set
of figures constitutes an adequate index of the true crime rate.
The kind of inadequacy of all crime statistics can be illustrated by
brief discussion of these reasons. 6
First, the number of crimes known to the police certainly is
much smaller than the number actually committed. Just how much
smaller is, of course, open to question. But in one period the
number of cases of shop-lifting discovered in three Philadelphia
department stores was greater than the total number of thefts of
all kinds in the entire city which were known to the police. 7
Second, the number of crimes recorded as "known" to any
police department may be only a fraction of the crimes actually
known to them. This is illustrated by a study made some years ago
in Chicago. The number of robberies "known to the police" in
Chicago increased from 1,263 to 14,544 between 1928 and 1931
and burglaries increased from 897 to 18,689 in the same period.
This tremendous increase was not due to a "crime wave," to an
influx of gangsters or to increased activity on the part of Chicago
hoodlums. Instead, the change was due almost completely to
changes in recording practices following an investigation of the
police department. 8
Third, it is highly probable that the ratio of crimes committed
to crimes known to the police varies with the nature of the offence.
• Thorsten Sellin, "The Significance of Records of Crime", The Law Quarterly
Review, 67 (October, 1951), pp. 496-504.
6 For more detailed discussion, see Edwin H. Sutherland and Donald R. Cressey,
Principles of Criminology, Sixth Edition, Philadelphia: Lippincott, 1960, pp. 27-30.
7 Thorsten Sellin, Research Memorandum on Crime in the Depression, New York: Social
Science Research Council, 1937, p. 69.
B Virgil W. Peterson, "An Examination of Chicago's Law Enforcement Agencies",
Criminal Justice (january, 1950), pp. 3-6.

It has been pointed out that only an insignificant proportion of

sex offences are reported to the police. 9 But most murders come
to the attention of the police and are recorded as known. As Sellin
has said:
We cannot use the total recorded criminality. We must extract from the
total the data for only those offences in which the recorded sample is large
enough to permit the assumption that a reasonably constant relationship
exists between the recorded and the total criminality of these types. We may
make that assumption when the offence seriously injures a strongly embraced
social value, is of a public nature in the sense that it is likely to come to the
attention of someone besides the victim and induces the victim or those who
are close to him to cooperate with the authorities in bringing the offender
to justice. 10
A further complication in this regard arises because some crimes

become known to the police only if a victim complains, while
other offences become known by direct observation on the part of
the police. Cases of drunken driving, for example, usually get into
police records only if observed by a policeman, while cases of
burglary usually become known to the police as a result of a
report by the victim. The ratio of offences committed to offences
known probably is greater for those offences that get into police
records only when observed by the police than for those which
get into the reports as a result of complaints by victims. For
example, the amount of drunken driving not appearing in a set
of crime statistics probably is greater than the amount of burglary
not appearing in similar sets of statistics.
Fourth, the definition of crime varies from time to time and
from place to place. Reports on general crime rates, rather than
on rates among specific offences, are then inadequate for long-
range comparative purposes. Further, agencies doing the re-
cording may classify offences in an unsystematic and irregular
way, so that variation in the rate of a particular offence is created
when none in fact exists.u For example, in some jurisdictions
the same act is sometimes recorded as "embezzlement," some-
times as "larceny by bailee," sometimes as "confidence game"
9 A. C. Kinsey, W. B. Pomeroy and C. E. Martin, Sexual Behaviour in the Human Male,

Philadelphia: Saunders, 1948, p. 392.
10 "The Significance of Records of Crime", op. cit.
11 For more detailed discussion of this "categorization problem", see Samuel A.
Stouffer, "Indices of Psychological Illness", in Paul F. Lazarsfeld and Morris Rosen-
berg, Editors, The Language ofSocial Research, Glencot" : The Free Press, 1955, pp. 63--65.
and sometimes as "forgery," making it impossible to get a close

estimate of the amount of variation in time of anyone of these
offences. 12
Fifth, when comparisons are made, the mere number of crimes
known to the police is not sufficient. What is needed is statements
of rates-the number of crimes in proportion to the number of
population or in proportion to some other base. But determi-
nation of this base is sometimes almost as difficult as determination
of the crime ratio itself. If a large proportion of the population of
one city is young males, then it is incorrect to compare the crime
rate of that city with the rate of a city in which a large proportion
of the population is old women. Moreover, sometimes even
accurate data on the population base is not sufficient. For
example, automobile theft might reasonably be calculated in
proportion to the number of automobiles, as well as in proportion
to the number of males and females of various ages. While the
rate of crimes against persons (such as assault and murder) might
reasonably be calculated in proportion to the number of persons
in an area, the rate of crimes against property (such as larceny
and burglary) probably should be calculated against the amount
of property of various kinds in the area. IS
Two important studies have shown that a large number of
crimes are "hidden," in the sense that they do not appear in any
set of crime statistics. In the first study, 1,020 men and 678 women
were asked to check which of forty-nine listed offences they had
ever committed.14 An effort was made to distribute the question-
naire to a balanced religious and racial cross section of the
population, but systematic sampling procedures were not used.
Most of the respondents were New Yorkers and. the group of
subjects contained an excess of persons from upper socio-economic
classes. Ninety-one per cent of the respondents admitted that
they had committed at least one offence, excluding juvenile de-
linquencies, for which they could have beenjailed or imprisoned.
Sixty-four per cent of the men and twenty-seven per cent of the
women said they had committed at least one felony, a class of
12 See below, pp. 109.
13 See Stuart Lottier, "Distribution of Criminal Offences in Metropolitan Regions",
Journal ofCriminal Law and Criminology, 29 (May-June, 1938), pp. 37-50.
14 J. S. vVallerstein and C.J. Wyle, "Our Law-Abiding Law-Breakers", Probation,
25 (March-April, 1947), pp. 107-112.
serious offences usually punishable by at least one year III the

state prison.
In the second study, "white-collar crimes" were investigated. 15
These are crimes committed by persons of respectability and high
social status in the course of their occupations. It was found that
prosecution for this kind of crime often can be avoided because of
the political or financial importance of the individuals involved,
because of the difficulty of securing evidence or because of the
apparent triviality of the crimes. Also, the actions are punishable
under criminal laws, but their perpetrators are often tried in the
hearings of administrative commissions and in the civil courts,
rather than being subjected to the regular criminal court pro-
cedures. In a comment on the frequency of such crimes-in the
form of misrepresentation of financial statements, bribery,
embezzlement, tax fraud, misrepresentation in advertising, etc.-
Sutherland said:
The manufacturers of practically every class of articles used by human
beings have been involved in legal difficulties ... with more or less frequency
during the last thirty years, including the manufacturers of the surgical
instruments with which an infant may be assisted into the world, the bottle
and nipple from which he may secure his food, the milk in his bottle, the blanket
in which he is wrapped, the flag which his father displays in celebration of the
event, and so on throughout life until he is finally laid away in a casket which
was manufactured and sold under conditions which violated the law. 16
More specifically, Professor Sutherland's study of the white-

collar crimes of the 70 largest mining, manufacturing and mercan-
tile corporations in the United States indicated that in a period
of forty years every corporation had violated at least one of the
laws outlawing restraint of trade, misrepresentation in adver-
tising, infringement of patents, trade-marks and copyrights,
violations of war-time regulations and some miscellaneous activi-
ties. The corporations had a total of 307 adverse decisions on
charges of restraint of trade, 222 adverse decisions on charges of
infringements, 158 adverse decisions under the National Labor
Relations Law, 97 adverse decisions under the laws regulating
advertising and 196 adverse decisions on charges of violating
other laws. Generally, the official records iIidicated that the
16 Sutherland, op. cit.

11 Edwin H. Sutherland, "Crime and Business", Annals qf the American Academy of
Political and Social Science, 217 (September, 1941), pp. 111-118.
corporations violated the laws regulating trade with great fre-

quency, but that such crimes do not appear in the official sets of
statistics used as "indexes" of the true crime rate in the United
States.
The Multiple Factor Approach

Later, we shall see that while Sutherland's principle of differ-
ential association and differential social organization does not
make good sense out of all criminological data, it seems to make
better sense out of more criminological facts than do some of the
other theories. One of the serious difficulties in current crimi-
nology arises from the failure of all but a handful of research
workers and theoreticians to determine whether Sutherland's
theory, or any other theory, does or does not make sense of the
available data. This negligence, we believe, stems in part from
the great emphasis placed on the multiple factor approach to the
study of crime and criminality.
Healy's emphasis on multiple causation in the cases of indivi-
dual delinquents,!' combined with the then-prevalent en-
deavours to discount the biological and physical conditions which
were said to produce crime and criminals, played an important
part in the development of this approach to the study of crime.
The revolt against the practice of stressing one cause for all crime
led to the production of extensive lists of "causes" of crime, each
one of which was at first said to produce a portion of all criminals.
This idea that crime is a result of multifarious influences has
persisted to the present, although the specific factors said to be
important in crime have shifted from time to time, progressively
becoming more social in nature. Also, there has been a shift to-
wards concentration on multifarious factors in the crimes of indivi-
dual offenders, rather than on multifarious factors producing high
crime rates.
When the multiple factor approach is used in the study of
individual cases it is maintained that one crime is caused by one
set of "factors" and that another crime is caused by a different set
of factors. Advocates of the approach pay little or no attention to
developing a system of theory which would try to make sense of
17 William Healy, The Individual Delinquent, Boston: Little, Brown, 1915, pp, 16,
33-125, 130-138.
the relationships that might be found. Rather, one delinquent is

said to have had to have become a delinquent because of the
influence of three or four factors, another delinquent is said to be
delinquent because of other factors, and no elaboration is viewed
as necessary. Some advocates of the approach hold that each
factor is of equal importance, but the majority argue that the
presence of one or two "important" factors or seven or eight
"minor" factors will produce criminality.
When the multiple factor approach is used in the study of
variations in crime and delinquency rates, the investigator
simply lists conditions that are associated with crime to a high or
low degree. The higher the degree of association, the more
importance ascribed to the particular condition. There is little or
no attempt to make sense out of the statistical findings by showing
that they are consistent or inconsistent with a theoretical scheme.
Professor Reckless has advocated the use of this method under the
name "actuarial approach":
The actuarial approach assumes that individuals have a greater or lesser
liability to be caught and reported as violators [of the criminal law] by virtue
of the position they occupy in society as determined by their age, sex, race,
nativity, occupational level and type of residence. The behaviour which is
studied is only that which is reported in contrast to that which is not recorded.
The liability is strictly that of becoming the sort of violator who is reported. 1S
Thus, the concern is with statistical association; there is no

imputation of "causal power" to any factor or set of factors. At
least as early as 1895, criminologists were warned not to impute
causal power to statistical correlations, on the dual grounds that
the connection between the alleged cause and the alleged effect
might merely be an accidental one and that "the coincidence of
two statistical facts does not in itself determine which of the two
is primary and which is secondary, nor whether both may not
have in common origin a third which, perhaps, is not included in
the scope of the investigation."19
Advocates of the multiple-factor approach often take pride in
their broadmindedness, arguing that explanations by psychiatrists,
sociologists, psychologists or members of any other academic
18 Walter C. Reckless, The Etiology qf Delinquent and Criminal Behaviour, New York:
Social Science Research Council, 1943, p. 74.
19 Frederick H. Wines, Punishment and Reformation, New York: Crowell, 1895,
p.267.
discipline are narrow and particularistic. 20 However, in the best

critique of the multiple factor approach to the study of individual
cases, Albert Cohen has made three principal points. 21
First, persons using the multiple factor approach confuse
explanation by means of a single theory or set of theory with
explanation by means of a single Jactor. No one now attempts to
explain crime by means of a single factor, but this does not mean
that crime cannot be explained by means of a single theory. A
single theory is a logical statement showing how variations in one
thing are linked up with variations in other things. Thus, theory
is concerned with many variables, but the variables are not
considered as causal agents. We make statements ofJact in terms
of values (high, low etc.) of variables: "The crime rate is high
among persons with low incomes, among young adults and among
males." But such statements of fact, whether concerned with the
relationship of the values of one variable to crime (income) or
with the relationship of the values of a number of variables
(income, age, sex) are not explanations of crime. A criminological
theory makes sense of the facts about crime and the best theory is
the one that makes the best sense out of the most facts.
Second, the "evil-causes-evil fallacy" usually characterizes the
multiple factor approach although it is not peculiar to it. The
fallacy is that results which we do not like (crime) must have
a.ntecedents which we do not like (alcoholism, psychopathic
personality, biological inferiority, lack of love by parents, over-
protection by parents, etc.). When this fallacy is present, "expla-
nations" of crime or almost any other social problem are likely to
be statements attributing causal power to a list of ugly and sordid
conditions which any "decent citizen" deplores. It is very difficult
to reason that crime (an evil) does not necessarily result from
something also considered evil, but instead might result from
something which most persons hold to be good. A "good thing"
like democracy, freedom or individualism can have unantici-
pated "evil" consequences like crime. Perhaps in criminology
there is a tendency to seek the "causes" of crime in evil ante-
20 Sheldon Glueck, "Theory and Fact in Criminology", British Journal of Delin-
quenz;y, 7 (October, 1956), pp. 92-109.
21 Albert K. Cohen, Juvenile Delinquency and the Social Structure, Unpublished
Ph. D. Dissertation, Harvard University, 1951, pp. 5-13; See also his Delinquent Boys:
The Culture of the Gang, Glencoe: The Free Press, 1955.
cedents because this practice will enable us to denounce crime

without hurting anyone's feelings and because it will enable us to
attempt to eliminate crime without changing any conditions
which we hold dear.
Third, "factors" are confused with "causes" and each factor is
assumed to contain within itself a fixed amount of crime-pro-
ducing power. The fact that a person has a low income is said to
have some crime-producing power; the fact that he is a young
adult pushes him further in the direction of crime; and the fact
that he is a male is the last straw. Sometimes the basis for im-
puting causal power to a factor in an individual case is high
statistical association between the factor and crime rates-if in
a city the areas of poor housing are also the areas of high crime
rates, a person who lives in a poor house is considered on the road
to crime. Statisticians have pointed out the fallacy of such reason-
ing and it is not part of the "actuarial approach." 22 But sometimes
the basis for imputing causal power to a factor is based on only
subjective judgments of the person doing the imputing and it
cannot be determined at all by outsiders. Often, for instance,
persons may say that poor eyesight, or red hair, or inability to
playa musical instrument is a "factor in" delinquency or crime
but no one can determine why such conditions are viewed as
"pushers." Nothing in the condition of poor eyesight jumps up
and pushes a person in one direction or another; if the condition
is important at all it must be related to some meaning which poor
eyesight has to the person and to the effects of this meaning on
some process which leads to crime. Specification of this process is
a theory of criminality.
The "Inadequacy" of Statistics

Complaints about the inadequacy of crime statistics seem to be
closely related to the great popularity of the multiple factor ap-
proach to the study of crime and criminality. Because the data are
commonly said to be inadequate, it is easy to dismiss attempts at
generalization and to subscribe to an approach which cannot be
subjected to an attack on the ground that the statistics do not
show what they seem to show. Moreover the practice of com-
22 William S. Robinson, "Ecological Correlations and the Behaviour of Indi-
viduals", American Sociological Review, 15 (June, 1950), pp. 351-357.
piling vague statistics dealing with only a few of the characteristics

of crime and criminals, and the practice of maintaining that
crime is caused by vaguely-defined multifarious influences, ap-
pear to have at least one important characteristic in common-
both serve a useful purpose by reconciling the widely divergent
ideological and theoretical commitments held by the many
persons who must deal with criminals. Perhaps it is the "inade-
quacy" of the statistics and the multiple factor approach which
enable men to work together in a fair degree of harmony even if
they have entirely contradictory commitments about what causes
crime and what should be done about it. This hypothesis about
the "usefulness and vagueness" in criminology can be made
clearer by examining the principal purposes for which criminal
statistics are compiled.
First, perhaps the most persistent demand for better statistics
comes from persons having short-range interests in evaluating
programmes which have been put into effect. A large proportion
of the research requests received by the United States Children's
Bureau, for example, are requests for help in evaluating some
programme or service-that is, for assistance in stating how
successful, effective or desirable the service is, as compared with
some ideal, some norm or some alternative. 23
Second, a persistent demand for better statistics is made by
persons having longer-range perspectives in prevention, law-
enforcement or corrections programme but who, nevertheless,
also are interested primarily in evaluation. As Herzog has said,
"Their emphasis would be less on a final rating of success or
failure and more on a set of correlations between elements of
treatment and elements of outcome; or between attributes or
problems of clients and outcome. 24
Finally, requests for better statistics are regularly requested by
persons with even broader theoretical interests in the etiology of
delinquent and criminal behaviour. The problem here usually is
conceived as one of obtaining accurate statistics on variation in
crime rates in time and by groups and geographic areas, so that
true differences can be discerned. If we do not know how much
23 See Elizabeth B. Herzog, "Research Planning. 1. One Type of Evaluative

Research", The Social Service Review (September, 1956), pp. 322-357.
24 Ibid., p. 325.
crime or delinquency an area has and can therefore only suspect

that its rates are higher or lower than the rates for some other
area, theoretical generalizations such as Sutherland's must con-
tinue to be quite hazardous.
Significantly, the kind and amount of statistics gathered for
one of these three purposes might be quite useless for the other
purposes. Further, the persons writing on the inadequacies of
statistics are only infrequently the persons who are gathering
them. Those who gather statistics may very well have purposes
quit different from those who examine the data, and for that
reason might consider the statistics to be inadequate.
Perhaps more detailed consideration of the major differences
among the three principal kinds of research which utilize crime
statistics, together with speculation as to reasons why inaccurate
statistics might be valuable for each kind of research, will indicate
that our "inadequate" statistics are not so inadequate after all.
(I) In administrative-evaluative research an immediate answer
to policy questions is sought. Statistics are compiled and inter-
preted in order to give an accounting to the administration of an
agency, to the public or to some special group. As Beattie has
pointed out:
In this country there is spent annually about two billion dollars in the direct
enforcement of the criminal law. It is highly essential that some rather exact
knowledge and information be available to indicate to the taxpayer, who has
to contribute this amount, whether or not the processes oflaw enforcement are
efficient and accomplish their purpose with fairness and justice to those who
are apprehended and prosecuted. 26
But "knowledge" and "information" of this kind do not stand

alone; they must justify a course of action. Thus, if an agency
handles a certain number of delinquents or provides certain
services to the community each year, it must specifY these
statistically, and the statistics must justify spending money on
these activities rather than on something else. If this is not done
the agency will go out of business. Such data as are collected by
the agency must, then, show efficiency or, at least, must not show
inefficiency in accomplishment of assigned goals.
Similarly, trends must be tabulated in order to anticipate
25 Ronald H. Beattie, Manual of Criminal Statistics, New York, American Prison
Association, 1950, p. 7.
future budgetary needs. Since budgets are almost never decreased

by the requesting organizations, statistics must be available for
backing up arguments for maintaining or increasing the current
appropriations. Consistently, most organizations concerned with
crime and corrections are subdivisions of state political adminis-
trations and, since politicians must be opposed to crime as well as
to sin and man-eating sharks, it is expedient to show that the
political administration in power and the sub-division administer-
ing a programme both are doing the best possible job.
It is important to note that statistical tabulations furnish
grounds for public esteem and professional reputation, as well as
information about programmes or about the condition of crime
in an area. These two are not the same. Because personal and
organizational needs supplement the societal needs which are
involved in programmes, "reporting" for the second purpose must
not threaten public or professional reputations. Further, if the
results of "evaluative research" do threaten the agency's or a
professional's reputation, those results are interpreted as incon-
clusive. The leaders of one evaluative research project in an
agency with which I was once associated analyzed statistics on
recidivism and concluded that one rather expensive phase of the
programme apparently was no more successful than a less ex-
pensive phase aimed at the same thing. The agency personnel
considering these data and conclusions had two principal re-
sponses: (a) The statistics and conclusions "must not get out of
the room," or the budgetary appropriation for the programme
might be cut off. (b) The statistics do not necessarily lead to the
conclusion that the expensive phase of the programme is not
justified; they are subject to all the general qualifications that
must be put on crime statistics, plus additional ones specific to the
agency.
The latter response is important. It indicates that the poor
quality of the statistics is useful to the agency. More generally, if
statistics were precise, complete and valid, agency programmes in
which many of us have vested interests in terms of promotions,
esteem and job security might not be continued. Agencies have
many purposes which are not spelled out in their official charters
or "statements of goals" -among other things they give profession-
al satisfaction and meet the personal needs of many employees.
We cannot, therefore, always afford to compile statistics which

indicate trends that cannot readily be interpreted as due to
deficiencies in the statistics, rather than to deficiencies in pro-
gramme, even if the evaluative research is to be conducted by an
outside agency. On the contrary, vague kinds of statistics, subject
to interpretation as "poor" if they do not support our programme
and to interpretation as "good" if they do support it, are desirable.
(2) In pre-evaluative research, as Herzog calls the second type,
the purposes of gathering and interpreting statistics are neither so
immediate nor so compelling as in administrative-evaluative
studies. The long-range objective is to test whether a programme
is doing what it is said to be doing, but the research is "purer" or
"more basic." The primary goal is to increase knowledge about
the effectiveness of a programme, ordinarily by means of testing a
theory with which the programme is thought to be consistent.
But validation is quite difficult if the programme involved is of a
"multiple factor" type, and studies therefore ordinarily must be
confined to only a small segment of the total programme. Rather
than evaluating a prison's programme, for example, the study
might be of the use of group therapy in the prison; the emphasis,
moreover, might not be on a final judgment of success or failure
of group therapy in the institution, but on certain techniques
which do or do not accomplish certain desired changes in
immates, on the types of personnel administering the therapy, on
the type of theory used, or on a host of other variables. The best
prison group therapy programme in the United States might,
thus, be shown to be afailure. "Best" is used here to indicate that
the programme is consistent with some theory or professional
standard, and "failure" is used to indicate that the conditions
under which it is administered make it impossible for it to be
effective. I t is in this sense that research of this kind is pre-
evaluative.
Again, since a research finding that, perhaps, all is not going
according to expectation might be interpreted as meaning that a
programme should not be given financial support, it is convenient
to see that the research cannot be definitive. Ultimately, this kind
of research might furnish grounds for public opinion about the
agency and, like adminstrative-evaluative research, it might be
politically dangerous. However, if it is restricted to a pre-evalu-
ative phase, any adverse professional, administrative or budgetary

decision based on it can be countered by an expose of the poor or
incomplete research design. Thus, both the multiple factor
idea and statistics which make it all but impossible to do even
definitive "pure" research on effectiveness of programmes are va-
luable.
(3) Etiological research is quite different in objectives from either
of the above types. Here the purpose is to make valid generali-
zations about differences and variations in crime rates; the
important statistical task is to keep law-enforcement procedures
and other variables constant. If this is done it can be determined
whether or not the crime rate in an area or group is increasing or
decreasing and whether an area or group has a disproportionate
share of crime. If procedures for compiling statistics do not
account for the variation, then the path is open for generalizations
like Sutherland's and ultimately, perhaps, modification of the
conditions which create the high crime rate.
But etiological research is dependent upon statistical compi-
lations which are rarely made with this kind of research in mind.
Since social science research has passed the phase in which data
gathered for non-scientific purposes can be used scientifically,26
compilers of statistics cannot be helpful to researchers in etiology
unless they are told what to compile. But a vicious circle is
created here, for researchers cannot tell them what to compile
unless they can be reasonably certain of the validity of the theory
which would be tested. This is particularly true among adherents
of the multiple factor approach. Theory, thus, has "to wait upon
the incidental availability of relevant data," 27 a condition existing
in part because theoreticians do not know what data are "rele-
vant." And, of course, even if they are asked, agencies cannot
secure the kind of statistics theoreticians might want. The organi-
26 In one of the earliest parole prediction studies, Burgess used the data that
happened to be recorded on envelopes containing case records of Illinois prisoners
to establish his experience tables. This was considered sufficient at the time, but at
present one of the principal criticisms of parole predication studies is that they try to
predict on the basis of factors which do not make theoretical sense. See the discussion
by Burgess in A. A. Bruce, E. W. Burgess and A.]. Ramo, The Workings of the Inde-
terminate-Sentence Law and Parole System in Illinois, Springfield, Ill., 1928; and Daniel
Glaser "A Reconsideration of Some Parole Prediction Factors", American Sociological
Review, 19 (june, 1954), pp. 335-341.
27 Merton, Social Theory and Social Structure, op. cit., p. 107.
zations collecting statistics are often officially committed to

changing the law-enforcement and corrections practices which
etiological researchers want held constant. If an experimental
change in police practices alters a city's crime rate, the police
cannot be expected to compile data indicating what the rate
might be if the practices had not been changed.
Moreover, theoreticians of various schools often have vested
interests in what is being compiled and etiological implications
are contained in routine tabulations. Organizations devoted to
preventing or reducing crime and delinquency are likely to com-
pile statistics indicating that deviation of this kind is caused by
whatever it is the agency is trying to correct with its programme.
Perhaps the great popularity of the multiple factor approach in
criminology can be attributed to the fact that an enormous
number of agencies with many different purposes are trying to
combat crime in many different ways. The tendency, of course,
is to say the opposite-that diverse agencies have been created to
handle multiple factors. For example, in a clinic whose official
purpose is prevention of delinquency, an impressive number of
children might in the course of a year be fitted with glasses or
have their teeth repaired. It is quite probable that the annual
report of the agency will include statistics on these activities and
will state or imply that delinquency is caused by conditions
stemming at least in part from "factors" such as defective eyesight
and poor teeth. On the other hand, if the agency is officially
established to administer recreational programmes in a slum
area, it is quite probable that "poor recreation" will be listed as a
"factor" in delinquency and that the number of bats and balls
issued in a year will be tabulated as evidence of delinquency
prevention.
Thus each school of theoreticians might want its statistics
compiled but does not necessarily enthusiastically support compi-
lations of materials which would support a contradictory theory.
So far as the kind of statistics collected is concerned, this leads to a
championing of the status quo, for theoretic interest tends to
concentrate in those areas where there is an abundance of a
pertinent kind of statistical data, and compilation of an entirely
different set of facts would be a threat to continuation of a
particular theory. For example, the differential association
theory, which I find useful for explaining differences of the kind

enumerated in the next chapter, would be seriously challenged
if we stopped collecting statistics on age, sex, etc. Also, signifi-
cantly, it would be seriously challenged by improved statistics
which indicated that there really is no difference between the
crime rates of the various categories.
Even more significantly, improved reliability of the statistics
collected can seriously threaten vested interests ifthey provide an
opportunity for conclusive tests of criminological ideas that we
champion for ideological reasons. As indicated earlier, when we
use the multiple factor approach we tend to select as causes of
crime and delinquency such "factors" as poverty, poor education,
lack of parental supervision, rejection and repressed hostility.
Such conditions can, presumably, be eliminated without changing
social conditions which all of us hold dear, and they can be safely
denounced. A crucial test by means of reliable statistics might
seriously challenge the notion that certain of the ideologically
important "factors" are also theoretically important, and this
means that vague statistics take on a certain value. Moreover, by
using a "multiple factor theory" a new ideologically-important
"factor" can readily be substituted for any similar ideologically-
important "factor" that were shown by reliable statistics to be un-
related to the etiology of crime and delinquency. Vague theory of
a "multiple factor" sort thus is useful to maintenance ofideological
commitments which might conceivably be threatened by reliable
statistics.
It may be concluded that in etiological research as well as in
evaluative and pre-evaluative research, the valuable statistics are
those which will not favour the ideas of anyone of the many
groups conducting research or administering programmes. It is
convenient to espouse theories of crime causation which can be
safely pronounced without hurting anyone's feelings and which
are stated in terms of conditions which "decent" citizens deplore,
and it is convenient to compile statistics which will neither slight
nor seriously challenge any group of theoreticians. Ultimately,
this means that the statistics compiled will please no-one. Just
as vague, common-sense and "umbrella" terms are useful to
interdisciplinary "crime commissions" and research teams be-
cause they reduce the area about which disagreement can be
expressed (thus indicating high degrees of consensus when in fact

no-one knows what his colleagues are taling about), so in compi-
lations of statistics vagueness is useful because it decrease the
range of points on which disagreements can occur.
CHAPTER III
SOME FACTS A THEORY MUST FIT
Despite all their limitations, the criminal statistics give infor-

mation which is important to our understanding of crime and
delinquency and to hypotheses and theories about them. Similari-
ties and differences in crime rates for certain categories of persons
are so consistent that it can reasonably be concluded that a gross
relationship between the category and crime exists in fact. In
these cases, it is practical to make the assumption that if that part
of an observed relationship which is due merely to the methods of
collecting statistics were eliminated, a real relationship would
still remain. After specifying this assumption, we can go ahead
and use the statistics. Even if they are gross, relationships which
consistently appear and which cannot be readily "explained
away" by citing the inadequacy of crime statistics, must be taken
into account in any theory of crime and criminality. There are at
least six types of such consistent relationships that are of great
theoretical significance to students of crime and criminality.
Variation by Age
Many varieties of statistics, in many juridictions, in many
different years, collected by many types of agencies, uniformly
report such a high incidence of crime among young persons, as
compared with older persons, that it may reasonably be assumed
that there is a statistically significant difference between the rate
of crime among young adults and the rate among other age
groups. Statistics are likely to exaggerate the crime rate of young
adults. Old people may have prestige enough to avoid finger-
printing and arrest, and young children might not be arrested as
readily as either young adults or old adults, leaving young adults
to bear the responsibility for more than their share of all the
crimes committed. But there does seem to be a difference, even if
FACTS A THEORY MUST FIT 51
it is not as great as the statistics indicated when they are taken at

face value. In this sense, there are two general relationships
between age and criminality.
(A) For all crimes, taken collectively, the age of maximum
criminality is in the adolescent period. American statistics
on crimes known to the police show the maximum age of
general criminality to be in the ages eighteen to twenty-four, but
these data probably are more subject to the bias indicated above
than are the English statistics, which show that males aged twelve
and thirteen have the highest crime rates of all male age groups,
and that females aged sixteen and seventeen have higher crime
rates than any other female age group.
(B) The age of maximum criminality is not the same under all
conditions. The extent to which the crime rate among young
persons exceeds the crime rate among other age groups varies by
offence, sex, place and time:
(I) The age of maximum criminality varies with the type of
crime. Automobile theft and burglary, for example, are concen-
trated in the age group 15-19, while murder, embezzlement, and
gambling offences are committed by persons who are much older.
(2) The age of maximum criminality varies by sex. Generally
speaking, females commit crimes at later ages than do males. In
1957, for example, 74 percent of the males arrested for larceny
were under 25 years of age, but only 69 percent of the females
arrested for the same offence were under 25. Yet sex offences,
narcotic drug offences, crimes against family and children, driving
while intoxicated, and homicide and forgery appear earlier in the
lives of women than in the lives of men. In 1957, 46 percent of the
females and 30 percent of the males arrested for forgery were
under age 25. 1
(3) The age of first delinquency varies from place to place. In
areas of high rates of delinquency, the children who become
delinquent do so at an earlier age than do the children living in
areas with low rates of delinquency.2
(4) The type of crime most frequently committed by persons
of various ages varies from place to place. In some areas of large
1 Federal Bureau ofInvestigation, Uniform Crime Reportsfor the United States, Volume
XXVIII, No.2, 1957, pp. 115-117.
2 See Ernest Mannheim, Youth in Trouble, Kansas City, Missouri: Department of
Welfare, 1945, pp. 66-67.
cities, delinquent boys aged twelve and thirteen commit burgla-

ries, while in other areas delinquent boys of those ages commit
petty larcenies or engage in gang violence. In rural areas, of-
fenders of any specified age are likely to be convicted of crimes
different from these committed by offenders of the same age who
live in urban areas.
(5) For all crimes, and for each specific crime, the rate de-
creases steadily from the age of maximum criminality to the end
of life. Thus, burglary and automobile theft decrease rather
regularly after age 15-19, as does the crime rate generally;
homicide decreases rather regularly after age 20-29, where it is
concentrated. Consistently, the number of first offenders per
1000 persons of any given age decrease regularly after age 15-19.
Some crimes decrease more dramatically with increasing age than
do others; for example, the evidence is fairly conclusive that
larceny decreases in old age more than do sex offences. 3
(6) The crime rates among different age groups vary from time
to time. The crime rate among old people seems to have declined
since the time of World War I, while the crime rate among
youngsters seems to have been increasing.
(7) Both the probability that a crime will be repeated and the
length of time between first and second offences vary with the age
at which the first offence is committed. Generally speaking, the
younger a person is when he commits his first offence, the higher
the probability that he will commit a second offence and the
shorter the interval between the first offence and the second
offence. 4
In sum, the available statistics on crime tell us that young
persons have higher crime rates than older persons but that there
are variations in the ratio of young persons to old persons in the
criminal population. Thus, crime rates vary with age, but in any
age group the rates vary with specific social conditions.
3 David O. Moberg, "Old Age and Crime" Journal of Criminal Law and Criminology,
43 (March-April, 1953) pp. 764-776; Vernon Fox, "Intelligence, Race and Age as
Selective Factors in Crime", Journal of Criminal Law and Criminology, 37 (July-August,
1946), pp. 141-152.
, Thorsten Sellin, "Recidivism and Maturation", National Probation and Parole
Association Journal, 4 (July, 1958), pp. 241-250; and Hermann Mannheim and Leslie
T. Wilkins, Prediction Methods in Relation to Borstal Training, London: Her Majesty's
Stationery Office, 1955, p. 64.
Variations by Sex
Sex status is of greater statistical significance in differentiating
criminals from noncriminals than any other trait. If an investi-
gator were asked to use a single trait to predict which persons in
a town of 10,000 population would become criminals, he would
make the least mistakes if the simply chose sex status and pre-
dicted criminality for the males and noncriminality for the
females. He would be wrong in many cases, for most of the males
would not become criminals, and a few of the females would be-
come criminals. But he would be wrong in more cases if he used
some other single trait, such as age, race, family background, or a
personality characteristic. For example, crime is concentrated in
the young age group, as we have just indicated, but it is not as
concentrated in this age group as it is in the male sex. Consequent-
ly, if the investigator used age to differentiate the criminals from
the noncriminals in an unknown population he would be incor-
rect more times than he would be if he used sex as a basis for the
differentiation. As is the case with age, there are two general
relationships to be observed between crime and sex status.
(A) The crime rate for men is greatly in excess of the rate for
women-in all nations, all communities with a nation, all age
groups, all periods of history for which organized statistics are
available, and for all types of crime except for these peculiar to
women, such as infanticide and abortion. In the United States at
present the rate of arrest of males is about ten times the rate of
arrest for females; about fifteen times as many males as females
are committed to correctional institutions of all kinds; and about
twenty times as many males as females are committed to the state
and federal prisons and reformatories housing serious offenders.
Even if the statistics are grossly biased against males, they can
reasonably be interpreted to mean that some excess of crime
among males is present. 5
(B) The extent to which the crime rate among males exceeds
the crime rate among females is not the same under all conditions.
There are variations in the sex ratio in crime, just as there are
variations in the age ratio in crime:
(I) The extent to which the rate for males exceeds the rate for
5 Cj. OUo Pollak, The Criminality of Women, Philadelphia: University of Pennsyl-
vania Press, 1950, pp. 44-45, 154.
females varies from one nation to another. In Algeria, Tunis,

Japan, and Ceylon male criminals and delinquents are three to
four thousand times as numerous as female criminals. 6 In
Australia, Western Europe, the United States, and other nations
in which females have great freedom and are viewed as almost
equal in social standing to men, the rate for females is closest to
the rate for males, although males remain greatly overrepresented
in the criminal population.
(2) The extent to which the rate for males exceeds the rate for
females varies with the social positions of the sexes in different
groups within a nation. An analysis of statistics in pre-war Poland
indicated sex ratios ranging from 176 to 1163 in forty-two groups
categorized according to age, province of residence, rural-urban
residence, religion, and civil status. 7 In the United States, the
ratio is nearer to 100 among Negroes than it is among whites, and
it is probable that Negro males and females more closely resemble
each other in social position than do white males and females.
In 1957, the sex ratio among Negroes committed to New York
state prisons was 1075, but the ratio among whites was over 3000. 8
(3) The extent to which the rate for males exceeds the rate for
females varies with the size of community of residence. In
American cities the crime rate of females is closer to the crime
rate of males than is the case in rural areas and small towns. In
Massachusetts "towns"-a technical term referring to communi-
ties which have elected to retain the town meeting form of
government but which generally have populations of less than
15,000-the sex ratio for arrests of persons for crimes against the
person, such as assault, homicide and robbery, was 2500 in 1957;
but the sex ratio in cities for the same offences was only 1700.
For offences against property the sex ratio was 3600 in towns and
only IIOO in cities. 9
(4) The extent to which the crime rate among males exceeds
6 E. Hacker, Kriminalstatistische und Kriminalaetiologische Berichte, Miskole: Ludwig,

1941.
7 L. Radzinowicz, "Variability in the Sex Ratio of Criminality", Sociological
Review, 29 (January, 1937), pp. 76--102. The sex ratio always is expressed as the number
of males per 100 females. A ratio over 100, thus, means that males exceed females,
while a ratio less than 100 means that females exceed males.
8 New York State Commission of Correction. Thirty-First Annual Report, 1957. New
York: Sing Sing Prison Press, 1958, p. 405.
• Statistical Reports qf the Commissioner of Correctian, 1957. Boston: Massachusctt~
the crime rate among females varies with age. In the United
States, the sex ratio among persons committed to penal insti-
tutions tends to increase with increasing age. At ages 15-17 the
sex ratio is about 1300, while at 60-64 it is about 2500. Similar
variations have been found in Sweden and eight other European
nations.lo
the crime rate among females varies with area of residence within
a city. Generally, the higher the crime rate of an area the lower
the sex ratio in crime. However, it has been shown that some
areas with high delinquency rates also have unusually high sex
ratios among their delinquents.u
the crime rate among females varies with time. There is some
evidence that the sex ratio is decreasing. Females were 5 percent
of the persons under age 18 whose arrests were reported to the
FBI in 1938, 10 percent in 1947, and 12.7 in 1957. In time of war,
when women take over some of the jobs usually performed by
men, and when women in other ways tend to become more
nearly equal to men in social positions, the sex ratio in crime
declines.
(7) Among young criminals, the extent to which the crime rate
for males exceeds the crime rate for females varies with the degree
of integration in the family. Among delinquents from broken
homes the sex ratio is lower than it is among delinquents from
unbroken, "integrated" homes. 12 Further, there is some evidence
from specialized studies that the sex ratio in delinquency is lower
in families in which male children outnumber the females than in
families in which the numbers of each sex is more nearly equal. l •
To summarize, the available statistics indicate that males have
much higher crime rates than females but that the ratio of male
criminals to female criminals varies with specific social situations.
Department of Correction, 1958, p. 51. See also Jackson Toby, "The Differential
Impact of Family Disorganization", American Sociological Review, 22 (October, 1957),
pp.505-512.
10 See Gunnar Dahlberg, "A New Method of Crime Statistics Applied to the
Population of Sweden", Journal of Criminal Law and Criminology, 39 (September-
October, 1948), pp. 327-341.
11 Mannheim, op. cit., pp. 64-65.
10 Toby, op. cit.
13 R. F. Sletto, "Sibling Position and Juvenile Delinquency", American Journal of
Sociology, 39 (March, 1934), pp. 657-659.
Variations by Race
A number of studies have shown that in the United States,
Negroes are more likely to be arrested, indicted, and convicted
than are whites who commit the same offences. 14 Similarly, it has
been shown that Negroes have less chance than whites to receive
probation, a suspended sentence, parole, commutation of a death
sentence, or a pardon. 16 Thus, almost any "index" of the crime
rate is likely to exaggerate the rate for Negroes, as compared with
the rate among whites. However, it is also true that many crimes
committed by Negroes against other Negroes receive no official
attention from the police or criminal courts, and this practice of
overlooking some crimes offsets to some unknown degree the bias
in other arresting, reporting, and recording practices. At least
three localized studies have shown that the racial membership of
the victim is of great importance in determination of the official
reactions to crimes committed by Negroes. 16 Like the situation
with age and sex, there are two general relationships between
crime and race, though here our observations are confined to the
United States.
(A) The official statistics indicate that the crime rate of
Negroes exceeds the rate among whites. The number of arrests of
Negroes per lOO,OOO Negroes is about three times the number of
arrests ofwhites per 100,000 whites. The rate of commitment ofNe-
groes to state and federal prisons is about six times the white rate.
(B) The extent to which the crime rate among Negroes ex-
ceeds the crime rate of whites varies with social conditions. In
some conditions the rate for Negroes is not as far in excess of the
rate for whites as it is in other conditions, and in still other con-
ditions the rate for Negroes is lower than the rate for whites:
(I) The extent to which the rate for Negroes exceeds the rate
14 See, for example, Edwin M. Lemert and Judy Rosberg, "The Administration
of Justice to Minority Groups in Los Angeles County", University of California Publi-
cations in Culture and Society, Vol. II, No.1, 1948, pp. 1-28.
15 Thorsten Sellin, "Race Prejudice in the Administration of Justice", American
Journal of Sociology, 41 (September, 1935), pp. 212-217; and Sidney Axelrad, "Negro
and White Institutionalized Delinquents", American Journal ofSociology, 57 (May, 1952,
pp. 569-574.
16 Guy B. Johnson, "The Negro and Crime", Annals of the American Academy of
Political and Social Science, 217 (September, 1941), pp. 93-104; James D. Turner,
Differential Punishment in a Bi-Racial Community, Unpublished Master's Dissertation,
Indiana University, 1948; andJames D. Turner, Dynamics ofCriminalLaw Administration
in a Bi-Racial Community of the Deep South. Unpublished Ph. D. Dissertation, Indiana
University, 1956.
for whites varies with region of the United States. The excess is
highest in the Western states and lowest in the Southern states,
with Northern states occupying an intermediate positionY In
Philadelphia in 1954, Negroes made up twenty percent of the
population but accounted for fifty percent of the arrests ;18 and in
Michigan at about the same time Negroes were seven percent of
the population and about forty percent of the prison population. 19
The differences in rates are not distributed in the same way for all
offences; for homicide, for example, the difference is greatest in
the South and least in New England.
(2) The extent to which the crime rate of Negroes exceeds the
crime rate of whites varies with sex status. In 1954, 177 Negro
women were committed to New York state prisons for each 100
white women committed, but 123 white men were committed for
each 100 Negro men. Although rates cannot be calculated, be-
cause the exact numbers of the four categories of persons in the
general population of the state are unknown, it is probable that
the excess of crime among Negro women was much greater than
the excess among Negro men.
crime rate of whites varies with offence. When based on imprison-
ment rates, the excess is greatest for assault and homicide and
lowest for rape. One study indicated that in Virginia the rates for
forgery and for drunken driving actually are lower among Negroes
than among whites. 20 Another study indicated that large cities
with a high percentage of Negroes in their populations had
higher rates for murder than did large cities with only small
percentages of Negroes. 21 It also is known that the excess of crime
among Negroes is higher for second offences than it is for first
offences.
17 The high ratio in the West may be due in part to the fact that Negroes in that
area tend to be unduly concentrated in the young adult group and in cities, both of
which have high crime rates.
18 William H. Kephart, "The Negro Offender", American Journal of Sociology, 60
(July, 1954), pp. 46--50.
19 Vernon Fox andJoan Volakakis, "The Negro Offender in a Northern Industrial
Area", Journal of Criminal Law and Criminology, 46 (January-February, 1956), pp. 641-
647.
20 Workers' Writers Programme, The Negro in Virginia, New York: Hastings House,
1940, p. 341.
U James E. McKeown, "Poverty, Race and Crime", Journal of Criminal Law and
Criminology, 39 (November-December, 1948), pp. 480--483.
(4) The extent to which the crime rate of Negroes exceeds that
of whites varies with the area of residence within a city. Studies of
Houston and Baltimore show that the Negro delinquency rate is
lowest in those areas having the greatest proportion of Negroes in
their populations, and highest in those areas with relatively low
proportions of Negroes to whites. 22 Although these data do not
necessarily indicate that the excess ofcrime is greatest in areas where
whites and Negroes are not segregated, this is probably the case.
crime rate of whites varies with time. There is no precise evidence
on this point, but it seems probable that the amount of excess has
been increasing. An earlier study indicated that in three decades
the Negro juvenile delinquency rate in Chicago increased seven
times, while the Negro population increased only three times. 23
The comparable rate and population increases for whites is not
available.
In short, such statistics as are available indicate that Negroes
have higher crime rates than whites but that the ratio of crime
rates among Negroes to crime rates among whites varies with
specific social situations.
Variations by Nativity
During the years when immigration was at its height, it was
generally believed that there was an undue amount of crime
among the foreign born. While assimilation of vast numbers of
immigrants is no longer a serious social problem in the United
States, the data on crime among immigrants remains a sociological
problem of great theoretical significance. The following two
observations about nativity and crime can be made.
(A) Such statistics as are available indicate that the crime rate
among the foreign born in the United States is less than that of
native whites with the same age, sex, and rural-urban distribution.
When computed on the basis of population numbers alone, the
crime rate of immigrants is about half that ofwhites. 24 Similarly,
.2 The Houston Delinquent in His CommuniV' Setting, Houston: Bureau of Research,
Council of Social Agencies, 194-5, pp. 22-25; Bernard Lander, Towards an Understanding
ofJuvenile Delinquency, New York: Columbia University Press, 1954, p. 82.
23 Earl R. Moses, "Community Factors in Negro Delinquency", Journal of Negro
Education, 5 (April, 1936), pp. 220-227.
24 See Arthur Lewis Wood, "Minority Group Criminality and Cultural Inte-
gration", Journal ofCriminal Law and Criminology, 37 (March-April, 1947), pp. 498-510.
in Australia the crime rate of native Australians is about two

times the rate for the immigrants arriving since World War 11.25
When correction is made for the age and sex distributions of the
two groups, the crime rates become more nearly equal but still
show less crime among immigrants than among native whites.
(B) The extent to which the crime rate of native whites ex-
ceeds the rate among immigrants is not the same under all social
conditions:
(I) The extent to which the crime rate among native whites
exceeds the crime rate of the foreign born various with offence.
Certain types of crime are characteristic of one immigrant group,
while another type of crime is characteristic of a different immi-
grant group. Usually, these same types of crime are also character-
istic of the home countries. Italy has a high rate of homicide, and
Italian immigrants in the United States also have a high homicide
rate. Drunkenness is rare in Italy but frequent in Ireland; Italian
immigrant groups have low arrest rates for drunkenness, but
Irish immigrants have rather high rates for that offence. For
some types of crime, such as embezzlement and other white-
collar crimes, the crime rate among immigrants is extraordinarily
low.
(2) The extent to which the crime rate of native whites ex-
ceeds the rate among foreign born persons varies from one immi-
grant group to another. In 1955, Eastern European immigrants
in Australia had crime rates about three times as high as Southern
European immigrants and almost twice as high as Northern
European immigrants. 26 Similarly, in the United States in the
years when immigration was at its height, persons ofIrish nativity
had crime rates three to five times as high as German immigrants.
The crime rate among Japanese immigrants is exceptionally
low. 27
ceeds the rate among the foreign born varies from one native
white group to another. The native white sons of immigrants tend
25 Commonwealth Immigration Advisory Council, Third Report of the Committee

Established to Investigate Conduct of Migrants, Canberra: Commonwealth Government
Printer, 1957, p. 4.
2. Ibid., p. 18 .
• 7 Norman S. Hayner, "Delinquency Areas in the Puget Sound Area", American
Journal of Sociologr, 39 (November, 1933), pp. 314-328.
to have crime rates higher than those of their fathers but lower
than other native whites. In 1933 the rate of commitment of
immigrants to state and federal prisons was lower than the com-
mitment rate for the sons of immigrants in all but one of the
twenty-eight states for which data were available. 28 Also, sons of
immigrants tend to commit crimes characteristic of the receiving
country, while, as indicated, the immigrants themselves commit
crimes characteristic of the home country. For example, the sons
ofIrish immigrants convicted by the New York Court of General
Sessions in 1908-9 had a homicide rate about half as high as the
rate among the Irish convicted in the same court, and about twice
as high as the rate among native whites of native parentage.
However, the conviction rate for gambling was about twice as high
for the sons of immigrants as for immigrants, and about one and
one-halftimes as high for native whites of native parentage as for
the sons of immigrants. 29
(4) The amount of the excess of crime among the native born
varies with age. Among immigrants who arrive in the United
States when they are in early childhood, the crime rate is higher
than among immigrants arriving in middle age. 30 Young immi-
grants take on the relatively high crime rate of native whites to a
greater extent than do middle-aged immigrants.
ceeds that ofimmigrants varies with the length of time the immi-
grants have been in the host country. Both immigrants and their
sons tend to take on the crime rate of the specific part of the host
country in which they locate. A study of crime rates in France in
the nineteenth century indicated that migrants moving from one
province to another changed their crime rates in the direction of
the rate ofthe host province, whether the rate in the host province
was higher or lower than the rate in the province from which
they migrated. 31 In the first five years of residence in an area of
high delinquency in Los Angeles, five percent of the children in a
Mollaccan immigrant group appeared in the juvenile court; after
28 Donald R. Taft, "Nationality and Crime", American Sociological Review, 1 (Oc-
tober, 1936), pp. 724-736.
29 See Edwin H. Sutherland and Donald R. Cressey, Principles of Criminology, Sixth
Edition, Philadelphia: Lippincott, 1960, p. 146.
30 C. C. Van Vechten, "The Criminality of the Foreign-Born", Journal of Criminal
Law and Criminology, 32 (July-August, 1941), pp. 139-147.
31 Henri Jo1y, La France Criminelle, Paris; Cerf, 1889, pp. 45-46.
five more years, 46 percent appeared; and after another ten

years 83 percent of the children came before the court.32 The
delinquency rate increased with length of stay in the area, pre-
sumably because the immigrant group was assimilating American
culture, including its delinquency rates. It is safe to assume, on
the basis of the study of France, that if the immigrant group had
had a high crime rate and had settled in an area of low delin-
quency rates their rate would have decreased with the increased
length of stay.
Variations by Size of Community

Statistics from many countries, and in many periods of time,
indicate that urban areas have higher crime rates than rural
areas. Again, two general types of relationship between crime and
size of community can be observed.
(A) Official statistics indicate that the number of serious crimes
per 100,000 population increases with the size of the community.
In 1958, robberies known to the police increased from 14.1 in
American towns ofless than 10,000 population to 112.4 in cities
with populations over 250,000. Offences committed by rural
criminals might not be reported or recorded as readily as offences
committed in urban areas, but the urban rate generally so far
exceeds the rural rate that it is reasonable to conclude that there
is in fact a great excess of crime in urban places. Moreover, a
large proportion of urban crime is overlooked, and it is not at all
certain that this proportion is any less than the proportion of rural
crimes that is overlooked.
(B) The extent to which the crime rate in urban areas exceeds
the crime rate in rural areas is not the same under all conditions.
In some rural areas the crime rate, especially for some types of
offences, is higher than the rate in urban areas:
(1) The amount of the excess of crime in urban areas varies by
offence. In American cities over 250,000, murder and rape rates
are about three times as high as the rates in towns of 10,000;
burglary and larceny rates are about two times as high; automo-
bile theft about four times as high, and robbery about eight times
as high. A half-century ago, Lombroso noted that certain Italian
32 Pauline V. Young, "Urbanization as a Factor in Juvenile Delinquency",
Publications of the American Sociological Society, 24 (1930), p. 162-166.
villages had acquired reputations for special crimes. Artena, for

example, had a highway robbery rate thirty times as high as the
rest of Italy and had been noted as the home of robbers for seven
centuries. 33 Other towns were noted for burglary and swindling,
reducing the amount of excess of crime in urban areas. Certain
crimes like theft of cattle and grain are, of course, peculiarly rural.
(2) The amount of the excess of crime in urban areas varies by
area. In states containing rural mining and logging towns, the
amount of excess is less than in states where rural communities
are engaged in agriculture. 34 In an earlier period, frontier towns,
river towns, and resort towns were noted for high crime rates,
despite the fact that they were not large in size. A study of the
Chicago area by Sutherland indicated that burglaries and
robberies of banks did not decrease steadily with distance from
the city, as was the case for burglary and robbery of chain stores
and drug stores; banks in the suburbs had higher rates of robbery
and burglary than banks in the centre of the city.3D Similarly,
Lottier found that in the Detroit area, murder, assault, rape and
robbery were more concentrated in the city than were burglary,
automobile theft and larceny.36
(3) The amount of excess of crime in urban areas varies in
time. There is evidence that as improved communication and
transportation have reduced the differences between urban and
rural districts, the differences in the crime rates of the two areas
have decreased. In the United States, the rural rate has since
about 1945 been increasing more rapidly than the urban rate.
Similar trends have been noted for European countries.
Variations by Social Class

The reliability of the official statistics on the socio-economic
class backgrounds of criminals has been questioned even more
severely than have statistics on variables like age, race and area of
residence. Many persons maintain that the law enforcement pro-
33 C. Lombroso, Crime. Its Causes and Remedies, Boston: Little, Brown, 1911, p. 23 .
•4 Helen L. Yoke, "Crime in West Virginia", Sociology and Social Research, 16
(January, 1932), pp. 267-273; M. G. Caldwell, "The Extent ofJuvenile Delinquency
in Wisconsin", Journal ofCriminal Law and Criminology, 32 (July-August, 1941) pp. 145-
159; Paul Wiers, "Juvenile Delinquency in Rural Michigan", Journal ofCriminal Law
and Criminology, 30 (July-August, 1939), pp. 211-222.
35 Sutherland and Cressey, op. cit., p. 156.
36 Ibid.
cesses tend to select working class persons, just as they tend to

select Negroes. Thus it is believed that if a member of the working
class and a member of the upper class are equally guilty of some
offence the person on the lower level is more likely to be arrested,
convicted and committed to an institution. Further, some white-
collar crimes are not included in sets of official crime statistics.
Professor Reckless is confident that if statistical procedures could
be corrected the distribution of crime by social classes in the
United States would show a bi-modal curve, with high peaks for
members of the upper class and the lower class, and a low valley
for members of the middle class. 37 However, the statistics on
ordinary crime so consistently show an overrepresentation of
lower class persons that it is reasonable to assume that there is a
real difference between the behavior of the members of this class
and the members of other social classes, so far as criminality is
concerned. If this assumption is made, the following two general
observations are warranted.
(A) In the United States, membership in the working class
seems to increase the probability of becoming a criminal. Many
studies of the statistical relationship between crime and economic
conditions have been made, and most of the measures of economic
conditions are measures of social class status as well. Official
statistics indicate that the largest proportion of criminal popu-
lations come from the working class, and there is also some
evidence that the crime rate of working class persons exceeds the
crime rate of other persons. In institutionalized populations,
about two-thirds to three-fourths of the men, and about nine-
tenths of the women, are members of the working class. A Wiscon-
sin study found that 12 percent of the population of the state was
unskilled but that 33 percent of the parents of the state's insti-
tutionalized delinquent boys and 53 percent ofits institutionalized
delinquent girls were unskilled. 3s In a famous study of social
classes in a New England community it was found that the two
lower classes (lower-lower and upper-lower) constituted 57
percent of the city's population but accounted for 90 percent of
87 Walter C. Reckless, The Grime Problem, Second Edition, New York: Appleton-
Century-Crofts, 1955, pp. 28-30.
38 M. G. Caldwell, "The Economic Status of the Families of Delinquent Boys in
Wisconsin", American Journal of Sociology, 37 (September, 1931), pp. 231-239.
the arrests made during a seven-year period. 39 Many other

studies have shown similar over-representations of working class
people in criminal populations. One study found that prison
inmates rank themselves lower than they rank their fathers on
socio-economic status,40 perhaps indicating that the prisoners
were unable to maintain the family level of status, let alone im-
prove it through upward mobility. Studies of the areas of crimi-
nals' residence also indicate high crime rates in areas occupied by
persons of low socio-economic status.
(B) The extent of over-representation of working class persons
in the criminal population is not the same under all conditions.
In some situations, working class people have crime rates lower
than those of other classes:
(1) The ratio of working class persons to other persons in the
criminal population varies by social group. In the Japanese
colony in Seattle prior to World War II, delinquency and crime
rates were very low, despite the fact that the residents were of the
working class and were in as great poverty as residents of the area
surrounding the colony, who had high rates. 41 Members of the
working class who live in rural areas, similarly, have relatively
low crime rates. A study of the crime rates and the percentage of
persons on relief in Minnesota's 87 counties indicated a correlation
coefficient of only +
.21; on the other hand, crime rates and
degree of urbanization of the counties correlated +.72. 42 Females
living in areas where most residents are of low socio-economic
status have very low crime rates. Also, in extreme situations
members of some groups have literally starved to death rather
than violate laws. This was illustrated in the period of famine in
India, in 1943:
Through all these months the white Brahmin cattle wandered by the
hundreds through the streets of Calcutta, as they always have, stepping
placidly over the bodies of the dead and near-dead, scratching their plump
haunches on taxi fenders, sunning themselves on the steps of the great Clive
Street banks. No-one ever ate a cow; no-one ever dreamed of it. I never heard
89 William Lloyd Warner and Paul S. Lunt, The Social Life of a Modem Communiry,
New Haven: Yale University Press, 1941, pp. 373-377.
40 Harold Bradley and Jack D. Williams, Intensive Treatment Programme: Second
Annual Report, Sacramento: Department of Corrections, 1958, p. 16.
41 Hayner, op. cit.
42 Van B. Shaw, "The Relationship Between Crime Rates and Certain Population
Characteristics in Minnesota Counties", Journal of Criminal Law and Criminology, 40
(May-June, 1949), pp. 43-49.
of a Bengali Hindu who would not perish with all his family rather than taste.
meat. Nor was there any violence. No grocery stall, no rice warehouse, none
of the wealthy clubs or restaurants ever was threatened by a hungry mob.'s
Finally, it is obvious that most working class persons do not

become criminals and that, therefore, something other than
working class membership is involved. Even in areas of great
poverty and high crime rates, large proportions of the residents
do not become criminals.
(2) The ratio of working-class persons to other persons varies
by offence. Most studies showing high ratios of working class
persons have concentrated on crimes against property, such as
larceny and burglary. There is some evidence, however, that the
ratio is somewhat lower for sex offences, and in fact the crime
rates of the working class may be lower than those of other classes
for some sex offences. 44 Similarly, a study conducted in Detroit
indicated that working class persons are not as overrepresented in
the population of automobile thieves as they are in other delin-
quent and criminal populations. 45 And of course working class
persons have lower crime rates than other persons for offences
such as embezzlement, misrepresentation in advertising, vio-
lation of anti-trust laws, and issuing worthless stocks. The person's
position in the occupational structure determines the opportuni-
ties for some kinds of crime and also determines whether or not a
person will possess the skills necessary to perpetrating some types
of offences.
Conclusion
Although the above are only some of the social conditions with
which crime rates vary, the list is sufficiently long to enable us to
draw the important conclusion that crime is social behaviour that
is closely associated with other kinds of social behaviour. One set
of facts indicates that the crime rate is higher for young adults
than for persons in later life, higher for men than for women,
higher for Negroes than for whites, higher for native-born than
43 John Fischer, "Indias's Insoluble Hunger", Harper's Magazine, 190 (April, 1945),
pp. 438-445 .
•• Cj. A. C. Kinsey, W. B. Pomeroy, and C. E. Martin, Sexual Behavior in the Human
Male, Philadelphia: Saunders, 1948, pp. 327-393.
45 William W. WattenburgandJamesBaiistrieri, "Automobile Theft: A 'Favoured-
Group' Delinquency", American Journal of Sociology, 57 (May, 1952), pp. 575-579.
66 FACTS A THEORY MUST FIT
for foreign-born, higher in cities than in rural areas, and higher

for the working class than for other social classes. Such differences
may be described as ratios-the age ratio in crime, the sex ratio
in crime, etc. A second set of facts indicates that these ratios are
not constant. They vary in definite ways.
These ratios, and variations in ratios, make up some of the
facts which a general explanation of crime must fit. They may be
called definitive facts, for they define or limit the explanations of
crime that can be considered valid. For example, an explanation
which attributes crime to poverty helps make good sense out of
the overrepresentation of Negroes and working class persons in
the general criminal population, but the theory falls flat when it is
recalled that women, who are equally in poverty with men, have
very low crime rates; when it is recalled that immigrants, who are
probably at least as poor as their sons, sometimes have crime rates
lower than their sons; when it is recalled that even poor Negroes
and working class persons do not have high crime rates if they
live in rural areas, and so on. Similarly, an explanation of crime
in terms of a trait of aggression or other personality character-
istics must show that the characteristic is much more frequent
among men than among women, among Negro women than
white women, among young persons as compared to old persons,
etc.; and it also must show that the trait occurs infrequently
among immigrant groups, Negroes who live in segregated areas,
middle class persons, etc. Such "pathology theories" generally
organize and "make sense of" the ratios and variations in ratios,
but none of them seems to fit all the ratios and variations as well
as does the theory of differential association and differential social
organization. Since we rank theories by their ability to make the
best sense out of the most facts, Sutherland's theory must be
given a high rating, despite its defects. However, it is important
to note that while this theory makes good use of available sta-
tistics it must remain as tentative because we cannot be sure that
the relation ships it deals with are not mere statistical artifacts.
CHAPTER IV
DIFFERENTIAL ASSOCIATION, DIFFERENTIAL SOCIAL

ORGANIZATION, AND THE EPIDEMIOLOGY OF CRIME
Any theory purporting to explain a specific kind of social be-

haviour such as criminality and delinquency should have two
distinct but consistent aspects. First, there should be a statement
that explains the statistical distribution of the behaviour in time
and space (epidemiology), and from which predictive statements
about unknown statistical distributions can be derived. Second,
there should be a statement that identifies, at least by implication,
the process by which individuals come to exhibit the behaviour in
question, and from which can be derived predictive statements
about the behaviour of individuals. Concentration on either the
epidemiological segment or the individual conduct segment of a
theoretical problem is sometimes necessary, but it is erroneous
and inefficient to ignore the second segment, to turn it over to
another academic discipline, or to leave its solution to a special-
ized set of workers within a single discipline.
In some cases, data on both aspects of a problem are not avail-
able, so that a two-edged theory is impossible. For example, my
work on trust violators was concerned almost exclusively with the
process by which one becomes a criminal, but such concentration
was necessary because reliable data on the distribution of this type
of crime was not available. l Should data become available, then
the generalization about trust violators should be integrated with
a generalization about variations in trust violation. In other
cases, concentration on one phase of an explanation my be
merely a matter of interest or time. However, it might also be due
to an undesirable informal or formal division of labor-such as
that between sociologists and psychiatrists, or that indicated by
the recent development of a special Section on Social Psychology
within the American Sociological Association.
1 Donald R. Cressey, Other People's Money, Glencoe: The Free Press, 1953.
The need for integrated theories of epidemiology and indivi-

dual conduct is demonstrated in the work done on Merton's
theory of deviant behaviour. Over twenty years ago, Merton
presented a sociological statement purporting among other things,
to account for an excess of property crimes in the working class
population. However, he left unanswered (and to some extent
unasked) the question of why only a rather insignificant pro-
portion of working class persons become property offenders. 2
While a few sociologists paid attention to this theory about the
epidemiology of crime, psychiatric theory that is quite unrelated
to Merton's theory has continued to dominate explanation of
individual cases of working class (and other) criminality.3 Only in
the last five years has there been a significant attempt to identify
the processes by which the blocking of legitimate means for
achieving success, posed by Merton, might "work" to produce the
criminality of individual working class persons. Even here, the
most significant efforts have concentrated on variations in socially
structured opportunities for deviation, rather than on the social
psychological mechanisms involved in individual cases. 4
In an even more significant case, a sociological theory of the
epidemiology of crime has been neglected because it has been
viewed as only an alternative to psychiatric theories about the
process by which individuals become criminals. The implications
of Sutherland's theory for explanation of the high and low crime
rates of various categories of persons have been all but ignored,
just as the implications of Merton's theory for explanation of
individual conduct have been neglected. Yet it is clear that when
Sutherland introduced the idea of differential association he was
concerned with both phases of the general pro blem in criminology:
explaining the distribution of crime rates, and identifying the
process by which a person becomes a crimina1. 5 His idea has had
a profound effect on criminological and sociological thought,
2 Robert K. Merton, "Social Structure and Anomie", American Sociological Review,
3 (October, 1938), pp. 672-682.
3 Cj. Frank E. Hartung. "A Critique of the Sociological Approach to Crime and
Correction", Law and Contemporary Problems, 23 (Autumn, 1958), pp. 703-734.
4 Richard A. Cloward, "Illegitimate Means, Anomie, and Deviant Behavior",
American Sociological Review, 24 (April, 1959), pp. 164-176; and Richard A. Cloward
and Lloyd E. Ohlin, Delinquency and Opportunity, Glencoe: The Free Press, 1960.
5 We do not mean to imply that this is the only general problem in criminology. In
addition to the two phases of the general problem of etiology, criminologists are
concerned with the sociology of criminal law and the sociology of punishment.
THE EPIDEMIOLOGY OF CRIME 69
despite the fact that it has become a centre of controversy. Signifi-

cantly, the controversy has concentrated on the capacity of the
statement to portray accurately the process by which individuals
become criminals, 6 and its capacity to explain the distribution of
crime and delinquency has scarcely been studied or discussed.
Yet we shall see that, in a very real sense, Sutherland was trying
to do for criminology what Darwin did for biology. Although
such an observation is rather pretentious when the range of
phenomena included in the scope of Darwin's theory is compared
with the range included in Sutherland's, each man did try to
state a principle accounting for the presence or absence of "de-
viant" phenomona, and then also tried to specify the process by
which "deviancy" comes to be present in individual cases.
It is unfortunately true that only a very meticulous and thought-
ful reader of Sutherland's formal statement can discern that it is
concerned with making sense of the gross facts about crime rates,
rather than concentrating exclusively on the problem of ex-
plaining variations in the criminality of individuals. 7 However,
even casual examination of Sutherland's writings clearly indicates
that he was greatly, if not primarily, concerned with organizing
and integrating the factual information about crime rates. In his
account of how the theory of differential association developed,
he made the following two points, which are sufficient to establish
his concern for the epidemiology of crime.
More significant for the development of the theory were certain questions
which I raised in class discussions. One of these questions was, Negroes, young-
adult males, and city dwellers all have relatively high crime rates: What do
these three groups have in common that places them in this position? Another
question was, even if feeble-minded persons have a high crime rate, why do
they commit crimes? It is not feeble-mindedness as such, for some feeble-
• See below, pp. 81-89.

7 One of Sutherland's own students, colleagues, and editors has said "Much that
travels under the name of sociology of deviant behaviour or of social disorganization
is psychology - some of it very good psychology, but psychology. For example,
Sutherland's theory of differential association, which is widely regarded as pre-
eminently sociological, is not the less psychological because it makes much of the
cultural milieu. It is psychological because it addresses itself to the question: How do
people become the kind of individuals who commit criminal acts? A sociological
question would be: What is it about the structure of social systems that determines the
kinds of criminal acts that occur in these systems and the way in which such acts are
distributed within these systems?" Albert K. Cohen, "The Study of Social Dis-
organization and Deviant Behaviour", Chapter 21 in Robert K. Merton, Leonard
Broom, and Leonard S. Cottrell, Jr., editors, Sociology Today, New York: Basic Books,
1959, p. 462.
minded persons do not commit crimes. Later I raised another question which
became more important in my search for generalizations. Crime rates have a
high correlation with poverty if considered by areas of a city but a low corre-
lation if considered chronologically in relation to the business cycle; this
obviously means that poverty as such is not an important cause of crime. How
are the varying associations between crime and poverty explained?8
The hypothesis of differential association seemed to me to be consistent with

the principal gross findings in criminology. I t explained why the Mollaccan
children became progressively delinquent with length of residence in the
deteriorated area of Los Angeles, why the city crime rate is higher than the
rural crime rate, why males are more delinquent than females, why the crime
rate remains consistently higher in deteriorated areas of cities, why the juvenile
delinquency rate in a foreign nativity is high while the group lives in a deteri-
orated area and drops when the group moves out of the area, why second-
generation Italians do not have the high murder rate their fathers had, why
Japanese children in a deteriorated area of Seattle had a low delinquency rate
even though in poverty, why crimes do not increase greatly in a period of
depression. All of the general statistical facts seem to fit this hypothesis.·
In appears, then, that in writing about differential association

Sutherland was trying to say, for example, that a high crime rate
in urban areas can be considered the end product of social con-
ditions that lead to a situation in which relatively large pro-
portions of persons are presented with an excess of criminal be-
haviour patterns. Similarly, the fact that the rate for all crimes is
not higher in some urban areas than it is in some rural areas can
be attributed to differences in conditions which affect the proba-
bilities of exposure to criminal behaviour patterns. 10 The im-
portant general point is that in a multi-group type of social
organization, alternative and inconsistent standards of conduct
are possessed by various groups, so that an individual who is a
member of one group has a high probability of learning to use
legal means for achieving success, or learning to deny the im-
portance of success, while an individual in another group learns
to accept the importance of success and to achieve it by illegal
means. Stated in another way, there are alternative educational
proceS8es in operation, varying with groups, so that a person may
be educated in either conventional or criminal means of achieving
8 Edwin H. Sutherland, "Development of the Theory", in Albert Cohen: Alfred

Lindesmith, and Karl Schuessler, Editors, The Sutherland Papers, Bloomington: Indiana
University Press, 1956, p. 15.
9 Ibid., pp. 19-20. See also the statement on p. 5 above.
10 Cj. Henry D. McKay, "Differential Association and Crime Prevention: Problems
of Utilization", Social Problems, 8 (Summer, 1960), pp. 25-37.
success. Sutherland sometimes called this situation "differential

social organization" or "differential group organization," and he
proposed that "Differential group organization should explain
the crime rate, while differential association should explain the
criminal behaviour of a person. The two explanations must be
consistent with each other."ll
It should be noted that, in the quotations above, Sutherland
referred to his statement as both a "theory" and a "hypothesis,"
and did not indicate any special concern for distinguishing be-
tween differential association as it applies to the epidemiology of
crime and differential association as it applies to individual
conduct. In order to avoid controversy about the essential
characteristics of theories and hypotheses, we prefer to call differ-
ential association, as it is used in reference to crime rates, a
"principle." Because sociology seems to be dominated by a logic
and methodology derived from physics, through psychology,
sociologists are reluctant to label a statement "theory" unless it is
a generalization sufficiently detailed to permit derivation of
predictive hypotheses that can be put to test by gathering new
facts. Nevertheless, it might be argued that many "theories" in
sociology are in fact principles that order known facts about rates
-now called epidemiology-in some way, and that they only in
very general ways specify directions for accumulation of new
facts that might prove them wrong. Durkheim, for example,
invented what may be termed a "principle of group integration"
to account for, organize logically, and integrate systematically the
data on variations in suicide rates. He did not invent a theory of
suicide, derive hypotheses from it, and then collect data to de-
termine whether the hypotheses were correct or incorrect. He
tried to "make sense" of known facts about rates, and the princi-
ple he suggested remains the most valuable idea available to
persons who would understand the differences in the rates of
suicide between Protestants and Jews, urban dwellers and rural
dwellers, etc.
We suggest, similarly, that Sutherland's statement is a "princi-
ple of normative conflict" which proposes that high crime rates
occur in societies and groups characterized by conditions that
lead to the development of extensive criminalistic subcultures.
11 Sutherland, "Development of the Theory", op. cit., p. 21.
Sutherland made some attempt to account for the origins of these

subcultures,12 but he did not concentrate on this problem any
more than Durkheim concentrated on attempting to account for
the fact that Jewish families seemed more closely integrated than
non-Jewish families. He "made sense" of variations in crime rates
by observing that modern societies are organized for crime as well
as against it, and then observing further that crime rates are
unequally distributed because of differences in the degree to
which various categories of persons participate in this normative
conflict.
The Methodology of Darwinism

The value of general principles like "normative conflict" can be
further established by returning to a comparison of the general
approaches used by Darwin and Sutherland. While no one can
claim that Sutherland's contribution has been anywhere near as
important to science, or to humanity, as has Darwin's, a compari-
son of the general methodology of the two men makes clear what
it was that Sutherland was trying to do. Although Darwin's
contribution is called the "theory of evolution" and Sutherland's
is called the "theory of differential association," both had two
distinct parts. There is a remarkable similarity in the goals of the
two "theories," the logic on which they are based, and the defects
in them. Darwin invented the principle of natural selection, with
its implication of evolution, to account for the strange distribution
of "deviant" biological specimens and the forms of plant and
animal life. Next, he tried to specify the process by which this
principle of natural selection "works" in individual cases. Suther-
land invented the principle of normative conflict to account for
the strange distribution of high and low crime rates; he then tried
to specify the mechanism by which this principle "works" to
produce individual cases of criminality. The mechanism proposed
is differential association:
The second concept, differential association, is a statement of [normative]
conflict from the point of view of the person who commits the crime. The two
12 Edwin H. Sutherland and Donald R. Cressey, Principles of Criminology. Sixth

Edition, Philadelphia: Lippincott, pp. 82-92.
kinds of culture impinge on him or he has association with the two kinds of
cultures and this is differential association. ls
Darwin had three principal advantages over Sutherland. First,

his emphasis was on the "epidemiological" part of his theory,
rather than on the "individual conduct" part. His principle of
natural selection ordered a wide range of facts that had been
minutely detailed by thousands of careful observers. He knew
quite precisely what facts his principle had to fit. For at least a
century prior to Origin of Species, observation of the wonders of
nature had been almost a national pastime in England. Great
numbers of persons who, like Darwin, had little formal training
in science were recording observations of biological and physical
phenomena.1 4 In the fifty years before Origin, at least a half dozen
persons, including Darwin's grandfather, tried to put order into
all these data by formulating something like a principle of natural
selection. Mter publication of Origin of Species, the principle be-
came a "hit" because it stirred up religious controversy, but also
because thousands of amateur scientists could, like the profes-
sionals, check it against their own small world of observations and
agree or disagree.
In contrast, Sutherland presented his theory to a world that
knew little about crime and cared little about understanding it.
Twenty-five years ago, the study of crime probably was more
popular than it is at present, but detailed, precise, observations
were being made by only a handful of persons. As today, careful
observations were being made largely by academic sociologists,
and the amateurs in the field were more concerned with doing
something about crime than they were in knowning about it.
Moreover, much work in criminology was, and still is, sporadic
and slipshod. As we indicated in Chapter II, lack of criminal
statistics that are known to be reliable, together with theoretical
emphasis on the multiple factor notion, have resulted in a situation
such that we cannot be sure that the "facts" about delinquency
and crime are facts at all.
13 Sutherland, "Development of the Theory", op. cit., pp. 20-2l.
14 The popularity of scientific concern was a social movement growing out of
Calvinism, which admonished its followers to observe God's laws by observing His
works, the wonders of nature. Fashionable English ladies carried pocket microscopes,
which they would train on flowers and insects while strolling through the garden. See
Gerald Dennis Meyer, The Scientific Lady jn England, 1650-1750. Los Angeles: The
University of California Press, 1955.
Sutherland tried to induce order in what facts we have, sparse

as they may be. His principle organized only a narrow range of
observations which were not always valid, and which were known
to only a handful of dedicated souls.
Moreover, Sutherland handicapped himself by presenting the
principle as an appendage to a well-established textbook, and by
not explicitly trying to show in a formal statement how the princi-
ple helped to integrate and organize the existing data on crime.
He needed to confront the reader with an overwhelming number
of valid observations that somehow seemed less likely to be mere
happenstance occurrences, after his principle was stated, than
they had seemed before it was stated. This was not done by
Sutherland, and it has not been done by anyone else. It seems
likely that Sutherland did not try to promote his principle be-
cause of a characteristic he had in common with Darwin-ex-
treme modesty. It is conceivable that he did not completely com-
mit himself to his principle, in the form of a major publication
like Origin, for the same reason that Darwin published four mono-
graphs on barnacles, none of them containing any reference to his
principle, between the time he formulated the principle and the
time he published it.15 Sutherland was well aware of the failure
of the previous theories about crime, and he did not want to get
too committed to his own formulation. 16
Second, it was to Darwin's advantage that Origin of Species
eventually attracted the attention of his professional colleagues;
Sutherland's theoretical work is still so unknown even among
sociologists that in at least two instances the words "differential
association" have been invented as concepts describing pheno-
mena quite unrelated to crimeY Publication in a textbook, as
compared to a monograph, probably had some effect on this
difference. Also, there is a tendency among sociologists to think of
criminology as a distinct discipline, rather than observing that
criminologists like Sutherland are interested in data on crime for
16 It is quite possible that Darwin never would have published his principle had it
not been independently formulated by Alfred R. Wallace, who threatened to scoop
him. See Garrett Hardin, Nature and Man's Fate, New York: Rinehart, 1959, pp. 42-45.
16 Sutherland, "Development of the Theory", op. cit., p. 17.
17 Ronald Freedman, Amos H. Hawley, Werner S. Landecker, and Horace M.
Miner, Principles of Sociology, First Edition, New York: Holt, p. 235-238; David Gold,
"On Description of Differential Association", American Sociological Review, 22 (August,
1957), pp. 448-450.
the same theoretical reasons that other sociologists are interested

in data on industry, family life and politics. Sutherland's princi-
ple remains unknown to most psychiatrists, psychologists, and
social workers.
The third, and by far the most important, advantage Darwin
had over Sutherland was a set of research workers who appeared
on the scene to correct him as well as criticize him. In simple fact,
Sutherland's Mendel, Fisher, and Wright have not appeared. It
turned out that Darwin was quite wrong after all. His principle
of natural selection became one of the most important ideas in the
history of man, but it was founded on an erroneous conception of
the mechanisms by which heredity takes place in individual cases.
Darwin adhered to the incorrect but popular "paint pot" theory
that viewed heredity as a blending process, and because of this
adherance he eventually had to join the Lamarckian geneticists,
holding that mysterious particles called "pangenes" are modified
by environmental conditions and are then gathered together to
form the hereditary elements of the sperm or egg. IS Although
Mendel "corrected" Darwin when he published his discovery in
1866, his work did not become known and understood until the
turn of the century. Since then, research on genetics has given
Darwin's principle what it most needed-mathematically precise
statements of the process by which natural selection "works." What
has remained of Darwin himself is his important first principle, the
principle of natural selection, and not his ideas about genetics.
There are no known published accounts of research that would
carefully quantify or in some other way induce exact precision in
Sutherland's statement of the process by which normative conflict
"works" to produce criminality in individual cases. The most
significant work has been done by Daniel Glaser and james
F. Short,jr. Although critics agree that the differential association
statement over-simplifies the process by which normative conflict
"gets into" persons and produces criminality, an acceptable
substitute that is consistent with the principle of normative
conflict has not appeared.
The Value oj Differential Association

We have suggested that Sutherland, like Darwin, tried to
18 Hardin, op. cit., p. 118.
formulate a principle that would organize available factual infor-

mation on a type of deviation and then tried to specify the process
by which that principle operates in individual cases of deviation.
In Chapter V we will show that Sutherland's critics have argued
that his specification of the latter process is incorrect, just as
Darwin's specification of the hereditary process was incorrect.
But inaccuracy in specifying the mechanism for becoming a
criminal does not necessarily negate the value of the general
principle, as the history of Darwinism has shown.
As an organizing principle, normative conflict makes under-
standable most of the variations in crime rates discovered by
various researchers and observers, and it also focusses attention on
crucial research areas. 19 The principle does not make good sense
out of all the facts, but it seems to make better sense out of more
of the facts than do any of the alternative theories. Probably we
should not expect the principle to fit all the observations to which
it Inight be applied. As the physicist-philosopher Phillipp Frank
has said, "There is certainly no theory which is in complete
agreement with all our observations. If we require complete
agreement, we can certainly achieve it by merely recording the
observations." 20
On the other hand, it also seems safe to conclude that differ-
ential association is not a precise statement of the process by
which one becomes a criminal. The idea that criIninality is a
consequence of an excess of intimate associations with criminal
behaviour patterns is valuable because, for example it negates
assertions that deviation from norms in simply a product of being
emotionally insecure or living in a broken home, and then indi-
cates in a general way why only some emotionally insecure
persons and only some persons from broken homes commit
crimes. Also, it directs attention to the idea that an efficient
explanation of individual conduct is consistent with explanations
of epidemiology. Yet the statement of the differential association
process is not precise enough to stimulate rigorous empirical test,
19 ('J. Llewellyn Gross, "Theory Construction in Sociology: A Methodological

Inquiry", Chapter 17 in Llewellyn Gross, editor, Symposium on Sociological Theory,
Evanston: Row, Peterson, 1959, pp. 548-555.
20 Philosophy of Science, Englewood Cliffs, New Jersey: Prentice-Hall, 1957, p. 353;
quoted by Daniel Glaser, "Differential Association and Criminological Prediction",
Social Problems, 8 (Summer, 1960), pp. 6-14.
and it therefore has not been proved or disproved. This defect is

shared with broader social psychological theory. As Schrag has
pointed out, "The individual internalizes the norms of his group,"
and "Stimulus patterns that are active at the time of a response
eventually acquire the capacity to elicit the response," are illus-
trations of assertions which cannot be confirmed or denied but
which stand, at present, as substitutes for descriptions of the
process by which persons learn social behaviour. 21 Criminological
theory can be no more precise than the general sociological
theory and general social psychological theory of which it is a
part.
It is important to observe, however, that the "individual
conduct" part of Sutherland's statement does order data on
individual criminality in a general way and, consequently, might
be considered a principle itself. Thus, "differential association"
may be viewed as a restatement of the principle of normative
conflict, so that this one principle is used to account for the
distribution of criminal and non-criminal behaviour in both the
life of the individual and in the statistics on collectivities. In this
case, both individual behaviour data and epidemiological rate
data may be employed as indices of the variables in the principle,
thus providing two types of hypotheses for testing it. 22 Glaser has
recently shown that differential association makes sense of both
the predictive efficiency of some parole prediction items and the
lack of predictive efficiency of other items. 23 In effect, he tested
the principle by determining whether parole prediction pro-
cedures which could have proven it false actually failed to prove it
false. First, he shows that a majority of the most accurate pre-
dictors in criminological prediction research are deducible from
differential association theory while the least accurate predictors
are not deducible at all. Second, he shows that this degree of
21 Clarence Schrag, "Some Foundations for a Theory of Correction", Chapter 8
Donald R. Cressey, Editor, The Prison: Studies in Institutional Organization and Change,
New York: Holt, Rinehart and Winston, 1961, p. 314.
22 I am indebted to Daniel Glaser for calling this point to my attention.
2. "Differential Association and Criminological Prediction, op. cit. See also Daniel
Glaser, "A Reconsideration of Some Parole Prediction Factors", American Sociological
Review, 19 (June, 1954), pp. 335-341; and "The Efficiency of Alternative Approaches
to Parole Prediction", American Sociological Review, 20 (June, 1955), pp. 283-287; and
Daniel Glaser and Richard F. Hangren, "Predicting the Adjustment of Federal
Probationers", National Probation and Parole Association Journal, 4 (July, 1958), pp.
258-267.
78 THE EPIDEMIOLOGY OF CRIME
accuracy does not characterize alternative theories. Finally, he

notes that two successful predictors of parole violation-type of
offense and non-criminal employment opportunities-are not
necessarily deducible from the theory, and he suggests a modifi-
cation that would take this fact into account.
Future research on differential association might specify in
more detail the mechanisms by which one becomes a criminal,
but it probably will do so only if sociologists recognize the epi-
demiological principle with which the process is consistent. While
it might be argued that Darwin's "theory of evolution" can only
be illustrated, not tested, it is clear that genetics has been pro-
foundly affected by Darwin's scientific desire to generalize broadly
on his, and others' observations of the distribution of species. 24
Similarly, Sutherland's "theory of normative conflict" tends to be
tautological and might not be testable. Nevertheless, it is a
starting point for theory of criminal epidemiology, and its counter-
part, differential association, indicates in a general way the
process which should be closely studied as a first step to develop-
ment of efficient theory of individual criminal conduct.
2. Cj. Garrett Hardin, "The Competitive Exclusion Principle", Science, 13 I (April,

1960), pp. 1292-1298.
II
SOME CRITICISMS OF THE THEORY

CHAPTER V
SOME POPULAR CRITICISMS OF

The "differential association" part of Sutherland's theory in

contrast to the "differential social organization" part, purports to
identify the general process by which persons become criminals.
This part of the theory, like Sutherland's subsequent work and
the work of those who have used his theory, is social psychological
in emphasis, and it relies heavily upon the theories of George
Herbert Mead. 1 The essential idea here is that all criminal con-
duct is learned in a process of social interaction involving an
"excess" of associations with criminal behaviour patterns. This
basic idea and the manner in which Sutherland expressed it has
been the recipient of criticism from many quarters. Identification
of some of the defects that various critics have found in Suther-
land's statement should make his principle clearer.
Five principal types of criticism have been advanced in the
literature. It would be incorrect to assume that a criticism ad-
vanced by many readers is more valid or important than one
advanced by a single reader, but commenting on every criti-
cism would take us too far afield. We can only mention, without
elaboration, some of the criticisms advanced by only one
or two authors. It has been stated or implied that the theory
of differential association is defective because it omits consi-
deration of free will, 2 is based on a psychology assuming rational
deliberation,3 ignores the role of the victim,4 does not explain the
origin of crime, 5 does not define terms such as "systematic" and
1 George H. Mead, Mind, Selfand Society, Chicago: University of Chicago Press, 1934.
2 Robert G. Caldwell, Criminology, New York: Ronald Press, 1956, p. 182.
3 S. Kirson Weinberg, "Theories of Criminality and Problems of Prediction",
Journal qfCriminal Law and Criminology, 45 (November-December, 1954), pp. 412-429.
4 Marshall B. Clinard, "The Sociology of Delinquency and Crime", in Joseph
Gittler, Editor, Review qf Sociology, New York: Wiley, 1957, p. 479.
5 Clarence R.Jeffery, "An Integrated Theory of Crime and Criminal Behaviour",
Journal of Criminal Law and Criminology, 49 (March-April, 1959), p. 537.
82 CRITICISMS OF THE THEORY
"excess,"6 does not take "biological factors" into account,7 is of

little or no value to "practical men,"s is not comprehensive e-
nough because it is not interdisciplinary,9 is not allied closely e-
nough with more general sociological theory and research,l° is too
comprehensive because it applies to non-criminals,l1 and assumes
that all persons have equal access to criminal and anti-criminal be-
haviour patterns.12 Some of these comments represent pairs of
opposites, one citicism contradicting another, and others seem
to be based on one or more of the errors described in Chapter I.
Still others are closely allied with the five principal types of criti-
cism, and we shall return to them.
One popular form of "criticism" of differential association is
not, strictly speaking, criticism at all. At least ten scholars have
speculated that some kinds of criminal behaviour are exceptional
to the theory. Thus, it has been said that the theory does not apply
to rural offenders,I3 to landlords who violated OPA regulations,14
to criminal violators of financial trust,15 to "naive check for-
• Arthur L. Leader, "A Differential Theory of Criminality", Sociology and Social
&search,26 (September, 1941), pp. 45-53; Caldwell, op. cit.; Marshall B. Clinard,
"Criminological Research", in Robert K. Merton, Leonard Broom, and Leonard
Cottrell, Editors, Sociology Today, New York: Basic Books, 1959, pp. 510-513; James
F. Short,Jr., "Differential Association and Delinquency", Social Problems, 4 (January,
1957), pp. 233-239.
7 Harry Elmer Barnes and Negley K. Teeters, New Horizons in Criminology, Third
Edition, Englewood Cliffs, New Jersey: Prentice-Hall, 1959, p. 159; Caldwell, op. cit.
p. 182; Howard B. Gill, "An Operational View of Criminology", Archives of Criminal
Psycho-dynamics, (October, 1957), pp. 289-291; Sheldon Glueck, "Theory and Fact
in Criminology", British Journal of Delinquency, 7 (October, 1956), pp. 98-99; Olaf
Kinberg, "Kritiska reflexioner over den differentiella associationhypotesen", Chapter
24 in Ivar Agge, Gunnar Boalt, Bo Gerle, Maths Heuman, Carl-Gunnar Janson, Olaf
Kinberg, Sven Rengby, Torgny Segerstedt, and Thorsten Sellin, Kriminologi, Stock-
holm: Wahlstrom and Widstrand, 1955, pp. 415-429.
8 Barnes and Teeters, op. cit., p. 210.
9 Ibid., p. 162; Caldwell, op. cit., p. 182; Gill, op. cit., p. 284; Glueck, op. cit., pp. 105,
108; Howard Jones, Crime and the Penal System, London: University Tutorial Press,
1956, p. 95.
10 Clarence Schrag. "Review rif Principles of Criminology", American Sociological Review,
20 (August, 1955), pp. 500-501.
11 Gill, op. cit., p. 284; Jeffery, op. cit., p. 537.
12 Richard A. Cloward, "Illegitimate Means, Anomie, and Deviant Behaviour",
American Sociological Review, 24 (April, 1959), pp. 164-176; James F. Short, Jr.,
"Differential Association as a Hypothesis: Problems of Empirical Testing", Social
Problems,8 (Summer, 1960), pp. 14-25.
13 Marshall B. Clinard, "The Process of Urbanization and Criminal Behaviour",
American Journal of Sociology, 48 (September, 1942), pp. 202-213; Marshall B. Clinard,
"Rural Criminal Offenders", American Journal of Sociology, 50 (July, 1944), pp. 38-45.
14 Marshall B. Clinard, "Criminological Theories of Violations of Wartime
Regulations", American Sociological Review, 11 (June, 1946). pp. 258-270.
15 Donald R. Cressey, "Application and Verification ofthe Differential Association
POPULAR CRITICISM 83
gers,"16 to "white-collar criminals,"17 to perpetrators of "indivi-

dual" and "personal" crimes,18 to irrational and impulsive
criminals,19 to "adventitious" and/or "accidental" criminals,20
to "occasional," "incidental," and "situational" offenders,21 to
murderers, non-professional shoplifters and non-career type of
criminals,22 to persons who commit crimes of passion,23 and to
men whose crimes were perpetrated under emotional stress. 24 It is
significant that only the first five comments-those referring to
rural offenders, landlords, trust violators, check forgers, and some
white-collar criminals-are based on research. It also is signifi-
cant that at least two authors have simply stated that the theory is
subject to criticism because there are exceptions to it; the kind of
behaviour thought to be exceptional is not specified. 26
The fact that most of the comments are not based on research
means that the "criticisms" actually are proposals for research.
Should a person conduct research on a particular type of offender
and find that the theory does not hold, a revision is called for,
providing the research actually tested the theory, or part of it.
As indicated, this procedure has been used in five instances, and
these instances need to be given careful attention. 26 But in most
cases, there is no evidence that the kind of behaviour said to be
exceptional is exceptional. For example, we do not know that
"accidental" or "incidental" or "occasional" criminals have not
Theory", Journal of Criminal Law and Criminology, 43 (May-June, 1952), pp. 43-62.
See Chapter VII below.
18 Edwin M. Lemert, "Isolation and Closure Theory of Naive Check Forgery",
Journal of Criminal Law and Criminology, 44 (September-October, 1953), pp. 293-307.
I. Marshall B. Clinard, Sociology of Deviant Behaviour, New York: Rinehart, 1957,
p. 240; Richard R. Korn and Lloyd W. McCorkle, Criminology and Penology, New York:
Holt, 1959, pp. 299-300.
18 Marshall B. Clinard, "Criminal Behaviour is Human Behaviour", Federal
Probation, 13 (March, 1949), pp. 21-27; "Research Frontiers in Criminology", British
Journal of Delinquency, 7 (October, 1956), pp. 110-122; Sociology of Deviant Behaviour,
op. cit., p. 229; and "Criminological Research", op. cit., p. 512.
19 Mabel A. Elliott, Crime in Modern Society, New York: Harper and Brothers, 1952.
p. 402; George B. Void, Theoretical Criminology, New York; Oxford University Press,
1958, pp. 197-198.
20 Clinard, "Criminological Research", op. cit., p. 511; Elliott, op. cit., p. 402;
Jeffery, op. cit.; Daniel Glaser, "Criminality The ories and Behavioural Images",
American Journal of Socio/Qgy, 61 (March 1956), p. 441.
21 Elliott, op. cit., p. 402; Clinard, "Criminological Research", op. cit., p. 512.
22 Ibid.
23 Jeffery, op. cit.
2. Elliott, op. cit., p. 347-348.
25 Barnes and Teeters, op. cit., p. 159; Taft, op. cit., p. 340.
2& See the discussion below, pp. 108-118.
gone through the process specified by Sutherland. Perhaps it is

assumed that some types of criminal behaviour are "obviously
exceptional." However, the theoretical analysis presented in
Chapter VI indicates that one type of behaviour that appears to
be obviously exceptional-"compulsive criminality"-is not ne-
cessarily exceptional at all. 27
A second principal kind of criticism attacks the theory because
it does not adequately take into account the "personality traits,"
"personality factors," or "psychological variables" in criminal
behaviour. This is real criticism, for it suggests that Sutherland's
statement neglects an important determinant of criminality.
Occasionally, the criticism is linked with the apparent assumption
that some kinds of criminality are "obviously" exceptional.
However, at least a dozen authors have proposed that Suther-
land's statement is defective because it omits or overlooks the
general role of personality traits in determining criminality.28
Sutherland took this kind of criticism seriously, and in an early
period he stated that his theory probably would have to be re-
vised to take account of personality traits.29 Later he pointed out
what he believed to be the fundamental weakness in his critics'
argument: "Personality traits," and "personality" are words that
merely specify a condition, like feeblemindedness, without show-
ing the relationship between that condition and criminality. He
posed three questions for advocates of "personality traits" as
supplements to differential association: (1) What are the person-
ality traits that should be regarded as significant? (2) Are there
personal traits, to be used as supplements to differential associ-
27 See Donald R. Cressey, "The Differential Association Theory and Compulsive
Crimes, "Journal of Crimina I Law and Criminology, 45 (May-June, 1954), pp. 49-64.
28 Barnes and Teeters, op. cit., p. 159; Milton L. Barron, The Juvenile in Delinquent
Society, New York: Knopf, 1954, p. 147; Clinard, "Criminological Theories of Viola-
tions of Wartime Regulations", op. cit., "Sociologists and American Criminology",
Journal of Criminal Law and Criminology, 41 (January-February, 1951), pp. 549-577;
"The Sociology of Delinquency and Crime", op. cit.; Sociology of Deviant Behaviour,
op. cit., pp. 204-205, 229, 240-241 ; Gill, op. cit., p. 286; Glueck, op. cit., p. 97; Kinberg,
op. cit.; Robert E. Lane, "'Vhy Businessmen Violate the Law", Journal of Criminal Law
and Criminology, 44 (July-August, 1953), pp. 151-165; Leader, op. cit.; S. F. Lottier
"Tension Theory of Criminal Behaviour," American Sociological Review, 7 (December,
1942), pp. 840-848; Martin H. Neumeyer, Juvenile Delinquen(y in Modem Society,
Second Edition, New York: Van Nostrand, 1955, pp. 152-153; Short, "Differential
Association as A Hypothesis", op. cit., p. 17; Void, ap. cit., p. 197.
2' Edwin H. Sutherland, "Development of the Theory", (1942) in Albert K. Cohen,
Alfred R. Lindcsmith, and Karl E. Schuessler, Editors, The Sutherland Papers, Blooming-,
ton: Indiana Cniversity Press, 1956, pp. 25-27.
ation, which are not already included in the concept of differ-

ential association? (3) Can differential association, which is
essentially a process oflearning, be combined with personal traits,
which are essentially the product of learning?30
Sutherland did not attempt to answer these questions, but the
context of his discussion indicates his belief that differential
association does explain why some persons with a trait like
"aggressiveness" commit crimes, while other persons possessing
the same trait do not. It also reveals his conviction that terms like
"personality traits," "personality," and "psychogenic trait com-
ponents" are (when used, with no further elaboration, to explain
why a person becomes a criminal) synonyms for "unknown
conditions. "
Closely allied with the "personality trait" criticism is the as-
sertion that Sutherland's statement does not adequately take into
account the "response" patterns, "acceptance" patterns, and
"receptivity" patterns of various individuals. 31 The essential
notion here is that differential association emphasizes the social
process of transmission but minimizes the individual process of
reception. Stated in another way, the idea is that the theory of
differential association deals only with external variables and
does not take into account the meaning to the recipient of the
various patterns of behaviour presented to him in situations which
are objectively quite similar but nevertheless variable, according
to the recipient's perception of them. One variety of this type of
criticism takes the form of asserting that criminals and non-
criminals are sometimes reared in the "same environment"-
criminal behaviour patterns are presented to two persons, but
only one of them becomes a criminal.
Sutherland was acutely aware of the social psychological
30 Edwin H. Sutherland, White Collar Crime, New York: Dryden, 1949, p. 272.
31 John C. Ball, "Delinquent and Non-Delinquent Attitudes Toward the Prevalence
of Stealing", Journal of Criminal Law and Criminology, 48 (September-October, 1957),
pp. 259-274; Caldwell, op. cit., p. 182; Clinard, "The Process of Urbanization and
Criminal Behaviour", op. cit., Sociology of Deviant Behaviour, op. cit., and "Criminological
Research", op. cit.; Glueck, op. cit.;J effery, op. cit.; Korn and McCorkle, op. cit., p. 298;
Leader, op. cit.; Neumeyer, op. cit., p. 152. Walter C. Reckless, The Crime Problem,
Second Edition, New York: Appleton-Century-Crofts, 1955, p. 109; and The Etiology
ofDelinquent and Criminal Behaviour, New York: Social Science Research Council, 1943,
p. 62; Harrison M. Trice, "Sociological Factors in Association with A.A.", Journal of
Criminal Law and Criminology, 48 (November-December, 1957), pp. 374-386; Void,
op. cit., p. 196; Weinberg, op. cit.
problem posed by such concepts as "differential response pat-

terns." Significantly, his proposed solution to the problem was
his statement of the theory of differential association. 32 One of the
principal objectives of the theory is to account for differences in
individual responses to opportunities for crime and in individual
responses to criminal behaviour patterns presented. To illustrate,
one person who walks by an unguarded and open cash register, or
who is informed of the presence of such a condition in a nearby
store, may perceive the situation as a "crime committing" one,
while another person in the identical circumstances may perceive
the situation as one in which the owner should be warned against
carelessness. The difference in these two perceptions, Sutherland
held, is due to differences in the prior associations with the two
types of definition of situation, so that the alternatives in be-
haviour are accounted for in terms of differential association. The
differential in "response pattern," or the difference in "receptivi-
ty" to the criminal behaviour pattern presented then, is accounted
for by differential association itself.33 In Chapter VII we shall
insist that one of the greatest defects in Sutherland's theory is its
implication that receptivity to any behaviour pattern presented is
determined by the patterns presented earlier, that receptivity to
those early presentations was determined by even earlier presen-
tations, and so on back to birth. 34 But this is an assertation that
the theory cannot be tested, not an assertion that it does not take
into account the "differential response patterns" of individuals.
If "receptivity" is viewed in a different way, however, the
critics appear to be on firmer ground. 35 Sutherland did not identi-
32 See Edwin H. Sutherland, "Susceptibility and Differential Association", in

Cohen, Lindesmith, and Schuessler, op. cit., pp. 42-43. See also Solomon Kobrin,
"The Conflict of Values in Delinquency Areas", American Sociological Review, 16
(October, 1951), pp. 653-661.
33 Cf. Ralph L. Beals, "Acculturation", in A. L. Kroeber, Editor, Anthropology
Today, Chicago: University of Chicago Press, 1953, pp. 641; and Richard Thurnwald,
"The Psychology of Acculturation", Amorican Anthropologist, 34 (October-December,
1932), pp. 557-569 .
.. See Cressey, "Application and Verification of the Differential Association
Theory", op. cit.
35 I am indebted to Albert K. Cohen for assistance with this paragraph and with
other points. Also, I am grateful to the following persons for suggested modifications
of the original draft of this chapter: Daniel Glaser, Sheldon Glueck, Michael Hakeem,
Frank Hartung, C. Ray Jefff'ry. Richard T. Morris. Melvin Seeman, Jamf's F. Short,
Jr., and George B. VoId.
fy what constitutes a definition "favourable to" or "unfavourable

to" the violation oflaw, but he recognized that the same objective
definition might be "favourable" or "unfavourable," depending
on the relationship between the donor and the recipient. Conse-
quently, he said that differential association may vary in "intensi-
ty," which was not precisely defined but "has to do with such
things as the prestige of the source of a criminal or anti-criminal
pattern and with emotional reactions to the associations." This
attempt at what is now called "reference group theory" merely
begs the question; it tells us that some associations are to be given
added weight, but it does not tell us how, or whether, early associ-
ations affect the meaning oflater associations. If earlier associations
determine whether a person will later identify specific behaviour
patterns as "favourable" or "unfavourable" to law violation, then
these earlier associations determine the very meaning of the later
ones, and do not merely give added weight to them. In other
words, whether a person is prestigeful or not prestigeful to an-
other may be determined by experiences that have nothing to do
with criminality and anti-criminality. Nevertheless, these experi-
ences affect the meaning (whether "favourable" or "unfavour-
able") of patterns later presented to the person and, thus, they
affect his "receptivity" to the behaviour patterns. 36
A fourth kind of criticism is more damaging than the first
three, for it insists that the ratio of learned behaviour patterns
used by Sutherland to explain criminality cannot be determined
with accuracy in specific cases. A minimum of eight authors have
stated this criticism in ~even different articles. 37 Short, for exam-
ple, has pointed out that the extreme difficulty of operationalizing
terms such as "favourable to" and "unfavourable to"; neverthe-
less, he has devised various measures of differential associations
and has used them in a series of significant studies. Glaser has
argued that "the phrase 'excess of definitions' itself lacks clear
denotation in human experiences," and Glueck has asked, "Has
36 This actually is the important point VoId was making in the quotation cited at
footnote 16 in Chapter I, above.
37 Ball, op. cit.; Clinard, "Criminological Research", op. cit.; Cressey, "Application
and Verification of the Differential Association Theory", op. cit.; Glaser, "Criminality
Theories and Behavioural Images, op. cit. ; Glueck, op. cit. p. 96; Lane, op. cit. ; Reckless,
The Etiology of Delinquent and Criminal Behaviour, op. cit., p. 63; Schrag, op. cit.; Short,
"Differential Association and Delinquency", op. cit.; "Differential Association as a
Hypothesis", op. cit.
anybody actually counted the number of definitions favourable to

violation oflaw and definitions unfavourable to violation oflaw,
and demonstrated that in the pre-delinquency experience of the
vast majority of delinquents and criminals, the former exceeds the
latter?" In my work on trust violation, I was unable with the
methods at my disposal to get embezzlers to identify specific
persons or agencies from whom they learned behaviour patterns
favourable to trust violation. My general conclusion was, "It is
doubtful that it can be shown empirically that the differential
association theory applies or does not apply to crimes of financial
trust violation or even to other kinds of criminal behaviour."38
I have been severly taken to task for not revising Sutherland's
statement in light of this conclusion. 39 My reasons for not doing
so have to do with the difference in the theory of differential
association considered as a general principle which organizes and
makes good sense of the data on crime and delinquency rates, as
compared to the theory considered only as a statement of the
precise mechanism by which a person becomes a criminal or a
delinquent. As we indicated in Chapter IV, a principle accounting
for the distribution of deviancy, or any other phenomenon, can
be valid even if a presumably coordinate theory specifying the
process by which deviancy occurs in individual cases is incorrect,
let alone untestable.
The fifth kind of criticism states in more general terms than the
first four that the theory of differential association over-simplifies
the process by which criminal behaviour is learned. Such criti-
cism ranges from simple assertions that the learning process is
more complex than the theory states or implies,40 to the idea that
the theory does not adequately take into account some specific
type of learning process, such as differential indentification. 41
Between these two extremes are assertions that the theory is
inadequate because it does not allow for a process in which crimi-
nality seems to be "independently invented" by the actor. I am
one of the dozen authors who have advanced this kind of criti-
38 See below, p. 118.

3. Caldwell, op. cit., p. 185.
40 See, for example, Ball, op. cit .
., See, for example, Clinard, "The Process of Urbanization and Criminal Be-
haviour", op. cit.; and Glaser, "Criminality Theories and Behavioural Images,op.
cit.
cism,42 and in this day ofrole theory, reference group theory, and
complex learning theory, it would be foolhardy to assert that this
type of general criticism is incorrect. But it is one thing to criticise
the theory for failure to specify the learning process accurately
and another to specify which aspects of the learning process
should be included and in what way.43 Clinard's and Glaser's
attempts to utilize the process of identification, and Weinberg's,
Sykes and Matza's, and my own efforts to utilize more general
symbolic interactionist theory, seem to be the only published
attempts that specifically substitute alternative learning processes
for the mechanistic process specified by Sutherland. Even these
attempts are, like Sutherland's statement, more in the nature of
general indications of the kind of framework or orientation one
should use in formulating a theory of criminality than they are
statements of theory.
4' Caldwell, op. cit., p. 183; Clinard, "The Sociology of Delinquency and Crime",
op. cit.; and "Criminological Research", op. cit.; Cressey, "Application and Verification
of the Differential Association Theory, op. cit.; and "The Differential Association
Theory and Compulsive Crime", op. cit.; Daniel Glaser, "Review of Principles of
Criminology", Federal Probation, 20 (December, 1956), pp. 66-67; "The Sociological
Approach to Crime and Correction", Law and Contemporary Problems, 23 (Autumn,
1958), pp. 683-702; and "Differential Association and Criminological Prediction:
Problems of Measurement", Social Problems, 8 (Summer, 1960), pp. 6-14; Glueck,
op. cit., pp. 93, 97; Korn and McCorkle, op. cit., p. 299; Leader, op. cit.; Short, "Differ-
ential Association as a Hypothesis", op. cit.; Gresham Sykes and David Matza,
"Techniques of Neutralization: A Theory of Delinquency", American Sociological
Review, 22 (December, 1957), pp. 664--670; Weinberg, op. cit.
43 Despite the fact that Sutherland described a learning process, it should be noted
that he protected himself by saying, "The process of learning criminal and anti-
criminal patterns involves all the mechanisms that are involved in any other lear-
ning".
CHAPTER VI
DIFFERENTIAL ASSOCIATIONS
AND COMPULSIVE CRIMES
Chapter I reported a number of interpretive errors that have

arisen because readers have not understood what Sutherland
apparently was trying to say, and Chapter V reported some of the
principal criticisms of Sutherland's theory. It is only necessary to
observe here that two of the important types of criticism are
related both to each other and to the question of whether a small
segment of all criminal behaviour, "compulsive crime," can be
handled by the theory of differential association.
One general kind of criticism holds that the theory over-
simplifies the process by which criminal behaviour is learned.
Taken in its best light, this criticism implies that Sutherland and
his followers have not taken full advantage of broader social
psychological research describing learning mechanisms. I t is true
of course, as we indicated earlier, that this may involve a mis-
guided faith in the more general work being done on learning in
sociology, social psychology, and psychology; it could easily be
argued that the theory of differential association is as precise as
possible in view of the present state of knowledge regarding social
learning. Nevertheless, most criminologists-especially those out-
side the United States and Scandinavia-have little training in so-
cial science and therefore have not taken advantage ofsuch general
social psychological theory as exists. For example, it is probably
lack of familiarity with "symbolic interaction" theory! in social
psychology which makes "compulsive crime" seem exceptional to
Sutherland's principle. The theory ofdifferential association is con-
sistent with this broader set of theory, and one of our objectives in
this chapter is to show how that set of theory can be used in con-
siderations of behaviour such as "kleptomania "and "pyromania."
1 See Arnold Rose, Editor, Human Behaviour and Social Processes: An Interaetionisl
"Jj,proach, Boston: Houghton Hiffiin, 1961.
COMPULSIVE CRIMES
A second general kind of criticism indicates that some kinds of

criminal behaviour seem expectional to the theory, and "indivi-
dual," "personal," "irrational," "impulsive" and "compulsive"
behaviour are among the exceptions noted. In one sense, this is a
variation of the first kind of criticism, for the implication is that a
learning process quite different from the one Sutherland had in
mind, and thus from "symbolic interaction," is involved.
Use of concepts such as "crimes of passion" and "compulsive
crime" are likely to indicate a belief that some kinds of criminal
conduct are exceptional to the kind of associational process
discussed by Sutherland-and, more generally, by the followers
of the interactionist theory set in motion by George Herbert
Mead 2-because an alternative causal process is in operation.
In another sense, however, this kind of criticism might indicate a
conviction that some types of criminal conduct are "obviously
exceptional" to the theory of differential association, even if it is
not now possible to specify a precise process by which these types
of criminal conduct come into being. Sutherland's generalization
is stated in universal form, as a description of the etiology of all
criminal behaviour. Hence, the discovery of cases of behaviour
which is crime and whose genesis and development has not
followed the process Sutherland described will call for either a
modification of the generalization or a redefinition ofthe concept
"crime." 3
As indicated, the important point in Sutherland's theory-as
it pertains to individual criminality-is the principle that all
criminal behaviour is learned in a process of social interaction.
It therefore is important that we carefully examine all behaviour
to which the label "crime" is applied, but which does not seem to
have been learned in such interaction. 4 "Compulsive crime" is
the best example of such behaviour.
2 George H. Mead, Mind, Self and Society, Chicago: University of Chicago Press.
1934.
3 J\fany criminologists argue that crime is not a homogenous phenomenon and
that it is therefore unwise to attempt universal scientific generalizations about it.
While there is merit in this argument, it is not pertinent to the present discussion.
See Donald R. Cressey, "Criminological Research and the Definition of Crimes",
American Journal of Sociology, 56 (May, 1951), pp. 546-551.
• Some apparent exceptions were pointed out by Sutherland himzelf. See his "The
Relation Between Personal Traits and Associational Patterns", in Walter C. Reckless,
The Etiology of Delinquent and Criminal Behaviour, New York: Social Science Research
Council, 1943, pp. 131-137. Similarly, crimes which in the criminal law are based on
The Legal-Psychiatric Controversy

Before such re-examination can be made, however, it is neces-
sary to review briefly the three fundamental points in the legal-
psychiatric controversy about whether behaviour said to be
"compulsive" also is "crime." The issues are fairly clear. First,
in the criminal law, under the M'Naghten rules, the stigma
"insanity," not "crime," is applied to legally harmful behaviour
perpetrated under circumstances such that the defendant was
unable to distinguish between right and wrong. This is to say, in
more sociological terms, that he was unable to contemplate the
normative consequences of his acts. If it is observed that a so-
called "compulsive criminal" did contemplate the normative
consequences of his behaviour, the behaviour is classed as crime,
rather than as insanity.
Second, psychiatrists insist that some of the behaviour which
results in legal harm ("compulsive crime") has essentially the
same characteristics as does compulsive behaviour generally. As a
general category of neuroses ("psychasthenia," "anankastic re-
ctions," etc.), compulsive acts are described as irresistible be-
haviour which the person in question often recognizes as ir-
rational but is subjectively compelled to carry out. 5 Such acts are
considered as irrational because they are thought to be prompted
by a subjective morbid impulsion which the person's "will" or
''judgment'' or "ego" cannot control.6 In other words, behaviour
described as compulsive is thought to be completely determined
strict liability rather than on proof of criminal intent almost immediately appear as
possible exceptions. One distinguished student of the criminal law holds, however,
that strict liability offences have so little in common with other crimes that they should
not be included in the ambit of the criminal law - Jerome Hall, Principles of Criminal
Law, Indianapolis: Bobbs-Merrill, 1947, pp. 280, 296. Thus, the apparent exception
seemingly calls for a revision of the concept "crime", rather than a revision of the
theory. Such a redefinition of the crime concept, incidentally, would have serious
effects on current generalizations about "white-collar crimes", many of which are of
a strict liability type.
5 Cj. Roy M. Dorcus and G. Wilson Shaffer, Textbook of Abnormal Psychology,
Baltimore: Williams and Wilkins, 1941, p. 364.
• Malamud, for example, indicates that psychasthenias or anankastic reactions
"have in common the fact that the patients feel themselves compelled by some inner
force and against their own will or reason to think, act or feel in an abnormal manner",
and that compulsive acts are "forms of behaviour which the person carries out
consciously without knowing the reason for such activity or for reasons which he
knows have no logical foundation". William Malamud, "The Psychoneuroses" in
J. McV. Hunt, Editor, Personality and the Behaviour Disorders, New York: Ronald Press,
1944, pp. 851-852. Cj. Franz Alexander and Hugo Staub, The Criminal, The Judge,
and The Public, New York: Macmillan, 1931, pp. 149-150.
COMPULSIVE CRIMES 93
by the inner impulse or compulsion, and while the genesis of the

compulsion might lie in social context, once it has been formed it
apparently operates as an entity, agent, or element in itself. Thus,
the overt act is considered as prompted entirely "from within,"
and present contact with values concerning morality, decency, or
correctness of the overt behaviour in no way affects the actor, in
the last analysis, either in deterring him from acting or in en-
couraging him to act. 7
It is argued by psychiatrists, then, that in cases of "compulsive
crime" the actor does know right from wrong and does contem-
plate the normative consequence of his acts (i.e., he recognizes
the behaviour as irrational, foolish, wrong, illegal, etc.) but never-
theless exhibits the behaviour because it is prompted from within
by a force which he is powerless to resist. Lorand, for example,
cites three case histories as evidence that "compulsive stealing" is
a subconscious act of aggression against the parents or parent
surrogates. He points out that there were faults in the critical
appreciation of the factors of reality, and that "all showed an
overwhelmingly strong instinctual drive which clouded the
function of critical faculty. They were unable, consciously, to
resist, and they could not prevent the breaking through of strong
drives from within which lead to stealing."g If the legal harm
resulting from such behaviour actually is crime, then it obviously
is exceptional to the theory of differential association. The
"criminality" in "compulsive crime" would depend not upon
former contracts with differential values concerning law-abiding-
ness, but upon a non-social agent or process.
Third, some jurists have adopted a position similar to that of
the psychiatrists. This apparent from the fact that the courts of
about fourteen of the American states hold that the consequences
to an actor of the perpetration of a legal harm can be avoided by
showing that while the defendant knew right from wrong his
behaviour was prompted by an "irresistible impulse."9 The
"irresistible impulse" and "compulsive crime" concepts seem to
have at their base the same assumptions inherent in the old
7 Cj. Gregory Zilboorg, "Misconceptions of Legal Insanity". American Journal !if

Orthopsychiatry, 9 (July, 1949), pp. 540-553.
8 Sandor Lorand, "Compulsive Stealing", Journal ofCriminal Psychopathy, I (January
1940), pp. 247-253. Cj. Malumud, op. cit.
e E. R. Keedy, "Irresistible Impulse as a Defense in Criminal Law", University of
Penn~vlvania Law Review. 100 (May, 1952) pp. 956-993.
faculty psychology concepts "moral perversion," "moral imbe-

cility," "inhibatory insanity," "affective mania," "monomania,"
etc., each of which implied a psychological disorder which has no
connection with the "intellect" or "knowing" or "reasoning"
faculties.lO Most modern psychiatrists claim that theirs is not a
faculty psychology since that which was formerly considered as
emotional and intellectual faculties is now considered as one-the
total personality. But while this "personality integration" theory
is affirmed by psychiatrists as they oppose the M'Naghten rules,
it is denied when they support notions of "compulsive crime" or
"irresistible impulse."ll While there is disagreement among
psychiatrists, it appears that most of them agree with the legal
theory of those jurisdictions allowing the irresistible impulse
defense, and many of them contend that those judges not allowing
it are backward, ignorant, or stubborn. Wertham is a noteworthy
exception. He writes that "the criminal law which makes use of
the conception of irresistible impulse is not an advance belonging
to the present 'scientific social' era. It is a throwback to, or rather
a survival of, the previous 'philosophical psychological' era. The
concept of irresistible impulse derives from a philosophical,
speculative, synthetic psychology. It forms no part of and finds no
support in the modern dynamic psychoanalytic study of mental
process."12 Elsewhere, the same author has stated that there is
nothing in the whole field of psychopathology which corresponds
to the irresistible impulse, and that compulsions play no role in
criminal acts.13 Bromberg and Cleckley have taken a similar
position: "The concept of sudden. 'irresistible impulses' in an
otherwise perfectly normal organism is unsupported by modern
psychiatric knowledge."14
A1entalistic Assumptions of Law and Psychiatry

This divergence in opinion and viewpoint is enhanced by the
10 See Lawson G. Lowrey, "Delinquent and Criminal Personalities", inJ. McV.
Hunt, Personality and the Behaviour Disorders, New York: Ronald Press, 1944, pp. 799-
801.
11 See the discussion by Jerome Hall, op. cit., pp. 523-524.
12 New York University School of Law, Social lvleaning of Legal Concepts - Criminal
Guilt, New York: New York University School of Law, 1950, p. 164.
IS Fredric Wertham, The Show of Violence, New York: Doubleday, 1949, pp. 13-14.
14 Walter Bromberg and Hervey M. Cleckley, The Medico-Legal Dilemma - A
Suggested Solution", Journal of Criminal Law and Criminology, 42 (March-April, 1952),
pp. 729-745.
fact that an assumption of "mind" is implicit in the psychological

orientations of both psychiatry and criminal law, so that each
discipline has a "mentalistic" approach to human behaviour. In
criminal law, the "right and wrong test" assumes the existence of
a mind which, when normal and mature, operates in such a way
that the human has conscious freedom to choose rationally
whether or not a crime shall be committed. The mind impels the
person only in the direction "he" wishes to be impelled. But a
mind which is immature or "diseased" cannot make intelligent
choices, and a defendant possessing such a mind is considered in-
capable of entertaining criminal intent. Such an assumption
tends to equate rationality and sanity, and it is necessary to fixing
"responsibility" for acts.I5 Although psychiatrists often denounce
this jurisprudential assumption on the ground that it ignores the
facts of science,16 their denunciation is possible only because of
emphasis on a different mentalistic construct. In writing of "com-
pulsive crime," at least, most psychiatrists assume a mind with
only one significant difference from the one assumed by most
jurists. Here, the mind is said to be subject to casual emotional
experiences, especially early sexual experiences, which give it
characteristics such that at the present moment of action it com-
pletely determines the person's choice-and, consequently, his
overt behaviour-in a manner which usually is completely un-
conscious and unknowable without the help of a psychiatrist. The
deeply hidden emotional forces of the mind are thought to compel
the actor even if "he" knows the action is illegal, and "he" has no
choice in whether the action shall be undertaken. The chief
difference between such psychological forces or "mainsprings of
action," and instinctive "mainsprings of action" is that the former
are "unconscious."17 Foote has pointed out that in spite of the
seductive appeal which is exerted by the hope of reducing human
15 For discussion of the differences between "responsibility", "causation", and

"accountability", see Arnold W. Green, "The Concept of Responsibility", Journal qf
Criminal Law and Criminology, 33 (January-February, 1943), pp. 392-394, and Brom-
berg and Cleckley, op. cit.
16 See, e.g., Edward Glover, The Diagnosis and Treatment of Delinquency", in
L. Radzinowicz and]. W. C. Turner, Editors, Mental Abnormality and Crime, Volume II
of English Studies in Criminal Science, London: Macmillan, 1949, pp. 279-280.
17 See Benjamin Karpman, "An Attempt At a Re-evaluation of Some Concepts
of Law and Psychiatry", Journal Criminal Law and Criminology, 38 (September-October,
1947), pp. 206-217.
behaviour to some simple and permanent order through finding

certain "basic" imperatives underlying it, criticism has negated
every specific naming of the "mainsprings of action."18
Mentalistic assumptions of both kinds must be clarified and
supplemented in order to determine whether "compulsive crimi-
nality" is exceptional to the theory of differential association. As
long as criminality is said to have its etiology in a rather mysteri-
ous "mind," "soul," "will," or "unconscious," there will be no
possible way for generalizations about criminality to be subjected
to empirical tests or observations which would settle the issue.
Also, so long as "compulsiveness," as traditionally described,
must be determined by specialists rather than judges or juries,
jurists will resist discussion of it in their courts, and the legal
psychiatric controversy will continue.
A Sociological Theory of Motivation

Behaviour traditionally considered as "compulsive crime" can
be handled and clarified without the assumption of "mind" or a
basic biological or psychic imperative by application of the
sociological hypothesis that there are differences in the degree to
which acts are controlled by the linguistic constructs (words or
combinations of words) which the actor has learned from his
social groups. Since the use oflinguistic constructs depends upon
contacts with social groups, this amounts to differences in the
degree to which the actor participates in group experiences. In
sociological "role theory," differences of this kind are considered
differences in the motivation of the actor, although this concept is
used in a sense quite different from the use in psychiatry. Moti-
vation here refers to the process by which a person, as a partici-
pant in a group, symbolically (by means of language) defines a
problematic situation as calling for the performing of a particular
act, with symbolically anticipated consumation and conse-
quences. 19 Motives are not inner, biological mainsprings of action
but linquistic constructs which organize acts in particular situ-
ations,20 the use of which can be examined empirically. The key
18 Nelson N. Foote, "Identification as a Basis for a Theory of Motivation", American
Sociological Review, 16 (February, 1951), pp. 14-22.
19 Ibid.
20 Cf. C. Wright Mills, "Situated Actions and Vocabularies of Motive", American
Sociological Review, 5 (December, 1940) pp. 904-913.
linguistic constructs which a person applies to his own conduct in

a certain set of circumstances are motives; the complete process
by which such verbalizations are used is motivation.
The great difference between this conception of motivation and
the notion that motives are biological or are deeply hidden in the
"unconscious" may be observed in the use of the concept "ration-
alization" in the two systems. In psychiatry it usually is said that
one "merely rationalizes" (ex post facto justification) behaviour
which "has really been prompted by deeply hidden motives and
unconscious tendencies."21 In the other system, which uses a non-
mentalistic conception of motivation, it is held that one does not
necessarily "merely rationalize" behaviour already enacted but
acts because he has rationalized. The rationalization is his motive.
When such rationalizations or verbalizations are extensively
developed and systematized the person using them has a sense of
conforming because they give him asense ofsupport and sanction. 22
An individual in our society, for example, may feel fairly comfort-
able when he commits an illegal act in connection with his
business for, after all, "business is business." But not all verbali-
zations are equally developed or systematized, and in some
instances the use of the verbalizations does not, therefore, receive
such extensive support and sanction. The individual in these
instances does not have a comfortable sense of conforming. The
person in the above example probably would not feel as comfort-
able if his illegal act were perpetrated according to the verbali-
zation "all businessmen are dishonest." Motives can be treated,
then, as "typical vocabularies (linguistic constructs) having
ascertainable functions in delimited societal situations"23 and, as
such, they may be examined empirically.
U sing this conception of motivation, it is immediately apparent
that not all behaviour is equally motivated; there are differences
in the degree to which behaviour is linguistically controlled.
Certainly some behaviour is performed with almost no social
referent, i.e., with the use of no shared verbalization. For instance,
21 Arthur P. Noyes, Text-book of Modem Psychiatry, Philadelphia: W. B. Saunders,

1940, p. 49. See also, for example, William A. White, Insanity and the Criminal Law,
New York: Macmillan, 1923, p. 9; and Benjamin Karpman, op. cit.
22 For an excellent discussion of this point, see A. R. Lindesmith and A. L. Strauss,
Sodal Psychology, New York: Dryden Press, 1949, pp. 307-310.
28 C. Wright Mills, op. cit.
behaviour which is physiologically autonomous is clearly non-

motivated since the release of energy appropriate to performing
the behaviour does not depend upon the application of a linguistic
construct. Similarly, if one's behaviour has been so conditioned
by his past experiences that he behaves automatically, in the way
that Pavlov's dogs behaved automatically at the sound of the bell,
he is not motivated. Genuinely fetishistic behaviour probably is
of this kind. However, it is equally certain that other behaviour
cannot be enacted unless the actor has had rather elaborate and
intimate contact with linguistic constructs which are, by defi-
nition, group products. Such behaviour is motivated, and it may
be distinguished from automatic behaviour by the fact that it has
reference to means and ends. If a person defines a situation as one
in which there are alternatives, if there is evidence of planning,
evidence of delaying small immediate gains for larger future gains
or evidence of anticipation of social consequences of acts, he is
motivated.
Application of the Motivation Theory to "Compulsive Crime"

When this theory of motivation is applied to the problem at
hand, it may be seen that if behaviour traditionally considered as
"compulsive crime" were clearly non-motivated or autonomous
then the legal-psychiatric controversy would have been resolved
long ago since, if such were the case, the behaviour easily could
be subsumed under the legal concept "insanity." If "compulsions"
"in" a person "came out" in the same way that his whiskers
"come out," then even in the most "anti-psychiatric" court there
would be no question of his legal responsibility, and his behaviour
would not, in fact, be designated as crime. In this case, there
would be no problem about whether the behaviour were ex-
ceptional to the differential association theory but, instead, the
behaviour would lie outside the definition of the phenomenon
(crime) with which the differential association theory is con-
cerned. Non-motivated behaviour of this kind which resulted in
legal harm would not be unlike the behaviour of a sleeping or
drugged person whose hand was guided by another to the trigger
of a gun aimed at a victim's head. Such behaviour is not planned
by the actor, and precautions against detection are not taken
because the ability to use the language symbols normally perti-
nent to the situation is absent or deficient. 24 If behaviour is non-

motivated the actor cannot possibly entertain "crimal intent."
In most cases now labelled "kleptomania," "pyromania," etc.,
however, the actors appear to be motivated in the same way that
other criminals are motivated. Consequently they are, in the
terminology of the criminal law, "responsible." They select
secluded places in which to perpetrate their acts, plan their
activities in advance, realize that they will be arrested if detected,
and do many other things indicative that there is a conscious
normative referent in their behaviour. Certainly most acts tra-
ditionally described as "compulsive crime" are clearly quite
different from autonomous behaviour having no normative com-
ponents, in spite of the fact that the two are usually assumed to be
identical or at least very similar. Possible it was observation of
the significant difference between the two that led Alexander and
Staub to argue that while the impulse in kleptomania, pyromania
and compulsive lying is an unconscious one, an impulse foreign to
the ego, yet the act is not completely unconscious, as is the case,
they hold, in compulsive neuroses generally.25 Gault uses klepto-
mania and pyromania as illustrations of the psychopathic person-
ality. That is they are considered as-
A form of outlet for a nature that is unbalanced by reason of the dominance

of the egocentric disposition ... They take what does not belong to them not so
much as a result of blind impulsion; not quite "blind impulsion" because the
kleptomaniac is at pains to conceal not only his act but the products of his
stealing ... We are forced to conclude that in the general run of instances of this
nature we are dealing with unreasoned impulsion to get goods to amplify one's
store, to gratify one's desire for possession and therefore magnify one's self. In
other words, here is the egocentric disposition. This language suggests a
purposive character of the behaviour - and it is so that it results in obtaining
goods. 26
24 Although it is not recommended, because it probably would lead to even more

confusion than now exists, the criminal law theory which exempts some persons from
liability because they lack "responsibility" might easily be restated in these terms.
Those persons - generally psychotics and very young children - who are now excused
on the ground that they cannot distinguish between right and wrong either have not
acquired or have lost the ability to control language symbols. In fact, on a level which
now seems unsophisticated, this principle was recognized in the criminal law as early
as the thirteenth century when Bracton formulated what erroneously has been called
the "wild beast test".
25 Franz Alexander and Hugo Staub, op. cit., pp. 95-97.
26 Robert H. Gault, Criminology, New York: D. C. Heath and Company, 1932,
pp. 163-166.
However, many other authorities claim that the evidence of

deliberation and intent in acts of pyromaniacs and kleptomaniacs
does not in and of itself signify sanity.27 Thus, while it is not easy
to classify precisely certain acts as motivated or non-motivated,
since men do not always explicitly articulate motives,28 the soci-
ological framework at least affords an opportunity to classify
correctly the great proportion of the acts ordinarily labelled
"compulsive crime." Using that framework, illegal conduct
which is motivated would be classed as "crime" and illegal acts
which were non-motivated would be classed as "compulsion" and
would fall within the legal category "insanity."
Non-Scientific Criteria of Compulsive Criminals

Accurate classification of this kind would be valuable, since in
the current system it appears that compulsive crime concepts are
no less "wastebasket categories" than is the "psychopathic person-
ality" concept. Casual observation indicates, at least, that the
application of the "compulsive crime" label often accompanies
the inability of either the subject or the examiner to account for
the behaviour in question in terms of motives which are current,popular,
and sanctioned in a particular culture or among the members of a particular
group within a culture. For example, one criterion, usually over-
looked, for designating behaviour "kleptomania" rather than
"theft" is apparent lack of economic need for the item on the part
of the person exhibiting the behaviour. This may be observed in
at least two different ways. First, the probability that the term
"kleptomania" will be applied to a destitute shoplifter is much
lower than the probability that it will be applied to a wealthy
person performing the same kind of acts. "Kleptomania," then,
often is simply a short-hand way of saying, as the layman does,
"That woman is rich and can buy almost anything she desires.
She does not need (economic) to steal. She must be crazy."29 An
27 Perry M. Lichtenstein, A Doctor Studies Crime, New York: Van Nostrand, 1934,
p. 182; Wilhelm Stekel, Peculiarities of Behaviour, New York: Liveright, 1924, Vol. I,
p.258.
28 C. Wright Mills, op. cit.
2D Compare: "In kleptomaniacs we have individuals who steal, but their stealing
has a number of important differences from ordinary theft. For one thing, the purely
predatory element present in common theft is lacking here. The subject steals not
because of the value and the money he gets from the stolen articles - that is, not for
their mercenary value - but entirely for what they mean to him emotionally and
symbolically. One often observes this in rich women who have no need for the article
interesting but erroneous assumption in such logic is that the

behaviour of normal persons committing property crimes is ex-
plainable in terms of economic need. 30 This assumption, coupled
with the empirical observation that wealthy persons sometimes
do commit major property crimes led to the erroneous conclusion
by two sociologists that such crimes must be prompted by "greed"
rather than "need."31 Among psychiatrists, contradiction of the
same assumption, through observation of the fact that wealthy
persons do sometimes commit minor property crimes, results in
the notion that larcenous behaviour of wealthy persons must be
"compulsive." The economic status of the observer probably is of
great importance in determining whether he thinks a person is
not in economic need and is consequently compulsive. Whether
or not the misconduct is considered "disproportionate to any
discernible end in view" 32 conceivably will depend a great deal
upon the attitudes of the examiner rather than upon those of the
offender. That is, a poor person might consider that a middle-
class person had no need to steal and that his stealing must be the
result of "greed" or a "compulsion," while a middle-class person
probably will entertain this notion only as it refers to upper-class
persons whose incomes far exceed his. If all psychiatrists were
poverty-stricken the proportion of shoplifters called "klepto-
maniacs" probably would be much higher than it is. And if it is
assumed that the larcenous behaviour of wealthy persons is, be-
cause they are wealthy, "compulsive" then there is little oppor-
tunity for determination of possible contacts with behaviour
patterns conducive to crime.
Second, the absence, from the observer's standpoint, of
economic need is used as a criterion for designating persons as
"kleptomaniacs" in cases in which the particular articles taken
they steal and, in point of actual fact, dispose of almost immediately after the article
has been stolen. While the symbolic nature of such stealing is often evident on the
surface, we not infrequently come across cases of stealing the nature of which is not
so obvious, so that one is puzzled to figure out whether we are dealing with kleptomania
or ordinary theft. Many such cases are found in our prisons". Benjamin Karpman,
"Criminality, Insanity and the Law", Journal of Criminal Law and Criminology, 39
(January-February, 1949), pp. 584, 605.
30 For evidence to the contrary, see E. H. Sutherland, White Collar Crime, New
York: Dryden Press, 1949.
81 Harry Elmer Barnes and Negley K. Teeters, New Horizons in Criminology, New
York: Prentice-Hall, 1945, p. 43. The conclusion does not appear in the Second
Edition (1951) of this book.
3. William Healy, The Individual Delinquent, New York: Little, Brown ,1918, p. 771.
appear to be of no immediate use to the subject. For example,

Alexander and Staub do not consider as kleptomanic the be-
haviour of a physician (a "neurotic criminal") who had been
taking medical books and supplies, but his "theft of porcelain
figures which were new and actually of no value is more in the
nature of a kleptomanic act," and Wallerstein has stated that a
case "was hardly kleptomania in the usual sense because the
articles were pawned or sold for money."33 Although they have
not been explored in this connection, Veblen's arguments about
the great desirability in our culture of acquiring money merely as
a means of accumulating economically "useless" goods might
have a bearing here. 34 If this thesis were followed it would seem,
at least, that there is little logical justification for designating as
"criminal" the behaviour of one who stole money with which to
buy "useless" goods while at the same time designating as "com-
pulsive" the stealing of the goods themselves.
This criterion also is used for designating even poor persons as
"kleptomaniacs" in instances of repetitive taking of what appear
to be economically useless goods. However, the fact that mere
repetitive taking need not indicate a compulsion may be illus-
trated by the case of a gang of boys who went from store to store
in a large city stealing caps. They would enter a store and each
boy would steal a cap, leaving his own on the counter. The group
would then move to another store where the stolen caps would be
left on the counter, and new ones would be stolen. This practice
would continue until the gang members became bored with the
game. 35 It is not difficult to distinguish such behaviour from what
is called kleptomania because it was perpetrated by a number of
boys acting together, because it appears to have been done as
play,36 and, most important, because it was perpetrated by boys
closely approaching the cultural stereotype of the delinquent or
33 Franz Alexander and Hugo Staub, op. cit., p. 168; James S. Wallerstein, "Roots
of Delinquency". Nervous Child, 6 (October, 1947), pp. 399-412.
34 See Thorsten Veblen: The Theory 0/ the Leisure Class, New York; Vanguard Press,
1938; and Louis Schneider, The Freudian Psychology and Veblen's Theory, New York:
King's Crown Press, 1948.
35 Clifford R. Shaw, 'Juvenile Delinquency - A Group Tradition", Bulletin o/the
State University 0/ Iowa, No. 23, N.S. No. 700, 1933, p. 5.
36 In discussing legal problems of kleptomania Halls says, "But mere repetition
tends to prove habit, not abnormality." Jerome Hall, op. cit., p. 517. See also H. M.
Tiebout and M. E. Kirkpatrick, "Psychiatric Factors in Stealing", American Journal
o/Orthopsychiatry,2 (April, 1932), pp. 114-123.
criminal. Even a psychoanalyst probably would not assume that

the caps were sex symbols or fetishes, although it is not incon-
ceivable that he would do SO.37 However, if one of the boys at a
later date repeated alone the same kind of thefts, the probability
that he would be labelled a kleptomaniac would be high. And if
the boy were a member of a wealthy family the probability of his
being labelled a kleptomaniac would be even higher.
In fire-setting cases, the absence of obvious economic need also
is used, in the traditional system of thought, as a criterion for
applying the term "pyromania," but here the absence of other
popular motives is used as well. If it can be determined easily
that one burned property in order to collect insurance, in order to
get revenge, or in an attempt to conceal a criminal act, or if there
is ground for believing that he had some other conventional
motive, then the probability that he will be designated a "pyro-
maniac" is low. However, if none of these is immediately ap-
parent, and especially in instances where the thing burned has no
great economic value, the probability that the term will be
applied is much higher. One investigator states that of all the
varieties of incendiarists, the pyromaniac is most difficult to de-
tect "because of the lack of motive." 38 Traditionally, pyromania
has been, then, like kleptomania, a residual category.39
3. Karpman reports the case of a man who burglarized women's apartments taking
both money and female intimate garments. By using this case as an illustration of
"fetishistic kleptomania" he puts great emphasis upon the taking of the garments and
almost ignores the taking of money. Benjamin Karpman, "Criminality, Insanity and
the Law", op. cit., Gault has made the following significant statement about such
practices: "The attempt that some have made to lay the foundation for cases if this nature in
repressed sex motives and to interpret the objects stolen as so many symbols that they have in relation
to the sex aspect of experience is a very unconvincing procedure ... The case is doubtless not so
simple. The sex urge is only one of the many that actuate the human organism. For
instance, let us assume that a clothespin or a rubber hose found among the stolen
goods ofa kleptomaniac is a symbol. Symbol of what? The answer, according to the
present writer's opinion, is that its symbolic character depends upon what the investigator is
interested infinding." Robert H. Gault, op. cit., pp. 163-166. Italics not in the original.
38 Camille F. Hoyek, "Criminal Incendiarism," Journal of Criminal Law and Cri-
minology, 42 (March-April, 1951), pp. 836-845.
39 Psychoanalysts make much of the assumed sexual symbolism in cases of non-
economically motivated incendiarism. Thus, in one case of repeated burning of grass
on vacant lots it was asserted that the lots symbolized the subject's father, that by
driving onto the lot with his father's automobile the subject identified himself with
his father and his sex organ and performed the act of incest on his mother, that the
subject's efforts to help in extinguishing the fires was symbolic of an unconscious wish
to atone for his sin, and that the splashing of water on the fire was a symbolic repetition
of a regression to the urethral phase of his libidinal development, in which phase the
subject was said to have had erections which were relieved by urinating. Ernst Simmel,
In contrast to what appears to be current practice, when soci-

ological theory of motivation is used the apparent inability of a
person to explain his actions to the police, to a psychiatrist, or
even to himself, is not considered sufficient for classifying those
actions "compulsive." Using that theory, it may be observed that
most criminals, in fact, when asked to explain their acts either
recite the popular motives involved or respond that they do not
know. For example, one might say, as did a criminal who had
stolen a whole truck-load of groceries, "I didn't want to take them
but I had to because I was hungry." This response may be com-
pared with that of a person arrested for taking small objects from
a store: "I didn't want to take them but I just had to take them,"
and with that of a person who had burned an automobile, "I just
wanted to stir up some excitement." As indicated above, such
rationalizations are not necessarily ex post facto justifications for
acts-and if they are not, then there is no logical justification for
classifying one person as a "thief" and the others as "compulsive."
Motives are circumscribed by the actor's learned vocabulary.
Sociological Role Theory Applied to "Compulsive Crime"

But sociological theory can do more than correctly classify the
large proportion of defendants said to be compulsive. The litera-
ture on role theory provides a framework not only for under-
standing the behaviour of such defendants, but for understanding
their inability to account for their behaviour as well. Closely re-
lated to the theory of motivation which has been outlined is that
aspect of role theory which deals with the relationship between
the person's identification of himself as a "social object" and his
subsequent behaviour. 40 In order to playa social role, one must
anticipate the reactions of others by taking the role implicitly be-
fore it is taken overtly. He must look at himself from another's
point of view. By hypothesizing the reactions of others, the person
looks upon himself as an object and, consequently, identifies him-
self as a particular kind of object. He then performs the role which
is appropriate to the kind of social object with which he has
identified. The vocabulary of motives employed in the per-
"Incendiarism", in K. R. Eissler, Editor, Searchlights on Delinquency, New York:

International Universities Press, 1941, pp. 90-101.
40 For a general discussion of this subject, see Mead, op. cit.
formance of the role also is a corollary of this self identification.

But at various times and in different situations the person may
identify himself differently, so that he is able to play many, often
even conflicting, social roles. Again, his identification of who he
is determines the roles he plays.41
For example, one might in the course of a day identify himself
as a father in one situation, as a husband in another situation, and
as a property owner in another situation. The motives employed in
the performance of each role will reflect his particular identifi-
cation. A similar phenomenon may exist in respect to so-called
"compulsive crime." For example, a person might in some situ-
ations identify himself as a kleptomaniac, since that construct is
now popular in our culture, and a full commitment to such an
identification includes the use of motives which, in turn, release
the energy to perform a so-called compulsive act. The more
positive the conviction that one is a kleptomaniac the more auto-
matic his behaviour will appear. The subject's behaviour in
particular situations, then, is organized by his identification of
himself according to the linguistic construct "kleptomania" or its
equivalent. In the framework of role theory, it is this kind of
organization which makes the behaviour recognizably recurrent
in the life history of the person. The fact that the acts are re-
current does not mean that they are prompted from within. It
means only that certain linguistic symbols have become usual for
the person in question.
If this theory were applied, we would not expect apprehended
shoplifters, some ofwhom conceive of themselves as kleptomaniacs,
to provide a logically consistent or even "correct" explanation for
their behaviour. For example, one who has behaved according to
a set of linguistic constructs acceptable to himself in one role
(kleptomaniac) might later discover that both the behaviour and
the constructs are unacceptable to himself in another role he is
playing or desires to play (father, property owner). In that case
there will be a high probability of denial to himself in the second
role that "he" behaved at all. His conception of himself from one
point of view results in denial of the action: "I wasn't myself
when (of if) I did that," "I wasn't feeling well that day." "I
couldn't be the criminal you seek-I couldn't do a thing like
U See the discussion by Nelson Foote, Ope cit.
that," etc. On the other hand, his continued conception of him-

self according to the symbolic constructs which were used in be-
having in the first place probably will result in open confession
that "he" behaved: "Stop me before I do it again," "I have no
control of myself," "That which is 'in' me comes out in situations
like that," "I did it and I'm glad," etc.
In interviewing persons who, according to role theory, have
identified with "compulsive criminals" of some sort we should not
expect them to realize that "who they are" depends upon lan-
guage symbols and, hence, upon arbitrary ascriptions by others.
One who identifies himself as a "kleptomaniac," for example, will
be prone to accept such a conception of himself as ultimate reality.
Even those observers or examiners who use the traditional notion
that compulsive behaviour is an expression of an "inner spring of
action" have considered the subject's conception of himself as
absolute rather than as a group product. As one sociologist has
pointed out, "Because our learning has more often than not been
perfected to the point where cognitive judgments in standardized
situations are made instantaneously, and the energy for perform-
ing the appropriate behaviour is released immediately, it has been
an easy mistake for many observers to suppose that the organic
correlates came first and even account for the definition of the
situation, rather than the reverse."42
In our present state of knowledge we cannot be entirely sure
how one gets committed to particular identities and motives in
the first place but, as indicated, the process is certainly one of
social learning. The differential association theory is a theory of
social learning specifically applied to delinquent and criminal
behaviour, and it contends that, in the terminology used above,
the identifications and motives of criminals are acquired through
direct, personal, contacts with persons sharing those identifications
and motives. As indicated in Chapters V and VII, this theory
may be defective because it does not precisely or adequately
describe or integrate all the aspects of the processes by which
criminality is learned. Nevertheless, it describes the processes by
which one becomes a "compulsive" criminal as well as it de-
scribes the processes by which one becomes a "non-compulsive"
criminal. "Compulsive criminality," as traditionally described is
.. Ibid.
not of such a nature that it is necessarily exceptional to the differ-

ential association theory.
Summary and Conclusions

(1) The assertion that "the latent forces of such phenomena as
compulsive stealing and fire-setting are understood"43 is not
warranted.
(2) If the traditional assumption that all "compulsive crime"
is motivated entirely "from within" is correct, then the use of the
words "compulsive" and "crime" together is erroneous. If the
behaviour were actually prompted "from within" it would be
subsumed under the legal concept "insanity," not "crime."
(3) Re-examination of "compulsive crime" concepts in the
framework of sociological theories of motivation, identification,
and role-playing indicates that most of the legally harmful be-
haviour traditionally labelled "compulsive" actually is "moti-
vated" and has a developmental history which is very similar to
that of other "motivated" behaviour. That legally harmful be-
haviour which is automatic ("non-motivated") cannot be con-
sidered as crime.
(4) Since the development processes in so-called "compulsive
criminality" are the same as the processes in other criminality,
"compulsive crimes" are not, because of something in their
nature, exceptional to the differential association theory. Upon
closer empirical examination it probably will be demonstrated
that criminality which traditionally has been assumed to be
"personal" is actually a group product, and this criminality will
become of more concern to the sociologist than has been the case
in the past .
•3 Robert M. Lindner, Stone Walls and l'vlen, New York: Odyssey Press, 1940, p. 323.
CHAPTER VII
AND TRUST VIOLATION
Although the positive emphasis in Sutherland's theory is that

crime and delinquency are social phenomena, produced by and
through social learning, much of the theory's popularity probably
has arisen from its negative implications, namely that criminality
and delinquency are not biological, psychological or climatic
phenomenon. Similarly, many of the criticisms reported in
Chapter V seem to be addressed more to this negative implication
than towards the theory's positive assertions. The general propo-
sition that individual criminality is learned in a process of differ-
ential association amounts to a "principle" for making sense of
the multiple factors in individual cases, or a "framework" for
study of those cases. However, the details of Sutherland's state-
ment, which he called a "theory" or "hypotheses," as well as a
"principle," have only rarely been specifically subjected to the
test of empirical research. To have continuing value, a theory
must not only provide a general framework within which hypo-
theses may be formulated, as this theory does, but it must also be
stated in such a way that it can be verified or rejected on the basis
of empirical findings.
A genetic explanation of criminal and delinquent behaviour,
the theory emphasizes that it is contact with delinquency or
criminal behaviour patterns which is the necessary condition for
criminality and that it is an excess of contacts of this kind which
produce criminality in the individual. As we have indicated ear-
lier, this is the principle of differential association. "A person be-
comes delinquent because of an excess of definitions favourable to
violation of law over definitions unfavourable to violation of
law," but associations may vary also in frequency, duration,
priority and intensity.
An adequate and complete test of this part of Sutherland's
TRUST VIOLATION 109
theory probably requires the development of a precise formula,

perhaps in mathematical form, stating the relationship between
favourable and unfavourable definitions of law violation and the
modalities affecting these definitions. The past behaviour of crimi-
nals then would be examined to see whether the formula applied
to them, and unknown persons would be examined and their
"criminality scores" computed by means of the formula. A
formula of this kind has not been developed, and preliminary
attempts have indicated that its development will be extremely
difficult. We have attempted, therefore, to test only certain
segments of the theory by determining whether or not certain
necessary elements in the criminal violation of financial trust l are
learned in association with criminal behaviour patterns and, if so,
whether the contacts with such patterns are additive, so that they
could be used in calculating the ratio of contacts with the two
kinds of behaviour patterns.
Two conditions which have been found to be necessary for the
criminal violation of financial trust are the possession of the
appropriate technical information and skill and the possession of
a verbalization ("rationalization," in Sutherland's terminology
and the terminology used in Chapter VI) which enables the
trusted person to adjust his conceptions of himself as a trusted
person with his conceptions of himself as a user of the entrusted
funds or property.2 Taking this as "given," the general problem,
from the point of view of the theory of differential association, is
to determine whether or not these two conditions can be present
in individual cases without the person's having had an "excess" of
associations with criminal behaviour patterns. The specific hypo-
theses tested are as follows: (1) Criminal behaviour patterns of
specific persons or agencies can be identified as the source from
1 This is a sociological concept of crime and it includes all cases in which a crime is
committed in the process of violating a position of trust which has been accepted in
good faith. Thus, almost all persons convicted of embezzlement and larceny by bailee
are included and, in addition, a proportion of those convicted of forgery, confidence
game and using the mails to defraud are included. See Donald R. Cressey, "Crimino-
logical Research and the Definition of Crimes", American Journal of Sociology, 56 (May,
1951), pp. 546-55l.
2 This is not a full statement of the theory of criminal violation of financial trust.
For a more complete statement see Donald R. Cressey "The Criminal Violation of
Financial Trust", American Sociological Review, 15 (December, 1950), pp. 738-743;
and Donald R. Cressey, Other People's Money; A Study in the Social Psyclwlogy rif Embezzle-
ment, Glencoe: The Free Press, 1953.
which the trust violator learns the techniques and information

necessary for the criminal violation of financial trust; (2) criminal
behaviour patterns of specific persons or agencies can be identified
as the source from which the trust violator learns the rationali-
zations necessary for that behaviour. It should be pointed out
that the testing of these hypotheses does not provide a test of the
complete differential association theory since the hypotheses do
not deal with problems of the quantity or quality of contacts with
anti-criminal behaviour patterns. Only those segments of the
theory pertinent to the learning of criminal techniques and
rationalizations are involved.
In a period of two years about 65 persons confined at the
Illinois State Penitentiary at Joliet, 20 confined at the California
Institution for Men at Chino and 40 confined at the United
States Penitentiary at Terre Haute, Indiana, were interviewed
frequently and at length in regard to their common offence of
criminal trust violation. Among other things, they were questioned
in detail about the acquisition of the techniques and rationali-
zations which had been used in the perpetration of their crimes.
On the basis of evidence found in interview materials gathered
from these men, the first hypothesis, that the techniques are
learned in association with identifiable criminal behaviour
patterns, was rejected. It was found that the devices, skills or
techniques used by trust violators to perpetrate the defalcation,
or to hide the defalcation once it has been perpetrated, are the
same skills which are necessary to carrying on the legitimate
routine of the work in the position of trust, and that contact with
criminal behaviour patterns is not necessary.3 The techniques
which are used by trust violators are either those which are known
by all persons of average intelligence or are learned by pro-
fessional and business persons in the course of the training and
experience necessary to obtaining and holding a position of trust.
With one exception, the persons interviewed stated that "any-
one" holding the position of trust which was violated could have
3 The techniques and skills used in trust violation are ordinarily used to conceal
the defalcation, not to perpetrate it. Since in a technical sense the violation of trust
consists of taking the entrusted funds or property with criminal intent, the use of a
technique such as the manipulation of accounts is merely secondary. Trust violators
and others, however, usually consider that the method of concealing the defalcation
is the method of violating the trust.
TRUST VIOLATION 111
violated it with equal ease, as far as techniques were concerned. 4

While in some instances an unskilled layman would not have
been able to violate the position of trust by using a certain techni-
que, neither would that person have been able to hold the position
of trust without possession of such skill. Many of the non-ab-
sconding violators expressed the idea that the violation was
possible because of favourable personal relations with persons
who should have checked on their work and that, consequently,
another person, whose personal relationships were not so favour-
able, might not have been able to use the same system for vio-
lation of the trust position. However, this cannot be considered as
a technique or skill and it is not so considered by the violators
interviewed. In a sense, statements of this kind amount to a mere
reiteration of the fact that one must be in a position of trust be-
fore he can violate that trust.
As to the techniques themselves, an accountant with consider-
able experience in public accounting and auditing expressed the
views of the other violators when he claimed that one could not
graduate from a course in accounting without being familiar with
the techniques which are used by accountants who violate
positions of trust:
I don't think you can identify the place at which a man learned to do his
illegal act. It would be the same with burglary as it is with embezzlement.
How would a burglar be able to tell you where he got the attitudes he had
when he went out? He might be able to say that he learned a few techniques
here and there but that is about all. In my case, I would have to say that I
learned all ofit in school and in my ordinary accounting experience. In school
they. teach you in your advanced years how to detect embezzlements, what to
check to detect them, what to do to prevent them, and you sort of absorb it.
There is no way of knowing exactly where you learn any habit. The same thing
applies to the question oflearning to use it for illegal rather than legal means.
It is just like a doctor performing abortions. In his medical training he must
learn to conduct the abortion, because many abortions are necessary for the
health of the mother. Maybe he will perform a few legitimate abortions and
then an illegitimate one. He has learned to conduct the illegitimate one in his
ordinary medical training but he could not identify the point at which he
learned that, because he would have to include all of his courses in physiology,
anatomy and everything else, as well as the specific technique. In my case I
did not use any techniques which any ordinary accountant in my position
could not have used; they are known by all accountants, just like the abortion
technique is known by all doctors.
• In the one case the violator claimed that he invented the system used for defalcation.
This is highly doubtful, however, since his system was one which is used in all parts of
the United States by persons in similar capacities.
The situation in trust violation, then, is not unlike the learning

of certain of the techniques essential to some other types of crime.
One who hunts or shoots a gun has the necessary technical skill
for murder; an automobile mechanic ordinarily has the technical
skill necessary for theft of a locked automobile; almost everyone
has the technical skill necessary for simple theft. In all of these
instances, and in trust violation as well, known techniques which
have been learned for legitimate purposes and in contact with
law-abiding persons and behaviour patterns conducive to obeying
the law, could be used for illegal rather than legal purposes. Be-
fore trust violation occurs it is not necessary for the trusted per-
sons to learn new techniques for committing the crime; instead, it
might be said that known techniques undergo a process of
"circumstantiation" so that they are used for crime.
The second hypothesis was formulated after it had first been
determined that trust violation occurs only when the trusted
person applies to his own conduct a key verbalization which
adjusts for him contradictory ideas and values regarding crimi-
nality on the one hand and integrity, honesty and morality on the
other hand. Since a rationalization that one is "borrowing"
rather than "stealing" or "embezzling" the entrusted funds, for
example, must be learned, it is inconceivable that it could be
present unless the individual using it had been in contact with
persons who presented it to him or had been in contact with some
other cultural source which gave him a general acquaintance
with it. Our second hypothesis is that such contacts are contacts
with the "criminal behaviour patterns"5 of specific persons or
agenczes.
Unlike the cultural conditions which exist in respect to the
presentation of the techniques and general information to the
individual necessary to the criminal violation of financial trust,
which may be characterized as conditions which inform the
individual that positions of trust can be violated, the cultural
conditions which exist in regard to the presentation of definitions
of situations in which positions of trust may be violated are con-
flicting. As we just indicated, a technique necessary to trust vio-
5 For the specific problem here, a criminal behaviour pattern may be considered
as a definition of a situation in which the criminal violation of financial trust
is appropriate.
TRUST VIOLATION 113
lation might be learned from an individual who presents it as a

technique for preventing or detecting trust violation. But it is not
possible for a verbalization which justifies trust violation to be
presented in this way since, by definition, such verbalizations
must be and are contradictory to the ideas inherent in the words
"trust," "trustworthiness," "honesty" and so forth. It is impossible
for one person to present to another an idea that there are con-
ditions under which positions of trust may be violated criminally
without presenting to him a criminal behaviour pattern.
For this reason, the general implication of our second hypothe-
sis, and of the differential association theory as it applies to the
rationalizations of trust violators, is necessarily correct. That is,
the rationalizations which are applied to the person's own con-
duct in the criminal violation of financial trust are learned in
association with criminal behaviour patterns. It is not possible for
trust violators to use rationalizations in the manner indicated
without first having come into contact with definitions of situ-
ations which to a greater or less degree sanction the criminal
violation of financial trust. 6 The important theoretical questions
remaining are those concerned with the quantity and quality of
such contacts and, therefore, with the specific source of the defi-
nitions of situations which are applied to the violator's own
conduct as rationalizations. If the source ofa verbalization can be
identified, then that half of the ratio which pertains to the learn-
ing of criminal behaviour patterns can be crudely calculated.
However, in the large proportion of cases, it is impossible, or at
least extremely difficult, to identify the source from which a
trusted person learns the fundamental cultural contradiction
inherent in the notion that non-violation of trust is an expected
norm but that there are conditions under which trust may be
violated. Our subjects were asked in at least four different ways to
identify this source, 7 but it could not be specifically identified by
• However, the fact that the violator must verbalize in such a way that criminal
behaviour seems somehow justified, is itself evidence of contact with different sets of
values and of the fact that he has not completely assimilated criminal values while
eliminating anti-criminal values.
7 At different times in the various interviews they were asked their opinions as to
(I) where they got the idea to "borrow" or, generally, to do what they did, (2) how
they happened to hit on the idea which eventually resulted in their incarceration,
(3) why they had not violated the position of trust at an earlier time, (4) whether they
had observed criminal or unethical practices, or discussion of such practices, on the
part of their trustors or other associates. Other questions soliciting opinions in regard
either the subject or the investigator in at least 80 per cent of the

cases. Rather than naming a specific source, the subjects referred
directly or indirectly to rather general cultural ideologies8 with
which they had informal contact at some vague period in their
lives. For this reason, that portion of the second hypothesis which
pertains to the identification of the specific sources of the ration-
alizations was rejected and a calculation of the differential associ-
ation ratio could not be attempted.
There are numerous reasons for the trust violators' inability to
designate the specific persons or agencies from which they learned
that there are conditions under which positions of trust may be
violated. In the first place, trust violators, even while in prison,
seldom identify with an ideal-type criminal, and their responses
to questions pertaining to criminality often are made with con-
sideration for the fact that they have had little or no contact with
ideal-type criminals prior to the defalcation. Most of the violators
interviewed did not consider the questions to ask about their
having had contact with, for example, persons who presented
business or professional ideologies which contain definitions in
which trust violation is an appropriate mode of response. When
to the source of the verbalizations also were asked in some cases, and in addition all
voluntary remarks pertinent to these questions were recorded.
8 "When rationalizations are extensively developed and systematized as group
doctrines and beliefs, they are known as ideologies. As such, they acquire unusual
prestige and authority. The person who uses them has the sense of conforming to
group expectations, of doing the 'right thing' ... Unscrupulous and sometimes criminal
behaviour in business and industry is justified in terms of an argument which begins
and ends with the assertion that 'business is business' ... The principal advantage of
group rationalizations or ideologies, from the individual's standpoint, is that they
give him a sense of support and sanction. They help him to view himself and his
activities in a favourable light and to maintain his self-esteem and self respect."
A. R. Lindesmith and A. L. Strauss, Social Psychology, New York: The Dryden Press,
1949, pp. 309-310.
An anthropologist has given us an example of the process by which such ideologies
produce non-conformity to the expected norms: "With reference to its code, any fair
sized community is bound to have its martinets and its outlaws. The average member
neither flouts tradition at all costs or follows its guidance through thick and thin:
he compromises, rendering obeisance to fine principles in the abstract and finding
excellent excuses for doing as he pleases in concrete circumstances. The Burmese are
Buddhist, hence must not take the life of animals. Fishermen are threatened with dire
punishment for their murderous occupation, but they find a loophole by not literally
killing the fish. 'These are merely put on the bank to dry, after their long soaking in the
river, and if they are foolish and ill-judged enough to die while undergoing the process
it is their own fault.' ... When so convenient a theory had once been expounded, it
naturally became an apology of the whole guild of fishermen." Robert H. Lowie,
An Introduction to Cultural Anthropology, Enlarged Edition, New York: Rinehart, 1940,
p.379.
TRUST VIOLATION 115
they were specifically questioned about this in later interviews,

the responses indicated that in most cases this idea was incompre-
hensible to them. An individual's failure to identify with an ideal-
type criminal resulted in a conception of himself as something
other than a criminal or an ideal-type violator, and questions
pertaining to the sources of rationalizations were interpreted to
be questions about non-criminal activities.
For example, persons who "borrowed" (criminal behaviour)
over a period of time but who did not, in their own estimation,
even "exceed the limits" of their own resources, did not consider
themselves as criminals in any sense of the term and hence did not
consider the conditions under which they had learned to "borrow"
as being at all important in criminal activities. Similarly, among
independent businessmen who converted "deposits", the belief
that their criminal behaviour was closely related to ordinary non-
criminal business practice made the situation such that the
violators did not consider the circumstances in which these
practices (and the rationalizations for them) were learned as
having provided them with definitions of situations in which trust
violation is sanctioned. Instead, they considered the definitions to
have been presented in much the same way that the techniques
necessary to violation are presented. In a strict sense, such business
activities as using entrusted funds which are "covered" by one's
own securities might be criminal also, but the violators considered
them to be so extensive that they felt that the situations in which
they are learned have nothing to do with criminality. Such
practices, or even criminal practices in a more obvious sense, are
sometimes presented to persons just beginning employment or
enterprise in a business, and the individual's success or continued
employment in such business is dependent upon his acceptance of
them. The individual either accepts the definitions which sanction
criminal behaviour or he is eliminated from the business by
competition from men who have accepted them. 9
Second, the entrance of a particular person into the aggregate
of persons who have violated positions of financial trust is un-
obstructed by the desires of such other persons. A trusted person
does not need the permission of others in order to become a trust
8 Sutherland reiates several such cases. Edwin H. Sutherland, White Collar Crime,
New York: The Dryden Press, 1949, pp. 235-239.
violator and the process which results in trust violation does not
require direct, personal contact and communication among trust
violators comparable, for example, to the communication among
professional thieves. Trust violators have no tutelage systems, no
special requirements for admission to practice and no special
argot. What is prevalent among the associates of trusted persons
is not an argot referring to illegal or quasi-legal practices, but
definitions, in ordinary language, which sanction criminality and
which, when personalized, become rationalizations. Identification
of specific persons as the ones who presented the verbalizations is
therefore difficult.
Third, contacts with definitions of situations in which trust
violation is sanctioned are not in fact necessarily made within a
particular business organization or with a specific person or
agency, but in most instances are obtained as a consequence of
the person's more general contacts with ideologies conducive to
criminal behaviour. Neither trust violators nor persons who are
not even criminals can ordinarily identify the specific source of a
philosophy which includes adherence to such ideologies. The
"Jean Valjean philosophy," for instance, is one which sanctions
criminal behaviour, yet most people who adhere to it, or say that
they adhere to it, probably cannot identify its source. Similarly,
the source of the idea that it is not criminal, or at least not
"completely criminal," to cheat a large impersonal corporation
usually cannot be identified. Specifically in trust violation we find
the same situation. Even before they accept positions of trust,
persons learn fundamental cultural contradictions in regard to
criminal and non-criminal behaviour in such positions, but they
usually do not know the specific source of the patterns which have
been learned and often do not even recognize that there is a
contradiction in the cultural ideologies which govern their
behaviour. In many instances the presentation of a definition
conducive to violation of trust is made in such a way that at the
time no stigma is attached either to taking over the definition or
rejecting it, and the individual is not specifically aware ofthe fact
that the ideology entails a question of criminality or non-crimi-
nality. Individuals develop general conceptions of what is
"proper" in certain situations for persons of their status and when
these situations appear they behave in terms of those conceptions.
TRUST VIOLATION 117
Regarding the second hypothesis, then, we have shown that

contact with criminal behaviour patterns is necessary to trust
violation inasmuch as conscious, learned rationalizations are
necessary to the crime, but we have found that the specific source
of those rationalizations cannot be identified precisely. Demon-
stration of the fact that there are contradictory definitions in
regard to trust violation is much less difficult than demonstration
of the exact source, let alone the quantity and quality, of contacts
with each of the types of definitions. The general emphasis of the
differential association theory, that the motives or rationalizations
necessary to trust violation have been learned and that the
absence of such learning will prohibit trust violation, is correct.
However, the associations of trusted persons apparently do not
vary in such a way that it may be said that just prior to the crime
there was a presentation of an excess of definitions favourable to
the criminal violation of financial trust. If there is such a variation
in associations it cannot be measured, since the source of the
significant behaviour patterns cannot be specifically identified. 10
It is highly probable that the associations with the patterns
ordinarily have taken place some time prior to the crime and have
involved assimilation of definitions of situations in which trust
violation is sanctioned. Hence, when the individual later perceives
that the situation in his case is included in the general category of
situations in which trust violation is sanctioned, he violates the
trust.
Conclusions
While the general contention of the differential association
theory, that criminality is learned, cannot be disputed, the more
specific idea that criminality and non-criminality depend upon a
ratio of contacts with criminal and anti-criminal behaviour
patterns is open to question in cases of crimes involving violation
of financial trust. In the first place, contacts with criminal be-
haviour patterns are not necessary to the learning of the technique
or skill used in trust violation. Second, while the present research
10 While no attempt to identify the sources of anti-criminal behaviour patterns
was made in this study, it is obvious that such identification would be even more
difficult than the identification of criminal behaviour patterns. In explaining non-
violation in terms of a ratio of contacts with criminal and anti-criminal behaviour
patterns one encounters difficulties very similar to those discussed.
has not indicated definitely that such contacts cannot be precisely

identified and weighted, it does appear that there is no practical
way that known violators' prior contacts with the rationalizations
necessary to trust violation can be observed so that one could
develop a formula to be used in either the determination or
prediction of trust violation in other cases. Only if we could
observe almost every association of a person during his entire life
could we say definitely that he had been exposed to an excess of
criminal behaviour patterns and this, in turn, could be done only
if the distinction between criminal and anti-criminal behaviour
patterns were carefully drawn. Since such observations cannot be
made, even by parents, it is doubtful that it can be shown
empirically that the differential association theory applies or does
not apply to crimes of financial trust violation or even to other
kinds of criminal behaviour.
Third, the general implication of the present findings is that
the differential association theory should be modified in such a
way that it is subject to empirical test. One possible modification
would be simply to delete those portions of the theory which refer
to the ratio between contacts with criminal and anti-criminal be-
haviour patterns. The important theoretical position that criminal
behaviour is necessarily learned behaviour would remain, and
this position could be verified empirically. Another possible modifi-
cation would entail the substitution of a different conception of
the process by which criminality is learned for the conception of a
differential in the quantity and quality of contacts with the two
varieties of behaviour patterns. For example, the present research
findings might be generalized to indicate that the verbalizations
used by all criminals play an exceedingly important role in the
determination of their criminality, and a search for the differences
in the typical vocabularies used by criminals and non-criminals
in specific situationsl l might reveal that it is the presence or
absence of a specific, learned, verbal label in a specific situation
which determines the criminality or non-criminality of a particu-
lar person. 12
11 Cf, C. Wright Mills, "Situated Actions and Vocabularies of Motive", American
Sociological Review, 5, December 1940, pp. 904-913.
12 Since the time of publication of the original version of this chapter, the theory of
differential association has been subjected to empirical test by other investigators,
with much more favourable results than the above. The following articles are of most
TRUST VIOLATION 119
relevance: John C. Ball, "Delinquent and Non-Delinquent Attitudes Toward the

Prevalence of Stealing", Journal oj Criminal Law, Criminology, and Police Science, 48
(September-October, 1957), pp. 259-274; Daniel Glaser, "A Reconsideration of
Some Parole Prediction Factors", American Sociological Review, 19 (June, 1954), pp.
335-351; Daniel Glaser, "Criminality Theories and Behavioural Images", American
Journal oj Sociology, 56 (March, 1956), pp. 422--444; Daniel Glaser, "Differential
Association and Criminological Prediction", Social Problems, 8 (Summer, 1960),
pp. 6-14; Robert E. Lane, "Why Businessmen Violate the Law", Journal ojCriminal
Law, Criminology and Police Science, 44 (July-August, 1953), pp. 151-165; Edwin M.
Lemert, "An Isolation and Closure Theory of Naive Check Forgery", Journal of
Criminal Law, Criminology, and Police Science, 44 (September-October, 1953), pp. 296-
307; Edwin M. Lemert, "The Behavior of the Systematic Check Forger", Social
Problems, 6 (Fall, 1958), pp. 141-149; Henry D. McKay, "Differential Association
and Crime Prevention: Problems of Utilization", Social Problems, 8 (Summer, 1960),
pp. 25-37; James F. Short, Jr., "Differential Association, Delinquency, and Self-
Conception Among a Group of Institutionalized Delinquents", paper read at the
annual meeting of the American Sociological Society, September, 1956; James F.
Short, Jr., "Differential Association and Delinquency", Social Problems, 4 (January,
1957), pp. 233-239; James F. Short, Jr., "Differential Association with Delinquent
Friends and Delinquent Behavior", Pacific Sociological Review, 1 (Spring, 1958), pp.
20--25; James F. Short, Jr., "Differential Association as a Hypothesis: Problems of
Empirical Testing", Social Problems, 8 (Summer, 1960), pp. 14-25; Gresham Sykes
and David Matza, "Techniques of Neutralization: A Theory of Delinquency",
American Sociological Review, 22 (December, 1957), pp. 664-670; StantonJ. Wheeler,
Social Organization in a Correctional Communiry, Unpublished Ph. D. Dissertation,
University of Washington, 1958. The following items, written before the original
version of this chapter, also treat differential association from an empirical perspective:
Marshall B. Clinard, "The Process ofU rbanization and Criminal Behavior", American
Journal of Sociology, 48 (September, 1942), pp. 202-213; Marshall B. Clinard, "Rural
Criminal Offenders", American Journal ofSociology, 50 (July, 1944), pp. 38-45; Marshall
B. Clinard, "Criminological Theories of Violations of Wartime Regulations",
American Sociological Review, 11 (June, 1946), pp. 258-270; Marshall B. Clinard, The
Black Market, New York: Rinehart, 1952.
III
USE OF THE THEORY IN CORRECTIONS

CHAPTER VIII
AND CORRECTIONAL GROUP THERAPY
Sociological theories and hypotheses have had great influence

on development of general correctional policies, such as pro-
bation and parole, but they have been used only intermittently
and haphazardly in reforming individual criminals. Since soci-
ology is essentially a research discipline, sociologist-criminologists
have devoted most of their time and energy to understanding and
explaining crime, leaving to psychiatrists and others the problem
of reforming criminals. Even the sociologists employed in cor-
rectional work have ordinarily comitted themselves to non-
sociological theories and techniques of reformation, leading the
authors of one popular criminology textbook to ask just what
correctional sociologists can accomplish which cannot be ac-
complished by other professional workers.1
Perhaps the major impediment to the application of soci-
ological theories lies not in the nature of the theories themselves
but, instead, in the futile attempt to adapt them to clinical use.
Strictly speaking, the now popular policy of "individualized
treatment" for delinquents and criminals does not commit one to
any specific theory of criminality or any specific theory of refor-
mation, but, rather, to the proposition that the conditions con-
sidered as causing an individual to behave criminally will be
taken into account in the effort to change him.
Historically, the policy of individualized treatment in cor-
rections developed as a reaction to the 18th-century attempt to
impose uniform penalties on criminals. It was, and still is, argued
that policies calling for uniform punishments are as obviously
ineffective as would be a policy calling for uniform treatment for
all medical patients, no matter what their ailments. The alterna-
1 Harry Elmer Barnes and Negley K. Teeters, New Horizons in Criminology, New
York: Prentice-Hall, Inc., 1951, p. 644.
124 USE OF THE THEORY
tive eventually proposed was a system in which the handling of

each case of criminality would include expert diagnosis, expert
prescription of therapy and expert therapy-just as clinical
medicine includes diagnosis, prescription and therapy. An at-
tempt is made to diagnose the cause of the criminality and to base
the technique of reform upon the diagnosis. Analogy with the
method of clinical medicine is obvious.
However, by far the most popular interpretation of the policy
of individualization at present is that the theories, as well as the
methods, of clinical medicine must be used in diagnosing and
changing criminals. It is the emphasis on this clinical principle that
has impeded the application of sociological theories like Suther-
land's and, it may be conjectured, success in correctional work.
One kind of personality theory views the individual as an es-
sentially autonomous agent who must submit to the rules and
regulations of society and other organizations. "Personality" is
considered an outgrowth of the effect that the "restrictions"
necessary to organization have on an individual's expression of
his own, pristine, needs. Consistent with this general theory of
personality development is criminological theory maintaining,
essentially, that criminality is a personal trait or characteristic,
the property if the individual criminal exhibiting it. An extreme
position is that criminality is a biological phenomenon. Much
more popular is the theoretical position that criminality is a
psychological defect or disorder, or a "sympton" of either. The
criminal is viewed as one unable to canalize or sublimate his
"primitive", antisocial impulses or tendencies,2 who is expressing
symbolically in criminal behaviour some unconscious wish or urge
arising from an early traumatic emotional experience, 3 or as a
person suffering from some other kind of defective trait or con-
dition. The general point is that a "healthy" or "adjusted"
personality is one which does not own criminality because it has
been permitted freely to express itself in numerous alternative
ways.
The general implication of this kind of theory is that since
2 Sheldon and Eleanor T. Glueck, Delinquents in the Making, New York: Harper &
Bros., 1952, ppl 162-63; see also Ruth Jacobs Levy, Reductions in Recidivism through
Theraf~Y, New York: Seltzer, 1941, pp. 16,28.
3 EdwinJ. Lukas, "Crime Prevention: A Confusion in Goal", in Paul W. Tappan,
Contemporary Correction, New York: McGraw-Hill Book Co., 1951, pp. 397-409.
CORRECTIONAL GROUP THERAPY 125
criminality is an individual disorder it should, like a biological

disease, be corrected on a clinical basis. Consistent with the most
extreme theoretical position is the notion that the criminal's
anatomy or physiology is to be clinically modified through
lobotomy, castration, modification of glandular functioning or
something else. More popular is the idea that criminality is to be
corrected through psychological attention. But in either case the
implication is that criminality should be treated clinically. In a
sense, criminality is viewed as analogous to an infectious disease
like syphilis-while group relationships of various kinds are
necessary to the disorder, the disorder can be eradicated in a
clinic, without reference to the persons from whom it was ac-
quired. 4 Because criminality is an expression of individual
psychological disorders, it is to be corrected by elimination of the
disorders. But because the disorders, in turn, spring from the
restrictions society has placed on "free" individuals, correction of
them must be in the form of modifYing the impact of the re-
strictions on the individual. Although this might have reformation
of society as one of its implications, individual criminals are to be
corrected clinically, by giving them relief grom the restrictions-in
the form of "ventilation," "catharsis," "acting out" and other
devices for removing "tensions," "aggression," "unconscious
tendencies and wishes," and other individual disorders. Indivi-
dual psychotherapy, as a system for reforming criminals, is
probably the best example of a current treatment programme in
which the major effort is to implement this clinical principle of
rehabilitation.
Gradually, the treatment methods based on this clinical
principle are being supplemented by a different kind of indivi-
dualized treatment. Sociologists and social psychologistst have
provided an alternative theory of personality, namely that the
behaviour, attitudes, beliefs and values which a person exhibits
are not only the products of group contacts but are the properties of
groups. The basic notion here is that the "organization" of social
interaction and "personality" are two facets of the same thing.
The person is viewed as a part of the social relationships and
values in which he participates; he obtains his satisfactions and,
• See, for example, Charles Berg, "The Psychology of Punishment", British Journal
of Medical Psychology, 20 (October, 1945), pp. 295-313.
in fact, his essence, from participation in the rituals, rules, schedules,

customs and regulations of various kinds which surround him.
The person (personality) is not separable from the social re-
lationships in which he lives. He does not submit to society's
restrictions; he behaves according to the rules (which are some-
times contradictory) of the large organization, called "society," in
which he participates; he cannot behave in any other way. If this
theory about personality is correct and the behaviour of an
individual is an intrinsic part of the groups to which he belongs,
then attempts to change the behaviour must be directed at
groups. John Dewey in 1922, and Dorwin Cartwright in 1951,
expressed this group relations principle for changing behaviour, as
follows:
To change the "working character" or will of another, we have to alter
objective conditions which enter into his habits. Our own schemes ofjudgment,
of assigning blame and praise, of awarding punishment and honour are part
of these conditions ... We cannot change habit directly: that notion is magic.
But we can change it indirectly by modifying conditions, by an intelligent
selecting and weighing of the objects which engage attention and which
influence the fulfilment of desires. 5
The behaviour, attitudes, beliefs and values of the individual are all firmly
grounded in the groups to which he belongs. How aggressive or cooperative
a person is, how much self-respect or self-confidence he has, how energetic and
productive his work is, what he aspires to, what he believes to be true and
good, whom he loves or hates and what beliefs and prejudices he helds - all
these characteristics are highly determined by the individual's group member-
ship. In a real sense, they are properties of groups and of the relationships
between people. Whether they change or resist change will, therefore, be
greatly influenced by the nature of these groups. Attempts to change them
must be concerned with the dynamics of groups. 6
In criminology, Sutherland's theory of differential association

is consistent with the general personality theory which views the
person as an integral part of the kinds of social relationships in
which he participates. Basically, this theory holds that criminal
behaviour, like other behaviours which an individual may ex-
hibit, is the property of groups. The implication is that the group
relations principle, not the clinical principle, will be most success-
ful in attempts to eradicate criminality from the total behaviour
of any person. As we have seen, Sutherland's statement holds that
5 John Dewey, Human Nature and Conduct, New York: Henry Holt, 1922, pp. 19-20.
• Dorwin Cartwright, "Achieving Change in People: Some Applications of Group
Dynamics Theory", Human Relations, 4 (1951), pp. 381-392.
persons become criminals principally because they have been

relatively isolated from groups whose behaviour patterns (in-
cluding attitudes, motives and rationalizations) are anti criminal,
or because their residence, employment, social position, native
capacities or something else has brought them into relatively
frequent association with the behaviour patterns of criminal
groups. A diagnosis of criminality based on this theory would be
directed at analysis of the criminals attitudes, motives and
rationalizations regarding criminality and would recognize that
those characteristics depend upon the groups to which the criminal
belongs.
The proponents of the group relations principle have not
abandoned the belief in the importance of giving attention to
individual cases, but the theory used in diagnosing, prescribing
and "treating" has been modified. If the criminality of an
individual depends upon group relations, then the prescription
for "treatment" must be a prescription for modification of his
individual group relations. Sutherland's theory implies that if
delinquents and criminals are to be changed, either they must-
become members of anticriminal groups, or their present pro-
criminal group relations must be changed. 7 Negatively, it is
suggested that the group relations which support criminality
cannot be directly modified in a clinic in the way that the condi-
tion of a person suffering from syphilis can be modified in a
clinic; they can be modified only by providing the criminal with
new social relations or in some way changing the nature of
present group relations.
Numerous policies or programmes arising in correctional work
during the last one hundred years have been explicitly or im-
plicitly based upon recognition of the necessity for modifying
criminality by changing the social relationships of the criminal.
Probation and parole, of course, are perhaps the best examples
of such programmes. In prison work it has been recognized since
the time of Maconochie, at least, that the offender can best be
trained for participation in law-abiding society by providing him
7 Cf Donald R. Taft, "The Group and Community Organization Approach to
Prison Administration", Proceedings of the American Prison Association, 72 (1942),
pp. 275-284; and George B. Void, "Discussion of Guided Group Interaction in COTTectional
Work by F. Lovell Bixby and Lloyd W. McCorkle", American Sociological Review, 16
(August, 1951), pp. 460-461.
with contacts with that society. Although prisoners continue to be

effectively separated from law-abiding groups and from most
contacts with the kinds of social relations in which they will be
expected to participate after release, the isolation of inmates has
gradually been reduced by the "prison reforms" undertaken for
humanitarian reasons during the last century. It has also been
reduced by systems of individualized treatment which, while not
explicitly acknowledging the importance of group associations on
criminality and reformation, have promoted social relations with
law-abiding groups.8
The Group Relations Principle and Correctional Group Therapy

"Group therapy" has recently become popular in prisons or,
at least, in the thinking of correctional workers, 9 and it is timely
to consider whether this kind of treatment incorporates the group
relations principle for reforming offenders. Because the word
"group" is included in the name, group therapy is often assumed
to be directed at modification of the criminality of individuals by
reintegrating them into the social relations of non-criminal
groups. Accordingly, group therapy is often believed to be
consistent with Sutherland's theory. However, it is conceivable
that most correctional group therapy has little or nothing to do
with the participant's social relations and that it is, instead,
merely a new system for applying the notion that criminality can
be "cured" or treated in a clinical situation.
At the outset, a distinction must be made between the use of
group sessions for treating psychiatric patients and the use of
group sessions for reforming prisoners. The current emphasis on
group therapy, also called "group psychotherapy" and "group
psychoanalysis," grew out of the difficulty of treating mental cases
individually during World War II. There has been little careful
research and experimentation on the suqject, but there is an al-
most unanimous opinion that group therapy is a remarkably
effective technique for treating mental patients. lO Although much
8 See Donald R. Cressey, "The Nature and Effectiveness of Correctional Tech-
niques", Law and Contemporary Problems, 23 (Autumn, 1958), pp. 754-771.
9 See Lloyd W. McCorkle, "The Present Status of Group Therapy in United States
Correctional Institutions", International Journal of Group Psychotherapy, 3 (January
1953), pp. 79-87.
10 Marshall B. Clinard, "The Group Approach to Social Reintegration", American
Sociolol!,irai Review, 14 (April 1949), pp. 257-262.
group therapy for mental patients is merely individual psycho-

therapy simultaneously administered to a number of persons,
many psychiatrists who have used group techniques have re-
cognized the value of the group itself as a medium of change in
individuals. Also, the use of group sessions has led some psychia-
trists to abandon the conventional, individualistic explanations of
mental disorders and to recognize the importance of group
participation to psychological adjustment. As a system for dealing
with mental patients, the principal contribution of group therapy,
in contrast to "simultaneous individual therapy," has been the
elimination of reduction of social isolation and egocentricity,ll or,
stated positively, the assimilation of the isolated or egocentric
patient into the clinical group.
But the group relations principle for reforming prisoners im-
plies that there must be more than an integration of the prisoner
in a clinical group. In treatment based on this principle, the aim
is not mere reduction ofisolation and belligerence among prisoners
as they operate in the prison situation (although this is important
to the smooth operation of a prison), but the provision of positive
contacts with groups which will directly or indirectly implant in
the prisoner the anticriminal values of the larger society. The
emphasis is on reform or reorientation of the prisoner, and mere
evidence of psychological stability in the prison situation is not
considered evidence of reform. Hence, if the clinical group
therapy techniques, even of the type which actually attempt to
use the group as a medium of change, were uncritically adopted
in prisons, there would be little positive emphasis on altering the
social relations which may be specifically conducive to an in-
mate's criminality.
Undoubtedly, many existing group therapy programmes in
prisons are, in fact, mere uncritical adoptions of the clinical
techniques. Because the clinical principle has great popularity,
one would expect prison group therapy programmes to emphasize
simultaneous individual therapy, in the form of lectures, music,
athletics, crafts, collective psychoanalysis or other activities in
which the therapy is administered to a collection of individuals.
11 s. R. Slavson, An Introduction to Group Therapy, New York: The Commonwealth

Fund, 1943, p. 1; William C. Menninger, Psychiatry in a Troubled World, New York:
The Macmillan Company, 1948, pp. 316-317.
For the same reason, one would expect that where the therapy
group was actually considered a medium of change it would be
used as it is in most clinical therapy.12 That is, it would be based
upon the clinical principle and would be used merely to reduce
individual isolation and to provide a permissive setting in which
belligerent inmates could "ventilate" suppressed hostilities to-
wards the police, court personnel, prison officials and others.
According to the proponents of the clinical principle, this "re-
forms" inmates by enabling them to rid themselves of certain
individual emotional disorders which are considered the causes of
their criminality. The relevant point is that this kind of correc-
tional group therapy does not incorporate the group relations
principle and is not, in fact, different in aims or objectives from
individual, clinical therapy. It could not be more efficient and it
might be less efficient than individual correctional therapy.
Because numerous reports have been written about it, and be-
cause its aims have been stated, New Jersey's correctional group
therapy programme may be used to indicate some of the contra-
dictions and problems which may appear when a group therapy
programme attempts to incorporate both the clinical principle
and the group relations principle. Significantly, the therapists in
this programme have tried to go beyond mere ventilation and
reduction of isolation and have attempted in a more positive
way to utilize the therapy group for reformation of offenders.
This may be inferred from the fact that "guided group inter-
action" was selected as the name for the programme "to avoid
confusion with the use of group therapy as practised by psychia-
trists, and to avoid any implication that all prison inmates are
mentally abnormal or unbalanced."13 Also, it is said that in
guided group interaction "the major emphasis is on the group
and its development rather than on a attempt at exhaustive
psychoanalysis of individuals in the group."14 Similarly, cogn-
izance of the group relations principle of reformation seems to be
12 McCorkle, op. cit., does not describe the various kinds of therapy now being used
in correctional institutions, but his classificatory categories seem to support the notion
that the orientation in most programmes designated as "group therapy" is indi-
vidualistic.
13 Lloyd W. McCorkle, "Group Therapy", in Paul W. Tappan, Editor, Con
temporary Correction. New York: McGraw-Hill, 1951, pp. 211-223.
14 Lloyd W. McCorkle, "Group Therapy in the Treatment of Offenders", Federal
Probation, 16 (December 1952), pp. 22-27.
indicated by the following definition and statement of the aims

of guided group interaction: "The use of free discussion in a
friendly supportive atmosphere to re-educate the delinquent to
accept the restrictions of society by finding greater personal
satisfaction in conforming to social rules than following delinq uent
patterns." 15
However, a description of the specific process or processes by
which the stated aims of guided group interaction are to be
accomplished has not been located. Rather, McCorkle suggests
in one place that the reader make his own interpretation of the
material presented. 16 Following that suggestion, we have observed
that the programme seems to be based, implicitly or explicitly, on
four principal expectations about the processes by which the
group sessions contribute to individual reformation. Despite the
fact that the stated aims of the therapy programme imply that it
does much more than attempt a clinical "cure" for criminality,
three of the expectations are consistent only with the clinical
principle, one with the group relations principle.
Getting beneath the surface. First, it is expected that "free dis-

cussion" of an inmate's problems and personality characteristics
by and with a therapist and an inmate group "in a friendly
supportive atmosphere" will both enable and force the inmate to
"get beneath the surface" and "face the facts" of his caseY In the
permissive group session the therapist does not "blame" or cen-
sure the inmate, but fellow inmates who have had experiences
similar to his will draw him out and will not permit him to lie,
bluff, "have his own way" or provide ex postfacto justifications for
his criminal behaviour. Presumably, the individual eventually
will accept his fellow inmates' friendly analyses, rejections and
denunciations of his behaviour and rationalizations more readily
15 Ibid. This statement is a revision of a definition used in earlier publications: "The
use of free discussion to re-educate the delinquent to accept the restrictions of society
and to find satisfaction in conforming to social norms". F. Lovell Bixby and Lloyd
W. McCorkle, "Applying the Principles of Group Therapy in Correctional Insti-
tions", Federal Probation, 14 (March 1950), pp. 36-40.
16 "Group Therapy", op. cit.
17 "The essence of group therapy is to help the individual adjust to reality", Bixby
and McCorkle, op. cit.; "[Guided group interaction] assumes the mutual give-and-
take of group discussion will stimulate the inmate to some insight into the relationship
between what takes place in this learning situation and his immediate problems of
living". McCorkle, "Group Therapy in the Treatment of Offenders", op. cit.
than he would accept analyses, rejections and denunciations of the

same behaviour and rationalizations by an outsider. However, in
at least one attempt at group therapy with prisoners this expec-
tation was not fulfilled-the sessions never got beyond the
"testing-out stage" because, among other things, the participants
would not accept the proposition that "bad luck" or a "bad
judge" was not the source of their predicament. Is
Acquiring insight. Second, it is expected that this guided group

stimulation to "face the facts" will give the individual psychcho-
logical "insight" and will "help him find himself" by enabling
him to see that his problems are identifiable in the psychological
terminology of the session leader. Such insight, combined with the
opportunity to ventilate, presumably will reform him. This is
obviously in keeping with the hypothesis, based on the clinical
principle, that if a delinquent or criminal becomes convinced that
his criminality is due to such attitudes or conditions as "feelings
of weakness," "resentment of authority," "feelings of guilt" or
"frustration," and is then able, through ventilation, to dissipate
the "tensions" and "anxieties" arising from these undesirable
attitudes, he will be reformed. 19
Accepting general restrictions of society. Third, it is expected that

the experience of accepting the analyses, arguments and re-
strictions of the inmate group will give the individual needed
practice in accepting the general "restrictions of society."2o This
expectation is consistent with the clinical assumption that there is
18 Jay W. Fidler, "Possibility of Group Therapy with Female Offenders", Inter-
national Journal of Group Psychotherapy, I (November 1951), pp. 330-336.
19 See S. R. Slavson, "Group Psychotherapy in Delinquency Prevention", Journal
of Educational Sociology, 24 (September 1950), pp. 45-51; and Justin K. Fuller, "Group
Therapy for Parolees", Prison World, 14 (July-August, 1952), pp. 9-11.
20 The similarity to group therapy used for reducing belligerence and isolation of
mental patients may be observed here. McCorkle has emphasized that "Group
sessions tend to destroy the routinized, almost vegetative existence of some inmates
by forcing into their awareness the immediate past and future". "Group Therapy",
op. cit. In discussing one of McCorkle's articles, a reviewer made the following signi-
ficant statement: "The future of group therapy in correctional work is bright because
it offers help to a greater number of individuals and permits the release of pent-up
hostility and aggression which, among the more aggressive groups, frequently breaks
out into open conflict". Samuel R. Kesselman, "Review of Contemporary Correction",
International Journal of Group Psychotherapy, 2 (April 1952), pp. 194-197. See, also,
Joseph Abrahams and Lloyd W. McCorkle, "Group Psychotherapy of Military
Offenders", American Journal of Sociology, 51 (March 1946), pp. 455-464.
a war between "the individual" and "society." One variety of

this idea in criminology is that the individual, because of an
instinctive something in him, "breaks through" the restrictions of
society and follows delinquent patterns. For reformation, the
something in him must be modified. 21 Another variety, possibly
the one used in guided group interaction, is that the criminal's
character makeup is egocentric rather than altruistic-he thinks
in terms of "I" rather than "we" and, consequently, follows
delinquent patterns. By guided group interaction this indivi-
dualistic makeup, which is considered as being in opposition to
the cooperative spirit of "society" and "group living," purportedly
is removed by the therapist and fellow inmates who will not let
the individual "get away with" the expression of his egocentric
character. 22 But, according to the group relations principle,
reformation depends upon incorporation into the criminal's
personality of the attitudes, motives and rationalizations charac-
teristic of the social groups which obey the laws. These values
must replace the values he has learned in other groups. In other
words, the problem is not one of "individual versus society" but
instead of one kind of values (criminal) versus other kinds of
values (anticriminal).23 Similarly, according to the group re-
lations viewpoint, "finding personal satisfaction in conforming to
social rules" would mean conformity to non-criminal or anticriminal
values or "rules," since everyone, criminal or not, finds satisfaction
21 Cj. Irving Schulman, "The Dynamics of Certain Reactions of Delinquents to

Group Psychotherapy", International Journal of Group Psychotherapy, 2 (October 1952),
pp. 334--343. Schulman says, "The term delinquent character disorder ... refers to
character development which leads to repeated aggressive infringements upon values
of the community from which the transgressor derives gratification. In such behaviour
the individual is responding to the demands of his instinctual drives with little respect
for reality".
22 The policy of imprisonment itself is in part based on the same notion. It often is
assumed that by incarcerating criminals we are "educating them to accept the
restrictions of society", i.e. teaching them as individuals that they cannot "get away
with" committing crimes.
23 That this point has been recognized in correctional group therapy may be
inferred from the following statement regarding boys in the New York Training
School: "They were convinced that everyone is dishonest; even the police, the
government officials and the judges took bribes. Thus, they sought to convince
themselves that they were not different from anyone else ... They needed persons with
socially acceptable standards and conduct with whom they could form a strong,
healthy relationship and with whom they could identify. Group therapy can be an
effective means of providing this need". Charles Gersten, "An Experimental Evalu-
ation of Group Therapy with Juvenile Delinquents", International Journal of Group
Psychotherapy I (November 1951), pp. 311-318.
in conforming to social rules. What we attempt to correct in our

prisons is not non-conformity or lack of satisfaction in conformity
to "social rules." 24 Instead, we are trying to modifY conformity
and satisfaction in conformity to norms and values of which we,
the law makers, do not approve. 25 If the orientation in guided
group interaction were primarily that of group relations, it would
be concerned with providing informal experience in conformity
to legal norms, not with individual rebellion against restrictions
and rules generally. 26
Gaining experience as a law-abiding person. A fourth implicit

expectation is consistent with the group relations principle. It is
expected that each participant in the group sessions gains experi-
ences in the role of a law-abiding person and that this experience
will carryover to the life outside the session and outside the
prison. Here, the reformative effect of the session is considered as
operating not on the inmate as his criminal behaviour and atti-
tudes are analyzed and denounced but, instead, on the inmate as
he does the analyzing or denouncing. As the participant attempts
to change the behaviour of others he necessarily recognizes that
behaviour as undesirable. And he must identify with and "take
the side" of anticriminal groups when he condemns law violation
and law violators. He becomes a reformer, rather than a re-
formee, and in denouncing the criminality of others he denounces
his own criminality.27 Possibly, the entire group will become
"anticriminal," thus supporting the new anticriminal views of
individual participants. Status in the group may be assigned
according to the degree of "pro-reform" behaviour which is ex-
hibited. If this occurs, there has been a real modification of the
social relations of each participant, and the group itself has be-
come an effective medium of change. The personal satisfaction
24 The inmates of the Federal Reformatory at Chillicothe have been found to be
more "rule-obeying" than college males or Catholic students. Robert A. Harper,
"Is Conformity a General or a Specific Behaviour Trait?" American Sociological Review,
12 (February 1947), pp. 81-86.
26 "To earn the right to belong he [the youth] will adopt whatever code of behaviour
the gang or group prescribes, regardless of how much it conflicts with society's
standards and demands." John R. Ellingston, Protecting Our Children from Criminal
Careers, New York: Prentice-Hall, 1948, p. 35.
26 Cj. Void, op. cit.
2? For a description of a group session in which this point is dramatically demon-
strated, see Gisela Konopka, "The Group Worker's Role in an Institution for Juvenile
Delinquents", Federal Probation, 15 (June 1951), pp. 15-23. See also Chapter IX, below.
which a participant now obtains from denouncing criminal be-

haviour and values actually is satisfaction in conforming to anti-
criminal social norms.
We may conclude that to some extent, perhaps almost acci-
dentally, guided group interaction is working for a transfer of the
inmate's allegiance from criminal to law-abiding groups and is,
thus, recognizing the importance of group relations and associ-
ations to reformation and criminality. However, while both-
guided group interaction and, presumably, other correctional
group therapy programmes might quite incidentally give in-
mates experience in the role of a law-abiding person, they are
severely limited in achieving permanent reformation because they
are also expected to provide therapy of the kind implied by the
clinical principle.
One serious limitation on group therapy as a programme for
applying the group relations principle is the rather sharp dis-
tinction, based on the clinical principle, which is drawn between
the therapy sessions and the remainder of the activities in which
inmates participate. Group therapy sessions rarely deal with
"natural" groups in the prison, i.e. groups of men united by
common interests and attitudes. Instead, inmates are individually
selected for participation on the basis of some trait or character-
istic such as offense, psychiatric diagnosis, or disciplinary reports,
and the group sessions are sharply defined as something distinct
and different from ordinary prison life. 28 Consequently, attitudes
and values acquired from taking the role oflaw-abiding person in
the group sessions may receive little support in the general prison
community. The inmate's friends receive him as a fellow criminal
and may ridicule his newly acquired "Square John" attitudes;
the guards and other officials force him to assume the role of a law-
breaker and may distrust his "reformed" demeanour. According
to the group relations principle, if the experience in the role of a
law-abiding person is to have a lasting reformative effect, it must be
supported by the social organization in which the reformee lives. 29
28 It has been explicitly recommended, in connection with group therapy for

psychiatric patients, that the participants should not communicate with each other
outside the therapy sessions. Hendrick Lindt and Max A. Sherman, "'Social Incognito'
in Analytically Oriented Group Psychotherapy", International Journal of Group
Psychotherapy, 2 (july, 1952), pp. 209-220.
2. Cj. Edwin A. Fleishman, A Stutfy of the Leadership Role of the Foreman in an Industrial
Another serious limitation is the individualistic emphasis upon

catharsis and ventilation of suppressed hostilities, which actually
amounts to sanction of behaviour in the role of a criminal, in
direct contrast to the group relations sanction of behaviour in the
role of a law-abiding person. The published excerpts of many
group therapy sessions, for example, reveal that the inmate is
given experiences which, if reenacted in free society, would result
in his immediate investigation for slander, blasphemy and dis-
turbing the peace! Similarly, when therapy is even in part based
upon the clinical principle the inmate participants must be made
to understand that the sessions are for the purpose of changing
them, not the network of social relations in which they live or will
live. That is, they must accept the proposition that criminality
consists of individual deviation from "social norms," rather than
group-supported conformity to procriminal values. If this clinical
notion is to be the basis of a therapy programme, then that
programme, to be most effective, probably should deal with in-
mates as individuals rather than as group members. But when the
therapy programme attempts to deal with inmates as group
members and, at the same time, retains the assumptions of a
clinical treatment programme, it loses its efficiency.
Finally, group therapy is limited in its achievements by the
fact that participants must learn that much behaviour which is
permitted in the group sessions will be subject to the usual
institutional disciplinary action if it is repeated outside the ses-
sion,3o a condition which is not conducive to the carryover of
even law-abiding experiences. This is not exclusively a conse-
quence of basing the group therapy on the clinical principle, for
custodial considerations also are involved. However, the be-
haviour and attitudes which result in disciplinary action are
Situation, Columbus: Personnel Research Board, Ohio State University, 1951. In

this study, foremen's leadership attitudes and behaviour were significantly changed
in a "human relations training course", but when the men were returned to the factory
situation and the old group relations they reverted to their old attitudes and behaviour.
The author's conclusion is very significant for group therapy: "Our results suggest
that the foremen may learn different attitudes for each situation. The attitude that is
'right' in the training situation may be different from the one that 'pays off' in the
industrial environment".
30 "Any behaviour the boys engage in after the therapy hour must be handled in
the ordinary fashion". JamesJ. Thorpe and Bernard Smith, "Operational Sequence
in Group Therapy with Young Offenders", International Journal of Group Psychotherapy,
2 (January 1952), pp. 24--34.
those brought out by the clinical aspects, not the group relations
aspects, of group therapy.31
The effect of the limitations imposed on prison group therapy
by consideration for the clinical principle is emphasized by com-
paring these sessions with the so-called "repressive-inspirational" 32
type of group therapy used quite successfully by Alcoholics
Anonymous. In this programme, the group relations principle is
explicitly utilized for producing change or "reform" in the
participants and there is little, if any, conflict with the assumptions
of clinical therapy. When he joins Alcoholics Anonymous, the
alcoholic immediately acquires an intimate, primary-type
membership in a "network of personal relations that are made up
of obligations, friendships and other personal influences."33 The
"offender" becomes a member of a group composed of persons
who have "reformed" and who try to induce him to do the same.
Quite incidentally, his social isolation or, possibly, belligerence is
reduced and he may be given rather casual counsel regarding
psychological problems. But the major emphasis is placed on a
continuing effort by the group to redefine his attitudes in respect
to drinking. If the group is successful, the alcoholic identifies with
and becomes integrated in a sympathetic but nevertheless strong-
ly "antialcohol" group in which status is assigned according to
the "antialcohol" behaviour which is exhibited. His attitudes and
motives about the use of alcohol are replaced by new attitudes
and motives, the attitudes and motives of the therapy group.
Then, unlike the situation in correctional group therapy, these
new attitudes and motives are reinforced, not instigated, as the
reformee becomes the reformer: "The member sees in his pro-
spective convert himself as he once was and, by teaching the
others, becomes his own therapist." 34 In this programme, then,
31 "Until individuals in the group learn to differentiate between tbeir role in the
session and other roles in the institution, problems and difficulties can arise ... Group
therapy may bring hostilities to the surface that had previously been suppressed.
Feelings may overflow in other relationships and individuals get into difficulties of
violating institutional rules and regulations". Bixby and McCorkle, op. cit.
32 Giles Thomas, "Group Psychotherapy: A Review of the Recent Literature",
Psychosomatic Medicine, (April 1943), pp. 166-180.
33 Freed Bales, "Types of Social Structure as Factors in 'Cures' for Alcohol Addic-
tion", Applied Anthropology, I (April-June 1942), pp. 1-13.
34 Ibid. See, also, Willis H. McCann and Albert A. Almada, "Round-Table
Psychotherapy: A Technique in Group Psychotherapy", Journal of Consulting Psycho-
logy, 14 (December 1950), pp. 421-435.
there is explicit acknowledgment of the notion that attitudes,

motives and overt behaviour are the product of social interaction.
There also is unwavering acceptance of the principle that if a
reformee is to abandon attitudes, motives and overt behaviour
patterns which are considered undesirable he must be assimilated
into a group which presents him with and supports new attitudes,
motives and behaviour patterns.
As an instrument for reforming prisoners, group therapy
probably will become more effective than it now is when specific
techniques are invented for developing and supporting anti-
criminal values and behaviour patterns in intimate, personal,
informal groups comparable to those used to reform alcoholics.
The development of these techniques probably will be concurrent
with recognition of the proposition that group therapy must be
based on a group theory of criminality and reformation, like the
theory of differential association, and not on an individualistic
theory. If the group relations principle is followed, group therapy
programmes will include entire prison populations and will
explicitly attempt to substitute a set of "pro-leform" social re-
lations for the "anti-reform" network of social relations which
now exists among inmates.
CHAPTER IX
DIFFERENTIAL ASSOCIATION AND

REHABILITATION
While the group relations principle of rehabilitation is generally

accepted by sociologists and social psychologists, there has been
no organized effort by sociologists-criminologists to experiment
with it or to base techniques of treatment on it. As indicated in
the previous chapter, sociologists have emphasized the idea that
they can make unique contributions to clinical diagnoses, and
they have advocated the development of a "clinical sociology"
which would enable us to improve these diagnoses. 1 But here an
impasse is reached, for if a case of criminality is attributed to the
individual's group relations, there is little that can be done
in the clinic to modify the diagnosed cause of the criminality.
Moreover, extra-clinical work with criminals and delinquents
ordinarily has merely extended the clinical principle to the
offender's community and has largely ignored the group-re-
lations principle. For example, in the "group work" of cor-
rectional agencies the emphasis usually is upon the role of the
group merely in satisfying the needs of an individual, so that the
criminal may be induced to join an "interest-activity" group,
such as a hiking club, on the assumption that membership in the
group will somehow enable him to overcome the defects or
tendencies considered conducive to delinquency. 2 Similarly,
Chapter VIII has shown that in correctional group therapy the
emphasis is on the use of a group to enable the individual to rid
himself of undesirable psychological disorders, not criminality.
Even in group-work programmes directed at entire groups, such
as delinquent gangs, emphasis often is on new and different
1 See Louis Wirth, "Clinical Sociology", American Journal of Sociology, 37 (July,
1931), pp. 49--66; and Saul D. Alinsky, "A Sociological Technique in Clinical
Criminology", Proceedings of the American Prison Association, 64 (1934), pp. 167-78.
2 See the discussion by Robert G. Hinckley and Lydia Hermann, Group Treatment
in Psycho-therapy, Minneapolis: University of Minnesota Press, 1951, pp. 8-11.
formal group activities rather than on new group attitudes and

values.
In the face of this situation, a set of rules was developed in 1955
as a guide to specific application of the differential association
theory and the group relations principle to the rehabilitation of
criminals. The rules were adapted from a more general statement
by Dorwin Cartwright, 3 but in the long run they come from the
sociological and social psychological literature on social move-
ments, crime prevention, group therapy, communications,
personality change and social change. They are designed to show
that there is a distinctive, non-clinical, body of theory which can
be used effectively by practitioners who would prevent crime and
change criminal and delinquent behaviour. Sutherland also had
this as a principal objective when he formulated his theory of
differential association. The set of rules is tentative, and is in-
tended to direct attention to areas where research and experi-
mentation should prove fruitful. Two underlying assumptions are
that small groups existing for the specific purpose of reforming
criminals can be set up by correctional workers and that criminals
can be induced to join them. The first five rules deal with the use
of anti-criminal groups as media of change, and the last rule
emphasizes, further, the possibility of a criminal group becoming
the target of change.
( 1) if criminals are to be changed, they must be assimilated into groups
which emphasize values conducive to law-abiding behaviour and, concurrent-
ly, alienatedfrom groups emphasizing values conducive to criminaliry. Since
our experience has been that the majoriry rif criminals experience great
difficulry in securing intimate contacts in ordinary groups, special groups
whose major common goal is the riformation rif criminals must be created.
This general rule, emphasized by Sutherland, was used by
Gersten, apparently with some success, in connection with a
group therapy programme in the New York Training School for
BoYS.4
( 2) The more relevant the common purpose of the group to the re-
formation of criminals, the greater will be its influence on the criminal
3 Dorwin Cartwright, "Achieving Change in People: Some Applications of Group
Dynamics Theory", Human Relations, 4 (1951), pp. 381-392.
4 Charles Gersten, "An Experimental Evaluation of Group Therapy with Juvenile
Delinquents", International Journal of Group Psychotherapy, I (November 1951), pp.
311-318.
REHABILITATION 141
members' attitudes and values. Just as a labour union exerts strong in-
fluence over its members' attitudes towards management but less influence
on their attitudes towards, say, Negroes, so a group organized for re-
creational or welfare purposes will have less success in influencing crimi-
nalistic attitudes and values than will one whose explicit purpose is to
change criminals. Interesting recreational activities, employment
possibilities and material assistance may serve effectively to at-
tract criminals away from pro-criminal groups temporarily and
may give the group some control over the criminals. But merely
inducing a criminal to join a group to satisfY his personal needs is
not enough. Probably the failure to recognize this, more than
anything else, was responsible for the failure of the efforts at
rehabilitation by the Cambridge-Somerville Youth Study
workers.5
( 3) The more cohesive the group, the greater the members' readiness to
influence others and the more relevant the problem of conformity to group
norms. The criminals who are to be reformed and the persons expected to
effect the change must, then, have a strong sense of belonging to one group;
between them there must be a genuine "we" feeling. The reformers,
consequently, should not be identifiable as correctional workers, probation
or parole officers, or social workers. This rule has been extensively
documented by Festinger and his co-workers. 6
( 4) Both reformers and those to be reformed must achieve status within
the group by exhibition if "pro-reform" or anti-criminal values and be-
haviour patterns. As a novitiate, the one to be reformed is likely to assign
status according to social position outside the group, and part of the re-
formation process consists of influencing him both to assign and to achieve
status on the basis of behaviour patterns relevant to reformation. If he
should assign status solely on the basis of social position in the
community, he is likely to be influenced only slightly by the
group. Even if he becomes better adjusted, socially and psycho-
logically, by association with members having high status in the
community, he is a therapeutic parasite and not actually a
• See Margaret G. Reilly and Robert A. Young, "Agency-initiated Treatment of
a Potentially Delinquent Boy", American Journal of Orthopsychiatry, 16 (October, 1946),
pp. 697-706; Edwin Powers, "An Experiment in Prevention of Delinquency", Annals
of the American Academy of Political and Social Science, 261 (January, 1949), pp. 77-88;
Edwin Powers and Helen L. Witmer, An Experiment in Preventing Delinquency - The
Cambridge-Somerville Youth Study, New York: Columbia University Press, 1951.
• L. Festinger et al., Theory and Experiment in Social Communication; Collected Papers,
Ann Arbor: Institute for Social Research, 1951.
member until he accepts the group's own system for assigning

status.
( 5) The most effective mechanism for exerting group pressure on
members will be found in groups so organized that criminals are induced
to join with noncriminals for the purpose qf changing other criminals.
A group in which criminal A joins with some noncriminals to change
criminal B is probably most effective in changing criminal A, not B;
in order to change criminal B, criminal A must necessarily share the values
of the anticriminal members. This process may be called "retroflexive
reformation" ; in attempting to reform others, the criminal almost
automatically accepts the relevant common purpose of the group,
identifies himself closely with other persons engaging in re-
formation, and assigns status on the basis of anticriminal be-
haviour. He becomes a genuine member of this group and at the
same time he is alienated from his previous pro-criminal groups.
This rule is used successfully by Alcoholics Anonymous to "cure"
alcoholism, it was used in the treatment of psychotics by McCann
and Almada, and its usefulness in criminology was demonstrated
by Knopka. 7 Ex-convicts were used in the Chicago Area Projects
which, generally, have been organized in accordance with this
rule, but the effect on the ex-convicts, either in their roles as
reformers or as objects of reform, appears not to have been
evaluated.
(6) When an entire group is the target qf change, as in a prison or
among delinquent gangs, strong pressure for change can be achieved by
convincing the members qf the need for a change. Rather than inducing
criminals to become members of pre-established anticriminal groups, the
problem here is to change anti-reform and pro-criminal subcultures so that
group leaders evolve from among those who show the most marked hospitality
to anticriminal values, attitudes and behaviour. Neither mere lectures,
sermons or exhortations by correctional workers nor mere redi-
rection of the activities of a group nor individual psychotherapy,
academic education, vocational training or counselling will neces-
sarily change a group's culture. If the subculture is not changed,
• Freed Bales, "Types of Social Structure as Factors in 'Cures' for Alcohol Ad-
diction", Applied Anthropology, I (April-June 1942), pp. 1-13; Willis H. McCann and
Albert A. Almada, "Round-Table Psychotherapy: A Technique in Group Psycho-
therapy", Journal of Consulting Psychology, 14 (December 1950), pp. 421--435; Gisela
Knopka, "The Group Worker's Role in an Institution for Juvenile Delinquents",
Federal Probation, 15 (June 1951), pp. 15-23.
REHABILITATION 143
the person to be reformed is likely to exhibit two sets of attitudes

and behaviours, one characteristic of the agency or person trying
to change him, the other of the subculture. 8 Changes in the sub-
culture probably can best be instigated by eliciting the cooperation
of the type of criminal who, in prisons, is considered a "right
guy."9 This principle was demonstrated in an experiment with
hospitalized drug addicts, whose essentially anti-reform culture
was changed, under the guise of group therapy, to a pro-reform
culture. 10 To some extent, the rule was used in the experimental
system of prison administration developed by Gill in the Massa-
chusetts State Prison Colony.l1
If one assumes, as we do, that the set of rules is consistent with
Sutherland's theory of differential association and that this
theory, in turn, is consistent with more general sociological and
social psychological theory, a test of the validity of the rules
would be a test of the more general formulations. Ideally, such a
test would involve treating the rules as a set of interrelated
hypotheses-careful observations would be made of the results
occuring from a programme rationally designed to utilize the
rules for changing criminals. To our knowledge, such a test has
not been made, although, as indicated, some of the rules were
used in programs long before they were stated in their current
form.12 As a "next best" test of the rules, we may study rehabili-
tation programmes which have used them, even if there was no
planned design for doing so. Such study amounts to an "ex post
facto experiment," by means of which the results of programmes
are analyzed as if experimental variables had been introduced at
the beginning.
One programme that has utilized the first five rules has been in
8 Edwin H. Fleishman, A Stuqy of the Leadership Role of the Foreman in an Industrial
Situation, Columbus, Personnel Research Board, Ohio State University, 1951.
9 See Hans Riemer, "Socialization in the Prison Community", Proceedings of the
American Prison Association, 67 (1937), pp. 151-55.
10 JamesJ. Thorpe and Bernard Smith, "Phases in Group Development in Treat-
ment of Drug Addicts", International Journal of Group Psychotherapy, 3 (January, 1953),
pp.66-78.
n Howard B. Gill, "The Norfolk Prison Colony of Massachusetts", Journal of
Criminal Law and Criminology, 22 (September 1937), pp. 389-95; see also Eric K. Clarke,
"Group Therapy in Rehabilitation", Federal Probation, 16 (December, 1952), pp. 28-
32.
12 See, however, Joseph A. Cook and Gilbert Geis, "Forum Anonymous: The
Techniques of Alcoholics Anonymous Applied to Prison Therapy", Journal of Social
Therapy, 3 (First Quarter, 1957), po. 9-13.
operation since 1958. 13 Analysis of this programme, reported

below, shows remarkable similarity to any programme which
could have been designed to implement the rules on an experi-
mental basis, and the results over the years can therefore be viewed
as at least a crude test of the rules, of Sutherland's theory, and of
more general sociological and social psychological theory re-
garding the relationship between behaviour and participation in
social relationships. It should be noted that since the rules are
interrelated, the parts of any programme attempting to implement
them or test them, even inadvertently, must necessarily be over-
lapping.
"Synanon," an organization of former drug addicts, and in-
cluding many ex-convicts, was founded in May, 1958, by a
member of Alcoholics Anonymous, with the assistance of an
alcoholic and a drug addict. Beginning in December, 1958, and
continuing until the present, Miss Rita Volkman, M.A., ob-
tained permission of the Synanon Board of Directors to make a
series of observations on the operations of the programme in
Synanon House, and to do a statistical analysis of background
material on some of the early residents. 14 We shall show how these
observations revealed the close relationship of the programme to
the first five rules, then we will report on the effectiveness of the
programme in accomplishing its primary goal-rehabilitation of
drug addicts and criminals. 15
The Subjects
Background data were available on only the first fifty-two
persons entering Synanon after July 1958. These records were
prepared by a resident who in July 1959 took it upon himself to
interview and compile the information. We have no way of
determining whether these fifty-two persons are representative of
13 The remainder of this chapter is taken from Rita Volkman and Donald R.
Cressey, "Differential Association and the Rehabilitation of Drug Addicts", American
Journal of Sociology, 69 (September, 1963), pp. 129-142. I am indebted to Miss Volk-
man for her permission to reprint this work here.
14 The Board at first was composed of the three original members. It is now made
up of the founder (an ex-alcoholic but a non-addict) and seven long-term residents
who have remained off drugs and who have demonstrated their strict loyalty to the
group and its principles.
15 See Rita Volkman, A Descriptive Case Stud] of S]nanon as a Primary Group Organi-
zation, Unpublished Master's thesis, Department of Education, University of Cali-
fornia, Los Angeles, 1961.
REHABILITATION 145
Table 1. AGE AND SEX
Age in years Males Females Total
N % N % N %
18-20 0 0 1 7 1 2
21-30 17 44 11 79 28 54
31-40 18 48 2 14 20 38
41-50 1 3 0 0 1 2
51-60 2 5 0 0 2 4
Total 38 100 14
100 52 100
Median ages: Males, 31.0; Females 27.5
all addicts. However, we believe they are similar to the 215

persons who have resided at Synanon for at least one month.
Age and sex are shown in Table I. Forty-four per cent of the
fifty-two were Protestant, thirty-five per cent Catholic, eight
per cent J ewish. 16 Racially, twenty-seven per cent were Negroid,
and there were no Mongoloids; nineteen per cent of the Cauca-
soids were of Mexican origin and thirteen per cent were ofItalian
origin. Educational attainment is shown in Table II. Although
the data on early family life are poor because the resident simply
asked "What was your family like?" it may be noted that only
five of the fifty-two indicated satisfaction with the home. Words
and phrases such as "tension," "arguing," "bickering," "violence,"
"lack of warmth," "went back and forth," and "nagged" were
common. 17
Table II. EDUCATIONAL ATTAINMENT
N %
Part grade school 1 2
Completed grade school 3 6
Part high school 24 46
Completed high school 11 21
Part college 13 25
Completed college o o
Total 52 100
16 In May, 1961, twenty per cent of the residents were Jewish.

17 Cj. Research Center for Human Relations, New York University, Family
Background as an Etiological Factor in Personality Predisposition to Heroin Addiction, New
York: Author, 1956.
The sporadic and tenuous occupational ties held by the group

are indicated in Table III. This table supports the notion that
addicts cannot maintain steady jobs because their addiction
interferes with the work routine, and it suggests, also, that these
members had few lasting peer group contacts or ties, at least so far
as work associations go. In view of their poor employment records,
it might be asked how the addicts supported their addictions,
which cost from $30 to $50 a day and sometimes ran to $100 a
Table III. LENGTH AND CONTINUITY OF EMPLOYMENT
Unsteady
No. years on Steady
(Discontinuous Total
one job (Continuous)
or sporadic)
Under I 36* 4 40
2-3 3 2 5
4-5 I 3 4
6 or over 2 I 3
42 10 52
* 67% of this category defined their work as "for short periods only".
day. Only four of the males reported that they obtained their
incomes by legitimate work alone; thirty (79%) were engaged in
illegitimate activities, with theft, burglary, armed robbery, shop-
lifting and pimping leading the list. One male and seven females
were supplied with either drugs or money by their mate or family,
and five of these females supplemented this source by prostitution
or other illegitimate work. Five of the women had no source of
income except that from illegitimate activities, and none of the
women supported themselves by legitimate work only. The
problem of rehabilitating a drug addict is also a problem of
rehabilitating a criminal.
Institutional histories and military service histories are consis-
tent with the work and educational histories, and they indicate
that the fifty-two members were not somehow inadvertently
selected as "easy" rehabilitation cases. The fifty-two had been in
and out of prisons, jails, and hospitals all over the United States.
Table IV indicates that ten males and one female had been con-
fined seven or more times; the mean number of confinements for
males was 5.5 and for females 3.9. The table seems to indicate
REHABILITATION 147
that for this group whatever value confinement in institutions

might have had, it clearly did not prevent further confinements.
In sum, the pre-Synanon experiences of the fifty-two residents
seems to indicate non-identification with pro-legal activities and
norms. Neither the home, the armed services, the occupational
world, schools, prisons nor hospitals served as links with the larger
and more socially acceptable community. This, then, is the kind
of "raw material" on which Synanon has been working.
Table IV. CONFINEMENTS IN INSTITUTIONS
Male Female Total*
Confinements N N N
1-3 9 6 IS
4-6 12 7 19
7-9 8 0 8
10-12 0 I I
13-15 2 0 2
Total
Confinements 166 59 225
* Three males indicated "numerous arrests", and four supplied no

information. These seven were not included in the tally.
THE PROGRAM
Admission. Not every addict who knocks on the door of Synanon

House is admitted. Nevertheless, the only admission criterion we
have been able to find is expressed willingness to submit one's self to
a group that hates drug addiction. Use of this criterion has un-
wittingly implemented the first of the five rules. The process of
assimilation and alienation called for in the rule begins the
moment an addict arrives at Synanon, and it continues through-
out his stay. The following are two leaders' comments on such
admission interviews.
(I) When a new guy comes in we want to find out whether a person has one
inkling of seriousness. Everybody who comes here is what we call a psycho-
pathic liar. We don't take them all, either. We work off the top spontaneously,
in terms offeeling. We use a sort of intuitive faculty. You know he's lying, but
you figure, "Well, maybe if you get a halfway positive feeling that he'll stay ... "
We ask him things like "What do you want from us"? "Don't you think you're
an idiot or insane"? Doesn't it sound insane for you to be running around the
alleys stealing money from others so's you can go and stick something up your
arm"? "Does this sound sane to you"? "Have you got family and friends
outside"? We might tell him to go do his business now and come back when
he's ready to do business with us. We tell him, "We don't need you". "You
need us". And if we figure he's only halfway with us, we'll chop offhis hair.
It's all in the attitude. It's got to be positive. We don't want their money.
But we may just tell him to bring back some dough next week. Ifhe pleads and
begs, - the money's not important. If he shows he really cares. If his attitude
is good. It's all in the attitude.
(2) Mostly, if people don't have a family outside, with no business to take
care of, they're ready to stay. They ain't going to have much time to think
about themselves otherwise ... Now, when he's got problems, when he's got
things outside, if he's got mickey mouse objections, like when you ask him
"How do you feel about staying here for a year"? and he's got to bargain with
you, like he needs to stay with his wife or his sick mother - then we tell him to
get lost. Ifhe can't listen to a few harsh words thrown at him, he's not ready.
Sometimes we yell at him, "You're a Goddamned liar"! If he's serious he'll
take it. He'll do anything if he's serious.
But each guy's different. If he sounds sincere, we're not so hard. If he's sick
of running the rat race out there, or afraid of going to the penitentiary, he's
ready to do anything. Then we let him right in ...
This admission process seems to serve two principal functions,

both consistent with the first rule. First, it forces the newcomer to
admit, at least on a verbal level, that he is willing to try to con-
form to the norms of the group, whose members will not tolerate
any liking for drugs or drug addicts. From the minute he enters
the door his expressed desire to join the group is tested by giving
him difficult orders-to have his hair cut off, to give up all his
money, to sever all family ties, to come back in ten days or even
thirty days. He is given expert help and explicit but simple criteria
for separating the "good guys" from the "bad guys" -the latter
use dope. Second, the admission process weeds out men and
women who simply want to lie down for a few days in order to
rest, in order to obtain free room and board, or in order to stay
out of the hands of the police. In the terms used by Lindesmith,
and also in the terms used at Synanon, the person must want to
give up drug addiction, not just the drug habit. IS This means that
he must at least say that he wants to quit using drugs once and for
all, in order to realize his potentials as an adult, and he must not
indicate that he merely wants a convenient place in which to go
through withdrawal distress so that he can be rid of his habit for a
short time because he has lost his connection, or for some other
18 Alfred R. Lindesmith, Opiate Addiction, Bloomington: Principia Press, 1947,
pp.44-66.
REHABILITATION 149
reason. He must be willing to give up all ambitions, desires, and

social interactions which might prevent the group from assimi-
lating his completely.
If he says he just wants to kick, he's no good. Out with him. Now we know
nine out often lie, but we don't care. We'd rather have him make an attempt
and lie and then get him in here for thirty days or so - then he might stick.
It takes months to decide to stay.
Most fish [newcomers] don't take us seriously. We know what they want,
out in front. A dope fiend wants dope, nothing else. All the rest is garbage.
We've even taken that ugly thing called money. This shows that they're
serious. Now this guy today was sincere. We told him we didn't want money.
We could see he would at least give the place a try. We have to find out if he's
sincere. Is he willing to have us cut off his curly locks? I imagine cutting his
hair off makes them take us seriously ...
Although it is impossible to say whether Synanon's selective
admission process inadvertently admits those addicts who are
most amenable to change, no addict has been refused admission
on the ground that his case is "hopeless" or "difficult" or that he is
"unreachable." On the contrary, before coming to Synanon,
twenty-nine of the fifty-two addicts had been on drugs for at least
ten years. Two of these were addicted for over forty years, and
had been in and out of institutions during that period. The
average length of time on drugs for the fifty-two was eleven years,
and fifty-six per cent reported less than one month as the longest
period of time voluntarily free of drugs after addiction and prior
to Synanon.
Indoctrination. In the admission process, and throughout his

residence, the addict discovers over and over again that the group
to which he is submitting is an anti-drug, anti-crime, anti-alcohol
group. At least a dozen times a day he hears someone tell him
that he can remain at Synanon only as long as he "stays clean,"
i.e., stays away from crime, alcohol, and drugs. This emphasis is
an unwitting implementation of the second rule.
Indoctrination makes clear the notion that Synanon exists in
order to keep addicts off drugs, not for purposes of recreation,
vocational education, etc. Within a week after admission, each
newcomer participates in an indoctrination session by a spontane-
ous group made up of four or five older members. Ordinarily, at
least one member of the Board of Directors is present, and he acts
as leader. The following are excepts from one such session with a
female addict. The organization's rules indicate the extreme

lengths to which it is necessary for the individual to subvert his
personal desires and ambitions to the anti-drug, anti-crime group.
Remember, we told you not to go outside by yourself. Whenever anybody
leaves this building they have to check in and out at the desk. For a while,
stay in the living room. Don't take showers alone or even go to the bathroom
alone, see. While you're kicking, somebody will be with you all the time. And
stay away from newcomers. You got nothing to talk to them about, except
street talk, and before you know it you'll be splitting [leaving] to take a fix
together. Stay out of the streets, mentally and physically, or get lost now.
No phone calls or letters for a while - if you get one, you'll read it in front of
us. We'll be monitoring all your phone calls for a while. You see, you got no
ties, no business out there any more. You don't need them. You never could
handle them before, so don't start thinking you can do it now. All you knew
how to do was shoot dope and go to prison.
You could never take care of your daughter before. You didn't know how
to be a mother. It's garbage. All a dope fiend knows how to do is shoot dope.
Forget it.
There are two easy illustrations of the anti-drug and anti-crime

nature of the group's subculture. First, there is a strong taboo
against what is called "street talk." Discussing how it feels to take
a fix, who one's connection was, where one took his shot, the
crimes one has committed, or who one associated with is severely
censured. One's best friend and confidant at Synanon might well
be the person that administers a tongue lashing for street talk,
and the person who calls your undesirable behaviour to the
attention of the entire group during a general meeting. Second, a
member must never, in any circumstances, identify with the
"code of the streets," which says that a criminal is supposed to
keep quiet about the criminal activities of other criminals. Even
calling an ordinary citizen "square" is likely to stimulate a spon-
taneous lecture, in heated and colorful terms, on the notion that
the people who are really square are those that go around as bums
sticking needles in their arms. A person who as a criminal has
learned to hate stool pigeons and finks with a passion must now
turn even his closest friend over to the authorities-the older
members of Synanon-if the friend shows any signs of noncon-
formity. Ifhe should find that a member is considering "sneaking
off to a fix somewhere," has kept pills, drugs, or an "outfit" with
him when he joined the organization, or even has violated rules
such as that prohibiting walking alone on the beach, he must by
Synanon's code relinquish his emotional ties with the violator and
REHABILITATION 151
expose the matter to another member or even to the total member-

ship at a general meeting. If he does not do so, more pressure is
put upon him than upon the violator, for he is expected to have
"known better." For perhaps the first time in his life he will be
censured for not "squealing" rather than for "squealing."19 He
must identify with the law and not with the criminal intent or act.
The sanctions enforcing this norm are severe, for its violation
threatens the very existence of the group. "Guilt by association"
is the rule. In several instances during a general meeting the
entire group spontaneously voted to "throw out" both a member
who had used chemicals and a member who had known of this
use but had not informed the group. Banishment from the group
is considered the worst possible punishment, for it is emphasized
over and over again that life in the streets "in your condition" can
only mean imprisonment or death.
Emphasis on the notion that the group's purpose is keeping
addicts off drugs is given in formal and informal sessions-called
"haircuts" or "pull ups" - , as well as in spontaneous denunci-
ations, and in denunciations at general meetings. The "synanon,"
discussed below, also serves this purpose. A "haircut" is a deliber-
ately contrived device for minimizing the importance of the
individual and maximizing the importance of the group, and for
defining the group's basic purpose-keeping addicts off drugs and
crime. The following is the response of a leader to the questions,
"What's a haircut? What's its purpose?"
When you are pointing out what a guy is doing. We do this through mecha-
nisms of exaggeration. We blow up an incident so he can really get a look at it.
The Coordinators [A Coordinator resembles an Officer of the Day.] and the
Board members and sometimes an old timer may sit in on it. We do this when
we see a person's attitude becoming negative in some area.
For a real haircut, I'll give you myself. I was in a tender trap. My girl split.
She called me on the job three days in a row. I made a date with her. We kept
the date and I stayed out all night with her. Now, shewas loaded [ using drugs].
I neglected-or I refused - to call the house. By doing this I ranked everybody.
You know doing something like that was no good. They were all concerned.
They sent three or four autos looking for me because I didn't come back from
work. You see, I was in Stage II.
X found me and he made me feel real lousy, because I knew he worked and
was concerned. Here he was out looking for me and he had to get up in the
morning.
19 See Lewis Yablonsky, "The Anti-Criminal Society: Synanon", Federal Probation,
26 (September, 1962), pp. 50-57; and Lewis Yablonsky, The Violent Gang, New York:
Macmillan, 1962, pp. 252-263.
Well, I called the house the next morning and came back. I got called in for
a haircut.
I sat down with three Board members in the office. They stopped everything
to give the haircut. That impressed me. Both Y and Z, they pointed out my
absurd and ridiculous behaviour by saying things like this - though I did not
get loaded, I associated with a broad I was emotionally involved with who was
using junk. I jeopardized my own existence by doing this. So they told me,
"Well, you fool, you might as well have shot dope by associating with a using
addict". I was given an ultimatum. If I called her again or got in touch with
her I would be thrown out.
("Why"?)
Because continued correspondence with a using dope fiend is a crime against
me - it hurts me. It was also pointed out how rank I was to people who are
concerned with me. I didn't seem to care about people who were trying to help
me. I'm inconsiderate to folks who've wiped my nose, fed me, clothed me.
I'm like a child, I guess. I bite the hand that feeds me.
To top that off, I had to call a general meeting and I told everybody in the
building what a jerk I was and I was sorry for acting like a little punk. I just
sort of tore myself down. Told everyone what a phony I had been. And then
the ridiculing questions began. Everybody started in. Like, "Where do you
get off doing that to us"? That kind of stuff. When I was getting the treatment
they asked me what I'd do - whether I would continue the relationship,
whether I'd cut it off, or if! really wanted to stay at Synanon and do something
about myself and my problem. But I made the decision before I even went in
that I'd stay, and cut the broad loose. I had enough time under my belt to
know enough to make that decision before I even came back to the house ...
Group cohesion. The daily programme at Synanon is also con-

sistent with the third rule stated above, and it appears to be an
unwitting attempt to implement that principle. In the daily
activities, cohesion is maximized by a "family" analogy and by
the fact that all but some "Third Stage" members live and work
together. The daily programme has been deliberately designed to
throw members into continuous mutual activity. In addition to
the free, unrestricted interaction in small groups called "syna-
nons," the members meet as a group at least twice each day.
After breakfast, someone is called upon to read the "Synanon
Philosophy," which is a kind of declaration of principles, the
day's work schedule is discussed, bits of gossip are publicly
shared, and the group or individual members are spontaneously
praised or scolded by older members. Following a morning of
work activities, members meet in the dining room after lunch to
discuss some concept or quotation. Stress is on participation and
expression; quotations are selected by Board members to provoke
controversy and examination of the meaning, or lack of meaning,
of words. Discussion sometimes continues informally during the
REHABILITATION 153
afternoon work period and in "synanons," which are held after

dinner (see below). In addition to these two types of session,
lectures and classes are held several times a week for all members
who feel a need for them, and they are conducted by any member
or outside speaker who will take on the responsibility. Topics have
included "semantics," "group dynamics," "meaning of truth,"
"Oedipus Complex." ,
Weekend recreational activities have emerged, and holidays,
wedding anniversaries, and birthdays are celebrated. Each
member is urged to "Be yourself," "Speak the truth," "Be honest,"
and this kind of action in an atmosphere that is informal and open
quickly gives participants a strong sense of "belonging." Since
many of the members have been homeless drifters, it is not sur-
prising to hear frequent repetition of some comment to the effect
that "This is the first home I ever had."
Also of direct relevance to the third rule is the voluntary charac-
ter of Synanon. Any member can walk out at any time; at night
the doors are locked against persons who might want to enter, but
not against persons who might want to leave. Many do leave. But
one condition holding addicts in the house once they have been
allowed to enter is a strong appeal to ideas such as "We have all
been in the shape you are now in," "Mike was on heroin for
twenty years and he's off." It is significant in this connection that
addicts who "kick" (go through withdrawal distress) at Synanon
universally report that the sickness is not as severe as it is in
involuntary organizations such as jails and mental hospitals. One
important variable here, we believe, is the practice of not giving
"kicking dope friends" special quarters. A newcomer "kicks" on a
davenport in the center of the large living room, not in a special
isolation room or quarantine room. Life goes on around him.
Although a member will be assigned to watch him, he soon learns
that his sickness is not important to men and women who have
themselves kicked the habit. In the living room, one or two
couples might be dancing, five or six people may be arguing, a
man may be practicing the guitar, and a girl may be ironing. The
kicking addict learns his lesson: these others have made it. This
subtle device is supplemented by explicit comments from various
members as they walk by or as they drop in to chat with him.
We have heard the following comments, and many similar ones,
made to new addicts lying sick from withdrawal. It should be

noted that none of the comments could reasonably have been
made by a rehabilitation official or a professional therapist.
"It's OK boy. We've all been through it before".
"For once you're with people like us. You've got everything to gain here
and nothing to lose".
"You think you're tough. Listen, we've got guys in here who could run
circles around you, so quit your bull".
"You're one of us now, so keep your eyes open, your mouth shut and try to
listen for a while. Maybe you'll learn a few things".
"Hang tough, baby. We won't let you die".
Status Ascription. The fourth rule is the most crucial of the set of
five, and it is in connection with this rule that Synanon seems
most effective. The house has an explicit program for distributing
status symbols to members in return for staying off the drug and,
later, for actually displaying anti-drug attitudes. The resident, no
longer restricted to the status of "inmate" or "patient" as in a
prison or hospital, can achieve any staff position in the status
hierarchy.
The Synanon experience is organized into a career of roles
which represent stages of graded competence, and at the end of
the career are roles which might later be used in the broader
community. Figure 1 shows the status system in terms of occu-
pational roles, each box signifying a stratum. Such cliques as exist
at Synanon tend to be among persons of the same stratum.
Significantly, obtaining these jobs of increased responsibility and
status is almost completely dependent upon one's attitudes to-
wards crime and the use of drugs. In order to obtain a job such as
Senior Coordinator, for example, the member must have de-
monstrated that he can remain free of drugs, crime and alcohol
for at least three to six months.
Equally important, he must show that he can function without
drugs in situations where he might have used drugs before he
came to Synanon. Since he is believed to have avoided positions
of responsibility by taking drugs, he must gradually take on
positions of responsibility without the use of them. Thus, he
cannot go up the status ladder unless his "attitudes" are right,
no matter what degree of skill he might have as a workman.
Evaluation is rather casual, but it is evaluation nevertheless-he
will not be given a decent job in the organization unless he
REHABILITATION 155
Figure I. DIVISION OF LABOR AND STRATIFICATION SYSTEM,

JUNE 1962
Graduates Board of Directors

(Work & live outside) (Policy makers)
III
Assistant directors Office manager

Attend school Project directors
Business manager Senior coordinators
Department chiefs Work outside
II
Hustling crew Nursery heads

Junior coordinators Office workers
Kitchen crew chief Service crew chief
I-BA
Automobile crew Laundry

Barber Library
Electricity Maintenance crew
Housecleaning Plumbing
Kitchen help Service crew
I-AB
Newcomers Non-workers
Sick kicking addicts
I-A
relinquishes the role of "con artist" and answers questions honest-

ly, expresses emotions freely, cooperates in group activities and
demonstrates leadership. In a letter to a public official in May,
1960, the founder explained the system as follows: 20
20 See Volkman, op. cit., pp. 90-96.
Continued residence [at Synanon], which we feel to be necessary to work

out the problem of inter-personal relationships which underly the addiction
symptom is based on adherence by the individual to standards of behaviour,
thinking and feeling acceptable to our culture. There is much work to be done
here, as we have no paid help, and each person must assume his share of the
burden. Increased levels of responsibility are sought and the experience of self
satisfaction comes with seeking and assuming these higher levels ...
An analogy with a family and the development of a child also

is used. Officially, every member is expected to go through three
"stages of growth," indicated by Roman numerals in Figure l.
Stage I has two phases, "infancy" and "adolescence." In the
"infancy" phase (I-A) the member behaves like an infant and is
treated as one-as he kicks the habit "cold turkey" (without the
aid of drugs) in the living room he is dependent on the others, and
he is supervised and watched at all times. When he is physically
and mentally able, he performs menial tasks such as dishwashing
and sweeping (I-AB) and then takes on more responsible po-
sitions (I-BA). In this "adolescence" phase he takes on responsi-
bility for maintenance work, participates actively in group
meetings, demonstrates a concern for "emotional growth," ming-
les with newcomers and visitors and takes on responsibilities for
dealing with them. In work activities, for example, he might drive
the group's delivery truck alone, watch over a sick addict, super-
vise the dishwashing or clean-up crews, or meet strangers at the
door.
Stage II is called the "young adult stage." Here, the member is
in a position to choose between making Synanon a "career,"
attending school, or going to work at least part-time. If he works
for Synanon, his position is complex and involves enforcing policy
over a wide range of members. In Stage III, "Adult," he moves
up to a policy-making position in the Board of Directors or moves
out of Synanon but returns with his friends and family for oc-
casional visits. He can apparently resist the urge to resort to drugs
in times of crisis without the direct help of Synanon .members.
One man described this stage by saying, "They go out, get jobs,
lose jobs, get married, get divorced, get married again, just like
everyone else." However, the group does maintain a degree of
control. Graduates are never supposed to cut off their ties with
their Synanon "family," and they are expected to return frequent-
ly to display themselves as "a dope fiend made good."
REHABILITATION 157
From Table V it is apparent that seniority in the form oflength

of residence (equivalent to the number of "clean" days) is an
important determinant of status. As time of residence increases,
responsibilities to the group-in the form of work and in the form
ofleadership-tend to increase. In June 1962, twenty-seven of the
105 members of Synanon were in Stage III. It should be noted
that while stage is associated with length of residence, advance-
ment through the stages is not automatic. The longer one lives at
Synanon, the "cleaner" he is, the more diffuse the roles he
performs, and the higher his status.
Table V. LENGTH OF RESIDENCE AND "STAGE" OF MEMBERS,
JUNE 1962
Length of Residence Stages

(in months) I II III N %
1-3 20 0 0 20 19
4-6 15 0 0 15 14
7-9 7 3 0 10 9
10-12 2 0 0 2 2
13-15 3 4 0 7 7
16-18 3 0 2 5 5
19-21 4 1 0 5 5
22-24 0 4 1 5 5
25 and over 0 12 24 36 34
54 24 27 105 100
It is also important to note that high status does not depend

entirely upon one's conduct within the house. Before he graduates
to Stage III a member must in some way be accorded an increase
in status by the legitimate outside community. This is further
insurance that status will be conferred for activities that are anti-
drug in character. In early 1960, the members began to take an
active part in legitimate community activities, mostly in the form
of lectures and discussion groups. Since Synanon's inception,
over 350 service groups, church groups, political groups, school
and college classes etc., have been addressed by speakers from
Synanon. Such speeches and discussions gain community support
for the organization, but they further function to give members a
feeling of being important enough to be honored by an invitation
to speak before community groups. Similarly, members are

proud of those individuals who have "made good" in the outside
community by becoming board members of the P.T.A., Sunday
School teachers, college students, and members of civic and
service organizations. Over thirty-five Synanon members are
now working full or part time in the community, holding a wide
range of unskilled (janitor, parking attendant), skilled (truck
driver, carpenter, electrician), white-collar (secretary, musician,
photographer), and executive (purchasing agent) posts.
Further, the legitimate status of the group has increasingly risen
during the last two years. Since the summer of 1960, an average
of 100 to 150 guests have attended open house meetings, and the
guests have included distinguished persons from all walks of
legitimate life. Famous psychiatrists, correctional workers,
businessmen, newspaper men and politicians have publicly
praised the work of the group. There have been requests for
Synanon houses and for Synanon groups from several communi-
ties, and Synanon projects are now being conducted at Terminal
Island Federal Prison and the Nevada State Prison. New houses
have been opened in Nevada and Connecticut. Recently, the
group has been featured in films, on television and radio shows,
and in national magazines. At least two books and a movie are
being written about it. Over five hundred citizens have formed an
organization called "Sponsors of Synanon." Even strong attacks
from some members of the local community and complicated
legal battles about zoning ordinances have served principally to
unite the group and maximize the esprit de corps.
The "synanon." Synanon got its name from an addict who
was trying to say "seminar." The term "Synanon" is used to refer
to the entire organization, but when it is spelled with a small "s"
it refers only to the meetings occurring in the evenings among
small groups of six to ten members. Each evening, all members
are assigned to such groups, and membership in the groups is
rotated so that one does not regularly interact with the same six
or ten persons. The announced aim of these meetings is to "trigger
feelings" and to allow what some members refer to as "a ca-
tharsis." The sessions are not "group therapy" in the usual sense,
for no trained therapist is present. Moreover, the emphasis is on
enforcing anti-criminal and anti-drug norms, as well as upon
REHABILITATION 159
emotional adjustment. These sessions, like the entire program,

constitute a system for implementation of the fifth rule for reha-
bilitation, although they were not designed to do so.
In the house, behaviour of all members is exposable to all
members. What a member is seen to do at the breakfast table, for
example, might well be scrutinized and discussed at his synanon
that evening. The synanon sessions differ from everyday honesty
by virtue of the fact that in these discussions one is expected to
insist on the truth as well as to tell the truth. Any weapon, such as
ridicule, cross-examination, or hostile attack is both permissible
and expected. The sessions seem to provide an atmosphere of
truth-seeking which is reflected in the rest of the social life within
the household, so that a simple question like "How are you?" is
likely to be answered by a five minute discourse in which the
respondent searches for the truth. The following discussion is from
a tape recording ofa synanon session held inJune 1961. It should
be noted that an "innocent" question about appearance, asked by
an older member who has become a non-criminal and a non-
addict, led to an opportunity to emphasize the importance of
loyalty to the anti-drug, anti-crime group.
"What are you doing about losing weight"?

"Why? Is that your business"?
"I asked you a question".
"I don't intend to answer it. It's not your business".
"Why do you want to lose weight"?
"I don't intend to answer it".
"Why"?
"Because it's an irrelevant and meaningless question. You know I had a
baby only three weeks ago, and you've been attacking me about my weight.
It's none of your business".
"Why did you call your doctor"?
"Why? Because I'm on a diet".
"What did he prescribe for you"?
"I don't know. I didn't ask him".
"What did you ask for"?
"I didn't. I don't know what he gave me".
"Come on now. What kind of pills are they"?
"I don't know. I'm not a chemist. Look the doctor knows I'm an addict.
He knows I live at Synanon. He knows a whole lot about me".
"Yeah, well, I heard you also talking to him on the phone, and you sounded
just like any other addict trying to cop a doctor out of pills".
"You're a God damned liar"!
"Yeah, well X was sitting right there. Look, does the doctor know and does
the Board know"?
"I spoke to Y [Board member]. It's all been verified".

"What did Y say"?
" I was talking to ... "
"What did Y say"?
"Well, will you wait just a minute"?
"What did Y say"?
"Well, let her talk".
"I don't want to hear no stories".
"I'm not telling stories".
"What did Y say"?
"That it was harmless. The doctor said he'd give me nothing that would
affect me. There's nothing in it. He knows it all. I told Y".
"Gh, you're all like a pack of wolves. You don't need to yell and scream at
her".
"Look, I heard her on the phone and the way she talked she was trying to
manipulate the doctor".
"Do you resent the fact that she's still acting like a dope fiend and she still
sounds like she's conning the doctor out ofsomething? She's dope fiend. Maybe
she can't talk to a doctor any differently".
"Look, I called the doctor today. He said I should call him if I need him.
He gave me vitamins and lots of other things".
"Now wait a minute. You called to find out ifyou could get some more pills".
"Besides, it's the attitude they heard over the phone. That's the main thing"
"Yeah, well they probably projected it onto me".
"Then how come you don't like anyone listening to your phone calls"?
"Are you feeling guilty"?
"Who said"?
"Me. That's who. You even got sore when you found out X and me heard
you on the phone, didn't you? You didn't like that at all, did you"?
"Is that so"?
(Silence)
"I don't think her old man wants her back".
"Well, who would? An old fat slob like that".
"Sure, that's probably why she's thinking of leaving all the time and
ordering pills".
"Sure".
(Silence)
"My appearence is none of your business".
"Everything here is our business".
"Look, when a woman has a baby you can't understand she can't go back
to normal weight in a day".
"Now you look. We're really not interested in your weight problem now.
Not really. We just want to know why you've got to have pills to solve the
problem. We're going to talk about that if we want to. That's what we're
here for".
"Look, something's bugging you. We all know that. I even noticed it in
your attitude towards me".
"Yeah, I don't care about those pills. I want to know how you're feeling.
What's behind all this. Something's wrong. What is it"?
(silence)
"Have you asked your old man if you could come home yet"?
(solftly) "Yes".
REHABILITATION 161
"What did he say"?

(softly) "He asked me how I felt. Wanted to know why I felt I was ready to
come home ... "
(silence)
(softly) "I did it out of anger. I wasn't very happy. (pause) A day before I
tried [telephoning him], and he wasn't there. (pause) Just this funny feeling
about my husband being there and me here. My other kid's there and this
one's here. (pause) A mixed up family".
"Why do you want to stay then? Do you want to be here"?
"No. I don't want to be here. That's exactly why I'm staying. I need to stay
till I'm ready".
"Look, you've got to cut them loose for a while. You may not be ready for
the rest of your life. You may not ever be able to be with those people".
(tears)
"I know ... "
After the synanon sessions the house is always noisy and lively.
We have seen members sulk, cry, shout, and threaten to leave the
group as a result of conversation in the synanon. The following
comments, everyone of which represents the expression of a pro-
reform attitude by the speaker, were heard after one session. It is
our hypothesis that such expressions are the important ones, for
they indicate that the speaker has become a reformer and, thus,
is reinforcing his own pro-reform attitudes every time he tries to
comfort or reform another.
"Were they hard on you"?

"I really let him have it tonight".
"I couldn't get to her. She's so damned blocked she couldn't even hear what
I was trying to tell her".
"Hang tough, man; it gets easier".
"One of these days he'll drop those defenses of his and start getting honest".
"Don't leave. We all love you and want you t4fget well".
At Synanon, disassociating with former friends, avoiding street

talk and becoming disloyal to criminals are emphasized at the
same time that loyalty to non-criminals, telling the truth to
authority figures, and legitimate work are stressed. We have no
direct evidence that hair-cuts, synanons, and both formal and
spontaneous denunciations of street talk and the code of the
streets have important rehabilitative effects on the actor, as well
as (or, perhaps even "rather than") on the victim. It seems rather
apparent, however, that one's own behaviour must be dra-
matically influenced when he acts in the role of a moral police-
man and "takes apart" another member.
The Results
Of the fifty-two residents described earlier, four are "graduates"
of Synanon, are living honest lives in the community, and are not
using alcohol or drugs. Twenty-three (44.2 %) are still in residence
and are not engaged in crime and are not using alcohol or drugs.
Two of these are on the Board of Directors and eleven are
working part or full-time. The remaining twenty-five left Synanon
against the advice of the Board and the older members. Infor-
mation regarding the longest period of voluntary abstinence from
drugs after the onset of addiction but prior to entering Synanon
was obtained on forty-eight of the fifty-two persons. Eleven re-
ported that they were "never" clean, six said they were con-
tinuously clean for less than one week, ten were continuously
clean for less than one month. A total of thirty-nine (81 %) said
they had been continuously clean for less than six months, and
only two had been clean for as long as a one-year period. Twenty-
seven (52%) of the fifty-two residents have now abstained for at
least six months; twelve of these have been clean for at least two
years and two have been off drugs continually for over three
years.
Between May, 1958 (when Synanon started), and May 1961,
263 persons were admitted or readmitted to Synanon. Of these,
190 (72%) left Synanon against the advice of the Board of
Directors and the older members. Significantly, 59 per cent of all
drop-outs occurred within the first month of residence, 90 per cent
within the first three months. Synanon is not adverse to giving a
person a second chance, or even a third or fourth chance: of the
190 persons dropping out, eighty-three (44%) were persons who
had been readmitted. Personnel who were readmitted behaved in
general like first admissions so far as dropping out is concerned;
sixty-four per cent of their drop-outs occurred within the first
month, ninety-three per cent within the first three months after
readmission.
Of all the Synanon enrollees up to August, 1962, 108 out of
372 (29 %) are known to be off drugs. More significantly, of the
215 persons who have remained at Synanon for at least one month,
103 (48%) are still offdrugs; of the 143 who have remained for at
least three months, 95 (66 %) are still non-users; of the 87 who
have remained at least seven months, 75 (86%) are non-users.
REHABILITATION 163
These statistics seem to us to be of most relevance, for they indi-

cate that once an addict actually becomes a member of the anti-
drug community (as indicated by three to six months of partici-
pation), the probability that he will leave and revert to the use of
drugs is low.
Conclusions
Synanon's leaders do not claim to "cure" drug addicts. They
are prone to measure success by pointing to the fact that the
organization now includes the membership of forty-five persons
who each were heroin addicts for at least ten years. Two of these
were addicted for more than thirty years, and spent those thirty
years going in and out of prisons, jails, the U. S. Public Service
Hospital, and similar institutions. The leaders have rather inad-
vertently used a theory of rehabilitation that implies that it is as
ridiculous to try to "cure" a man of drug addiction as it is to try
to "cure" him of sexual intercourse. A man can be helped to stay
away from drugs, however, and this seems to be the contribution
Synanon is making. In this regard, its "success" rate is higher
than that of those institutions officially designated by society as
places for the confinement and "reform" of drug addicts. Such a
comparison is not fair, however, both because it is not known
whether the subjects in Synanon are comparable to those confined
in institutions and because many official institutions do not
concentrate on trying to keep addicts off drugs, being content to
withdraw the drug, build up the addicts physically, strengthen
vocational skills, and eliminate gaps in educational backgrounds. 21
We cannot be certain that it is the group relationships at
Synanon, rather than something else, that is keeping addicts away
from crime and drugs. However, both the times at which drop-
outs occur and the increasing anti-drug attitudes displayed with
increasing length of residence tend to substantiate Sutherland's
theory of differential association and our notion that modifying
social relationships is an effective supplement to the clinical
handling of convicted criminals. Drug addiction is, in fact, a
severe test of Sutherland's sociological theory and our five
21 CJ. Harrison M. Trice, "Alcoholism: Group Factors in Etiology and Therapy",
Human Organization, IS (Summer, 1956), pp. 33-40. See also Donald R. Cressey, "The
Nature and Effectiveness of Correctional Techniques", Law and Contemporary Problems,
23 (Fall, 1958), pp. 754-771.
sociological rules for rehabilitation, because addicts have the

double problem of criminality and the drug habit. The statistics
on drop-outs suggest that the group relations method of reha-
bilitation does not begin to have its effects until newcomers are-
truly integrated into the anti-drug, anti-crime group that is
Synanon. Despite the defects in the theory of differential associ-
ation, reported in earlier chapters, it may be concluded that the
theory remains as a valid guide for those who would rehabilitate
criminals, as well as for those who whould explain delinquent and
criminal behaviour and the distribution of crime and delinquency
rates in society.
INDEX OF NAMES
A E
Abrahams, Joseph, 132 Eissler, K. R., 104
Adler, M.J., 4 Ellingston,John R., 134
Agge, Ivar, 82 Elliott, Mabel A., 25, 26, 28, 83
Alexander, Franz, 92, 99, 102
Almada, Albert A., 137, 142 F
Axelrad, Sidney, 56 Festinger, L., 141
Fidler, Jay W., 132
B Fischer,John,65
Bales, Freed, 137, 142 Fisher, R. A., 75
Balistrieri, James, 65 Fleishman, Edwin A., 135, 143
Ball,John C., 85, 87, 88, ll8 Fointsky, 30
Barnes, Harry Elmer, 26, 82-84,101,123 Foote, Nelson N., 95, 96, 105
Barron, Milton L., 28, 84 Fox, Vernon, 52, 57
Beals, Ralph L., 86 Frank, Phillipp, 76
Beattie, Ronald H., 43 Freedman, Ronald, 74
Beck, Bertram M., 24 Fuller, Justin K., 132
Berg, Charles, 125 Fuller, Richard C., 22
Bixby, F. Lovell, 127, 131, 137
Boalt, Gunnar, 82 G
Bracton,99 Gault, Robert H., 99, 103
Bradley, Harold, 64 Geis, Gilbert, 143
Bromberg, Walter, 94, 95 Gerle, Bo, 82
Broom, Leonard, 69, 82 Gersten, Charles, 133, 140
Bruce, A. A., 46 Gill, Howard B., 29, 82, 84, 143
Burgess, E. W., 4, 46 Gillin, John L., 31
Gittler, Joseph, 25, 81
C Glaser, Daniel, 26, 46, 75, 76, 77,83,86,
Caldwell, Robert G., 25, 26, 27, 62, 63, 87,88,89, 119
81,85,88,89 Glover, Edward, 95
Cartwright, Dorwin, 126, 140 Glueck, Eleanor T., 124
Cavan, Ruth S., 25, 26 Glueck, Sheldon, 25-28, 40, 82, 84-87,
Clarke, Eric K., 143 89, 124
Cleckley, Hervey M., 94, 95 Gold, David, 74
Clinard, Marshall B., 25-29, 81-85, Green, Arnold W., 95
87-89, ll9, 128 Gross, Llewellyn, 76
Cloward, Richard A., 68, 82 Guerry, A. M., 30
Cohen, Albert K., 40, 69, 70,84, 86
Cook,Joseph A., 143 H
Cottrell, Leonard S., 69, 82 Hacker, E., 54
Cressey, Donald R., 34, 60, 62, 67, 72, Hakeem, Michael, 86
77,82,84,86,87,89,91, 109, 128, Hall, Jerome, 92, 94, 102
144, 163 Hamel,30
Hangren, Richard F., 77
D Hardin, Garrett, 74, 75, 78
Dahlberg, Gunnar, 55 Harno, A. J., 46
Darwin, Charles, 69, 72-76 Harper, Robert A., 134
Dewey, John, 126 Hartung, Frank E., 68, 86
Dorcus, Roy M., 92 Hawley, Amos H., 74
Dunham, H. Warren, 25 Hayner, Norman S., 59, 64
Durkheim. Emile, 71, 72 Healy, William, 38, 101
166 INDEX OF NAMES
Hermann, Lydia, 139 Malamud, William, 92, 93

Herzog, Elisabeth B., 42, 45 Mannheim, Ernest, 51
Heuman, Maths, 82 Mannheim, Hermann, 52, 55
Hinckley, Robert G., 139 Martin, C. E., 35, 65
Hoyek, Camille F., 103 Matza, David, 89, 119
Hunt,J. Mc V., 92, 94 Mead, George Herbert, 81, 91, 104
Mendel, Gregor, 75
J Menninger, William C., 129
Janson, Carl-Gunnar, 82 Merton, Robert K., 28, 31, 46, 68, 69, 82
Jeffery, Clarence R., 27, 28, 29, 81, 83, Meyer, Gerald Dennis, 73
86 Michael, J., 4
Johnson, Guy B., 56 Mills, C. Wright, 96, 97, 100, 118
Joly, Henri, 30, 60 Miner, Horace M., 74
Jones, Howard, 82 Moberg, David 0., 52
Morris, Richard T., 86
K Moses, Earl R., 58
Karpman, Benjamin, 95, 97, 101, 103
Keedy, E. R., 93- N
Kephart, William H., 57 Neumeyer, Martin H., 27, 84, 85
Kesselman, Samuel R., 132 Noyes, Arthur P., 97
Kinberg, Olaf, 82, 84
Kinsey, A. C., 35, 65 o
Kirkpatrick, M. E., 102 Ohlin, Lloyd E., 68
Knauer, Mary, 25
Konopka, Gisela, 134, 142 P
Kobrin, Solomon, 86 Park, Robert, 4
Korn, Richard R., 27, 83, 85, 89 Pavlov, Ivan Petrovich, 98
Kroeber, A. L., 86 Peterson, Virgil W., 34
Pollak, Otto, 53
L Pomeroy, W. B., 35, 65
Landecker, Werner S., 74 Powers, Edwin, 141
Lander, Bernard, 58 Prins, 30
Lane, Robert E., 25, 26, 84, 87, 119
Lazarsfeld, Paul F., 35 Q
Leader, Arthur L., 27, 82, 84, 85, 89 Quetelet, A., 30
Lemert, Edwin M., 56, 83, 119
Levy, Ruth Jacobs, 124 R
Lichtenstein, Perry M., 100 Radzinowicz, L., 54, 95
Lindesmith, Alfred R., 4, 70,84,86,97, Reckless, Walter C., 25, 26, 28, 39, 63,
114, 148 85,87,91
Lindner, Robert M., 107 Reilly, Margaret G., 141
Lindt, Hendrick, 135 Rengby, Sven, 82
Liszt,30 Riemer, Hans, 4, 143
Locke, Harvey, 4 Robinson, William S., 41
Lombroso, G., 61, 62 Rosberg, Judy, 56
Lorand, Sander, 93 Rose, Arnold, 90
Lottier, Stuart, 36, 62, 84 Rosenberg, Morris, 35
Lowie, Robert H., 114
Lowrey, Lawson G., 94 S
Lukas, EdwinJ., 124 Schaffer, G. Wilson, 92
Lunt, Paul S., 64 Schneider, Louis, 102
Schrag, Clarence, 77, 82, 87
M Schuessler, Karl, 70, 84, 86
McCann, Willis H., 137, 142 Schulman, Irving, 133
McCorkle, Lloyd W., 27, 83, 85, 89,127, Seagle, William, 22
128, 130, 131, 132, 137 Seeman, Melvin, 86
McKay, Henry, 4,21,70,119 Segerstedt, Torgny, 82
McKeown,James E., 57 Sellin, Thorsten, 4,31,34,35,52,56,82
INDEX OF NAMES 167
Shaw, Clifford R., 102 V

Shaw, Van B., 64 Valjean, jean, 116
Sherman, Max A., 135 Van Vechten, C. C., 60
Short, James F. Jr., 25, 27, 75, 82, 86, Veblen, Thorsten, 102
87,89,119 Volakakis, Joan,S 7
Shulman, Harry M., 26 Void, George B., 25, 83, 84, 85, 86, 87,
Simmel, Ernst, 103 127, 134
Slavson, S. R., 129, 132 Volkman, Rita, 144, 155
Sletto, R. F., 55
Smith, Bernard, 136, 143 W
Staub,Hugo, 92,99, 102 Wallace, Alfred R., 74
Stekel, Wilhelm, 100 Wallerstein,J. S., 36,102
Stouffer, Samuel A., 35 Warner, William Lloyd, 64
Strauss, A. L., 97, 114 Wattenburg, William W., 65
Sutherland, Edwin H., 3-16, 11, 12,20, Weinberg, S. Kirson, 27, 81, 85, 89
21, 23-28, 31, 33, 34, 37, 38, 43, 46, Wertham, Fredric, 94
60, 62, 66, 68-75, 77, 78, 81, 84-90, Wheeler, Stantonj., 119
101, 108, 109, 115, 124, 126, 127, 128, White, William A., 97
140, 143, 144, 163 Wiers, Paul, 62
Sweetser, Lowell, 4 Wilkins, Leslie T., 52
Sykes, Gresham, 89, 119 Williams, Jack D., 64
Wines, Fredrick H., 39
T Wirth, Louis, 139
Taft, Donald R., 26, 28, 31,60,83, 127 Witmer, Helen L., 141
Tappan, Paul W., 124, 130 Wood, Arthur Lewis, 58
Tarde, Gabriel, 30 Wright, S., 75
Teeters, Negley K., 26, 82, 83, 84, 101, Wyle, C. J., 36
123
Thomas, W. I., 4 Y
Thomas, Giles, 137 Yablonsky, Lewis, 151
Thorpe,JamesJ., 136, 143 Yoke, Helen L., 62
Thurnwald, Richard, 86 Young, Pauline V., 61, 141
Tiebout, H. M., 102
Toby, Jackson, 55 Z
Tolman, Frank L., 30 Zilboorg, Gregory, 93
Trice, Harrison M., 25, 27, 85, 163
Turner,James D., 56
Turner,j. W. C., 95
• MART/NUS NIIHOFF - PUBLISHER - THE HAGUE
DONALD R. CRESSEY, PH. D.
Professor of Socio/08Y
University of California. Santa Barbara
DELINQUENCY, CRIME
AND
This is a book about the late Edwin H. Sutherland's criminological
theory, the theory of "differential association." It is intended primarily
for non-American readers, who on the whole are not as familiar with
the theory as are Americans, and it is composed principally of the author's
edited writings on differential association during the years 1952-1963.
III Search of Crimin%8Y, published by Professor Leon Radzinowicz in 1961,
verified the fact that in the years since criminology began, and espeCially
since the decline of the "Biological School" near the beginning of the
present century, the quantity of criminological research has increased
enormously, but adequate knowledge about crime and delinquency cau-
sation has not increased proportionately. This survey revealed, among
other things, that research currently being conducted in one crimino-
logical centre often is completely unrelated to research conducted in
another, principally because there are wide variations in definitions of
the subject matter of criminology, in conceptions of criminology's
objectives, and in "basic" theory about the etiology of crime and de-
linquency and the reformation of criminals. In view of the widespread
international cooperation in other areas of science, such a lack of con-
sensus in criminology is rather appalling.
However, criminology has moved forward considerably in the last half
century despite its limitations. With the passage of time, more and more
concentrated efforts have been made to integrate, by means of theory,
the findings of empirically-oriented criminologists. The objective here
is the production of logical generalizations which would "make sense"
MAR TINUS NIJHOFF PUBLISHER THE HAGUE
of the known facts about individual delinquent and criminal conduct,

and about variations in delinquency and crime rates. Sutherland's theory
is one of the most noteworthy efforts of this kind, and it has done
much to increase criminological consensus, especially in the United States.
The theory can be stated in the form of the following nine propositions:
1. Criminal behaviour is learned.
2. Criminal behaviour is learned in interaction with other persons in a
process of communication.
3. The principal part of the learning of criminal behaviour occurs within
intimate personal groups.
4. When criminal behaviour is learned, the learning includes (a) tech-
niques of committing the crime, which are sometimes very compli-
cated, sometimes very simple; (b) the specific direction of motives,
drives, rationalizations, and attitudes.
5. The specific direction of motives and drives is learned from definition
of the legal codes as favourable or unfavourable.
6. A person becomes delinquent because of an excess of definitions
favourable to violation of law over definitions unfavourable to vio-
lation of law.
7. Differential associations may vary in frequency, duration, priority and
intensity.
8. The process of learning criminal behaviour by association with crimi-
nal and anti-criminal patterns involves all of the mechanisms that are
involved in any other learning.
9. While criminal behaviour is an expression of general needs and values,
it is not explained by those general needs and values since non-
criminal behaviour is an expression of the same needs and values.
The author hopes that this volume will further increase criminological
consensus by bringing together in one place some analyses of one of the
most important ideas put forth by an American criminologist, who
anticipated much of the current sociological and social psychological
thought in criminology. The author hopes, also, that his comments on
and criticisms of the theory will lead to a more widespread concern for
its research implications than has been the case in the past.
About the author: B. S. in SOCiology, Iowa State University (1943); lecturer to Instructor
in Sociology, University of California, Los Angeles (1949-'50); Ph.D. in Sociology,
Indiana University (1950); Research Associate, California Institution for Men, Chino,
California (1950-'51); Research Associate, United States Penitentiary, Terre Haute,
Indiana (1951); Assistant Professor of Sociology, University of California, Los Angeles
(1951-'56); Research Associate, Center for Education and Research in Corrections,
University of Chicago (1955-'56); Associate Professor to Professor of Sociology and
Acting Dean, Division of Social Sciences, University of California, los Angeles
(1956-'61); Visiting Professorial Fellow, University of Cambridge, England (1961-'62);
Dean of the College of letters and Science, and Professor of Sociology, University
of California, Santa Barbara (1962- ).
1964. IX and 167 pages. roy. 8vo. Guilders 18.-

MAR TINUS NIJHOFF - PUBLISHER THE HAGUE
Aspects of Art ForBery

PAPERS READ BY
H. VAN DE WAAL (Leiden)

TH. wtiR TENBERGER (Freiburg i. Br.)
W. FROENTJES (Lei den)
:11 a SYI1lPOSIlIIn organized by the

lNSTITUTE cr CRIMIi'\A.L L\ \V Ai'\D CRI~I1'\OLOGY
of the University oj Leidel!
Contents
H. VAN DE WAAL
Forgery as a Stylistic Problem
THO~L\S \VtiR TEKBERGER

Criminological and Criminal-La\v Problems of the Forging
of Paintings
W. FROENTJES
Criminalistic Aspects of Art Forgery
1962. X and 54 pages. \Vith 13 plates. roy. 8\0. Guilders 6.50
= Strafrechtelijk
een Criminologische Onderzoekingen. l\ieu\\e Reeks (Penal and
Criminological Researches. New Series), Volume VI.
MAR TINUS NIJHOFF - PUBLISHER THE HAGUE
--------------------------
Sexual Crime Today

PAPERS READ by
MAX GRUNHUT (Oxford)

RUDOLF SIEVERTS (Hamburg)
JACOB M. VAN BEMMELEN (Leiden)
at a symposium or8anized by the

INSTITUTE OF CRIMINAL LAW AND CRIMINOLOGY
of the University oj Leidell
Contents
MAXGRUNHUT
Penal and Corrective Treatment of Sexual Offenders
J. M. VAN BEMMELEN
Sexual Crimes Today
Data on Sexual Delinquency in The Netherlands
RUDOLF SIEVERTS
The Evolution of Sexual Criminality in Germany
Annexes
1960. VIII and 86 pages. With 20 tables and 10 graphs. roy. 8vo.
Guilders 7.60
-Strafrechtelijke en Criminologische Onderzoekingen, Nieuwe Reeks (Penal and
Criminological Researches. New Series), Volume IV.
One guilder = ab. $ 0.278 = abo 2 sh = env Fr. 1,36 = ca DMW 1.10
Obtainable through any bookseller or direct from the publisher

1964 Book DelinquencyCrimeAndDifferentia

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1964 Book DelinquencyCrimeAndDifferentia

Uploaded by

Copyright:

Available Formats

DELINQUENCY, CRIME

DONALD R. CRESSEY, PH. D .

Copyright 1964 by Martinus Nijhoff, The Hague, Netherlands

This is a book about Edwin H. Sutherland's theory of differ-

semble the volume during my 1961-1962 year as Visiting Fellow

We wish to express our thanks to the following publishers,

American Journal oj Sociology, 69 (September, 1963), pp. 129-142

AN INTRODUCTION TO THE THEORY OF DIFFERENTIAL ASSOCIATION

I. A Statement of the Theory. . . . . . . . . 3

SOME CRITICISMS OF THE THEORY

V. Some Popular Criticism of Differential Association. 81

USE OF THE THEORY IN CORRECTIONS

VIII. Differential Association and Correctional Group

AN INTRODUCTION TO THE THEORY

A STATEMENT OF THE THEORY

The first formal statement of Edwin H. Sutherland's theory

such as economics, psychology and biology. It was his efforts to

behaviour to be explained is redefined, so that the case is ex-

edition of Principles of Criminology was short lived. For reasons

"First, the processes which result in systematic criminal-behaviour are

"Second, systematic criminal behaviour is determined in a process of

behaviour unless he has had training in that kind of behaviour,

"Third, dijferential association is the specific causal process in the

automobiles or the burglary of stores as reprehensible, and

"Fourth, the chance that a person will participate in systematic criminal

"Fifth, individual differences among people in respect to personal

"Sixth, cultural conflict is the underlying cause of dijJerential association

to behave. The attempts to explain the behaviour of a particular

"Seventh, social disorganization is the basic cause qf systematic criminal

forms. When a gang starts in a disorganized district of a city it

It should also be noted that the above statement was some-

"the very trivial criminal acts" were "systematic," but that he

significance of much of the knowledge acquired about crime and

human behaviour. However, an explanation of criminal be-

A motion picture several years ago showed two boys engaged

Two Types of Explanations qf Criminal Behaviour. Scientific

The objective situation is important to criminality largely to

Genetic Explanation of Criminal Behaviour. The following state-

haviour, just as a person does not make mechanical inventions

is not in contact with criminal behaviour during the time he is so

It is not necessary, at this level of explanation, to explain why

criminal behaviour and in that sense the crime rate is an ex-

Differential Association and Social Processes

Second, in any society, the two kinds of definitions of what is

than it is in some rural areas can be attributed to differences in

The criminal laws established in more recent times, however,

not be important to another, or a group may believe that uni-

or a part of a group is involved. Second, a second group, or an-

Some Literary Errors

individuals, however, the difficulty in communication seems to

associations, in comparison with associations with anti-criminal

well has recently been as critical of the word "systematic" as was

existing theory of 'differential association,'" and Jeffery's state-

ential association" is equivalent to "differential association with

80 Howard B. Gill, "An Operational View of Criminology", Archives of Criminal

SOME OBSTACLES TO GENERALIZING IN

In the years since criminology began, and especially since the

1 Frank L. Tolman, "Study of Sociology in Institutions of Learning in the United

grown the number of causes which can be described as social.

Measuring Crime Rates

affected by conditions which cannot readily be taken into account

Criminal Justice (january, 1950), pp. 3-6.

It has been pointed out that only an insignificant proportion of

A further complication in this regard arises because some crimes