Differential Association Theory - Criminology

Differential association theory, developed by Edwin H. Sutherland, proposes that criminal behavior is learned through social interaction in intimate personal groups and that an individual becomes criminal when definitions favorable to violation of law exceed definitions unfavorable to violation of law. As one of the most influential theoretical contributions in the history of Criminology Theories, differential association theory redirected the field away from biological and psychological explanations of crime toward the sociological analysis of learning processes, establishing the proposition that crime is a product of normal social learning rather than individual pathology. This article examines Sutherland’s intellectual development and the social context in which the theory emerged, presents the nine propositions in their final 1947 formulation, discusses the macro-level complement of differential social organization, surveys the empirical evidence, traces the theory’s influence on subsequent theoretical developments, and assesses its criticisms, contemporary relevance, and policy implications.

Outline

I. Introduction

II. Sutherland and the Origins of Differential Association

a. The Intellectual Context: Against Biological Determinism
b. The Development of the Theory: 1939 to 1947
c. Sutherland’s Broader Contributions: White-Collar Crime

III. The Nine Propositions

a. Propositions 1–3: Crime as Learned Behavior
b. Propositions 4–6: The Content and Direction of Learning
c. Propositions 7–9: The Principle of Differential Association

IV. Differential Social Organization

a. The Macro-Level Complement
b. Normative Conflict and Group Rates of Crime
c. The Link to Social Disorganization

V. Empirical Evidence

a. Early Tests: Short, Reiss, and the Operationalization Challenge
b. Matsueda’s Structural Equation Models
c. Bruinsma’s Reconsidered Test
d. The Peer Effect Literature

VI. Applications

a. White-Collar and Corporate Crime
b. Juvenile Delinquency and Gang Membership
c. Cybercrime and Online Learning Environments

VII. Criticisms and Limitations

a. Operationalization and Testability
b. The Tautology Problem
c. Individual Variation and the Role of Choice

VIII. Legacy and Theoretical Influence

a. Social Learning Theory: Akers’s Extension
b. Subcultural Theory and Cultural Transmission
c. Differential Association in Integrated Theories

IX. Policy Implications

a. Restructuring Associations
b. Altering Definitions
c. Community-Level Prevention

X. Conclusion

XI. References

Introduction

Why do some individuals become criminals while others, raised in similar circumstances, remain law-abiding? Is criminal behavior the product of biological predisposition, psychological defect, or social learning? Edwin Sutherland’s differential association theory answered these questions with a proposition that was revolutionary for its time and remains foundational for contemporary criminology: criminal behavior is learned through the same social processes by which all behavior is learned — through communication, interaction, and association with others in intimate personal groups.

Sutherland’s theory, first presented in 1939 and finalized in 1947, offered a general explanation of crime that applied across all offense types, social classes, and demographic groups. Unlike biological theories that attributed crime to hereditary defects or psychological theories that located its causes in personality disorders, differential association theory proposed that crime is a normal product of normal social processes — that anyone, regardless of biological constitution or psychological makeup, can become criminal if exposed to a sufficient excess of definitions favorable to violation of law. This proposition established sociological criminology on a firm theoretical foundation and provided the intellectual platform on which subsequent theories — most notably Ronald Akers’s social learning theory — would build.

Differential association theory within Criminology Theories occupies a distinctive position as both a historically foundational framework and a living theoretical tradition. Its core insight — that criminal behavior is learned through social interaction and that the balance of pro-criminal and anti-criminal influences determines individual behavior — remains embedded in virtually every contemporary sociological theory of crime and continues to inform criminological research and criminal justice policy.

Sutherland and the Origins of Differential Association

The Intellectual Context: Against Biological Determinism

Sutherland developed differential association theory in explicit opposition to the biological and psychological explanations of crime that dominated American criminology in the first decades of the twentieth century. During the 1920s and 1930s, prominent criminologists attributed criminal behavior to inherited biological defects (Lombroso’s born criminal), mental deficiency (Goddard’s feeblemindedness thesis), constitutional body types (Sheldon’s somatotyping), and psychological pathology. These theories located the causes of crime within the individual offender — in defective genes, abnormal brain structures, or disordered personalities — and implied that criminals were fundamentally different kinds of people from law-abiding citizens.

Sutherland rejected these explanations on both empirical and theoretical grounds. The empirical evidence linking biological and psychological characteristics to crime was weak and methodologically flawed, relying on selected samples of institutionalized offenders without adequate control groups. More fundamentally, individual-level explanations could not account for the most important features of the distribution of crime: why crime rates varied systematically across social groups, neighborhoods, and cultures, and why these patterns persisted over time despite complete turnover in the populations occupying high-crime areas. These ecological patterns — documented by Sutherland’s colleagues Clifford Shaw and Henry McKay in their studies of juvenile delinquency in Chicago — demanded a sociological explanation that focused on social processes rather than individual characteristics.

Sutherland was influenced by the Chicago School tradition of urban sociology, particularly the work of W. I. Thomas and Florian Znaniecki on social disorganization and the work of George Herbert Mead on symbolic interactionism. From Thomas and Znaniecki, Sutherland drew the idea that social behavior is shaped by the normative systems to which individuals are exposed. From Mead, he drew the insight that meaning is constructed through social interaction — that individuals develop their orientations toward the world through communication with others in their social environment. These influences converged in a theory that located the causes of crime not in the individual but in the social relationships and cultural environments that shape individual behavior.

The Development of the Theory: 1939 to 1947

Sutherland first presented differential association theory in the third edition of his textbook Principles of Criminology (1939). This initial formulation consisted of seven propositions and was framed within the broader concept of “systematic criminal behavior” — behavior that is learned, organized, and culturally transmitted through group interaction. The 1939 version was influenced by Sutherland’s understanding of social disorganization theory but was not yet fully developed as a general theory of criminal behavior.

The definitive formulation appeared in the fourth edition of Principles of Criminology (Sutherland, 1947), where Sutherland revised the theory into nine propositions that constituted a more precise and more general statement. The revision reflected Sutherland’s growing dissatisfaction with the concept of social disorganization, which he came to regard as vague and potentially tautological. In place of social disorganization, Sutherland introduced the concept of “differential social organization” — the proposition that society is organized both for and against crime, and that the relative strength of pro-criminal and anti-criminal organization in a community determines its crime rate. This reformulation shifted the macro-level complement of the theory from a focus on the absence of organization (disorganization) to a focus on the presence of competing forms of organization — a more precise and analytically productive conceptualization.

The 1947 formulation remained unchanged through subsequent editions of the textbook, which were revised after Sutherland’s death in 1950 by his student Donald Cressey. Cressey defended and elaborated the theory through several subsequent editions, addressing criticisms and clarifying ambiguities while preserving the nine propositions in their original form. The theory has remained one of the most discussed, debated, and empirically tested frameworks in the history of criminological thought.

De Fleur and Quinney (1966) contributed a formal reformulation of the theory using set theory notation, deriving the nine propositions from a set of underlying postulates and generating additional testable predictions. Their reformulation demonstrated that Sutherland’s verbal theory contained a logically coherent deductive structure and provided a strategy for empirical verification that influenced subsequent quantitative tests. Matsueda (1988) would later credit this formalization as an important step in establishing differential association as a testable scientific theory rather than merely a verbal framework.

Sutherland’s Broader Contributions: White-Collar Crime

Sutherland’s theoretical work was inseparable from his empirical research on white-collar crime — a concept he introduced in his 1939 presidential address to the American Sociological Association and developed in his landmark monograph White Collar Crime (1949). Sutherland defined white-collar crime as crime committed by a person of respectability and high social status in the course of his occupation, and he documented the extensive criminal violations committed by the seventy largest American corporations — violations including fraud, restraint of trade, patent infringement, and unfair labor practices.

The study of white-collar crime was both an empirical contribution and a theoretical demonstration. If crime were caused by biological defects, mental deficiency, or poverty — as the prevailing theories argued — then white-collar crime should not exist, because corporate executives and professionals are neither biologically defective, nor mentally deficient, nor impoverished. The existence of extensive criminality among the wealthy and well-educated demonstrated that the causes of crime must lie in social processes that operate across all social classes — precisely the kind of processes that differential association theory described.

White-collar crime also illustrated the theory’s proposition that criminal behavior is learned through interaction in intimate personal groups. Corporate executives learn to violate antitrust laws, falsify financial statements, and evade safety regulations not through contact with street criminals but through their associations with other executives in corporate environments where such practices are normalized, rationalized, and rewarded. The corporate culture — its definitions of acceptable conduct, its informal norms regarding legal compliance, and its patterns of reinforcement for profit-maximizing behavior regardless of legality — functions as the differential association context in which white-collar criminal behavior is learned.

The Nine Propositions

Propositions 1–3: Crime as Learned Behavior

The first three propositions establish the foundational premise that criminal behavior is learned through social interaction. Proposition 1 states that criminal behavior is learned — it is not inherited, nor does it arise spontaneously in individuals who have had no exposure to criminal patterns. This proposition directly rejected the biological determinism that dominated early criminology by insisting that criminal behavior, like all behavior, is acquired through experience and social contact rather than transmitted through genetic inheritance.

Proposition 2 states that criminal behavior is learned in interaction with other persons in a process of communication. Learning occurs through both verbal and nonverbal communication — through conversation, observation, gesture, and example. This proposition specifies the mechanism through which learning occurs: not through passive exposure to criminal stimuli (such as media representations of crime) but through active social interaction with other persons. Sutherland emphasized that the communication process involves the exchange of meanings — the shared definitions, interpretations, and evaluations through which individuals make sense of their social world.

Proposition 3 states that the principal part of the learning of criminal behavior occurs within intimate personal groups. Family members, close friends, and peer associates exert far greater influence on the individual’s behavioral development than do impersonal sources of information such as newspapers, movies, or — in the contemporary context — television, social media, and the Internet. This proposition reflects Sutherland’s grounding in the symbolic interactionist tradition, which emphasizes the primacy of face-to-face interaction in the construction of meaning and the development of behavioral orientations. Although the proposition has been questioned in the digital age, Sutherland’s core insight — that intimate, emotionally significant relationships carry more weight than distant or impersonal ones — remains empirically supported.

Propositions 4–6: The Content and Direction of Learning

Propositions 4 through 6 specify what is learned through the process of differential association. Proposition 4 states that when criminal behavior is learned, the learning includes both the techniques of committing the crime — which are sometimes very complicated and sometimes very simple — and the specific direction of motives, drives, rationalizations, and attitudes. This proposition recognizes that criminal behavior requires not only technical knowledge (how to pick a lock, how to forge a document, how to manufacture methamphetamine) but also a motivational and attitudinal orientation that supports the commission of the act. An individual who possesses the technical skill to commit a crime but who holds definitions unfavorable to its commission will not commit it; conversely, an individual who is motivated to commit a crime but lacks the necessary skills will be unable to do so.

Proposition 5 states that the specific direction of motives and drives is learned from definitions of the legal codes as favorable or unfavorable. In some social environments, the prevailing definitions support compliance with the law: stealing is wrong, violence is unacceptable, cheating is dishonorable. In other environments, definitions favorable to violation of law predominate: stealing from corporations is justified because they exploit their workers, violence is an appropriate response to disrespect, cheating is acceptable because everyone does it. These definitions may be general — broad moral orientations toward right and wrong — or specific — attitudes toward particular behaviors, laws, or situations.

Proposition 6 states that a person becomes delinquent because of an excess of definitions favorable to violation of law over definitions unfavorable to violation of law. This is the principle of differential association — the central proposition of the entire theory. An individual who is exposed to more definitions favorable to crime than definitions unfavorable to crime will become criminal; an individual whose balance of definitions tilts toward conformity will remain law-abiding. The “excess” is not simply a numerical count but reflects the frequency, duration, priority, and intensity of the associations through which definitions are acquired.

Propositions 7–9: The Principle of Differential Association

Propositions 7 through 9 qualify and elaborate the process of differential association. Proposition 7 states that differential associations may vary in frequency, duration, priority, and intensity. Associations that occur more frequently, last longer, begin earlier in life (priority), and involve more emotionally significant partners (intensity) will exert greater influence on the individual’s balance of definitions than associations that are infrequent, brief, recent, or superficial. These four modalities provide the framework for explaining individual variation in criminal behavior: individuals exposed to similar social environments may differ in their behavior because their specific patterns of association differ in frequency, duration, priority, and intensity.

Proposition 8 states that the process of learning criminal behavior by association with criminal and anti-criminal patterns involves all of the mechanisms that are involved in any other learning. This proposition asserts the generality of the learning process — crime is learned through the same mechanisms as any other behavior, not through a special criminal learning process. The proposition was deliberately left unspecified because Sutherland recognized that the mechanisms of learning were the subject of active debate within psychology and that committing to a particular learning theory might limit the generality of his own framework. This deliberate vagueness would later become one of the theory’s most criticized features and would motivate Burgess and Akers’s (1966) reformulation, which specified operant conditioning and observational learning as the concrete mechanisms.

Proposition 9 states that while criminal behavior is an expression of general needs and values, it is not explained by those general needs and values, since non-criminal behavior is an expression of the same needs and values. The desire for money motivates both legitimate work and theft; the desire for status motivates both academic achievement and gang leadership; the desire for excitement motivates both extreme sports and criminal adventure. General needs and values cannot explain why some individuals satisfy them through criminal means while others satisfy them through legal means — only the specific content of the individual’s differential associations can explain this.

Table 1. Sutherland’s Nine Propositions of Differential Association

Proposition	Statement	Level
1	Criminal behavior is learned	Individual
2	Learned in interaction through communication	Individual
3	Principal learning occurs in intimate personal groups	Individual
4	Learning includes techniques, motives, drives, rationalizations, attitudes	Individual
5	Direction of motives learned from definitions favorable/unfavorable to law	Individual
6	Person becomes criminal when definitions favorable to violation exceed unfavorable	Individual
7	Associations vary in frequency, duration, priority, intensity	Individual
8	Learning criminal behavior involves same mechanisms as any other learning	Individual
9	Criminal behavior expresses same general needs as noncriminal behavior	Individual

Differential Social Organization

The Macro-Level Complement

While the nine propositions explain individual criminal behavior, Sutherland recognized that a complete theory of crime must also explain why crime rates differ across groups, communities, and societies. The concept of differential social organization — introduced in the 1947 formulation as a replacement for social disorganization — provides this macro-level explanation. Differential social organization refers to the proposition that any community or society is organized both for and against criminal behavior: some groups, institutions, and cultural traditions support conformity to the law, while other groups, institutions, and traditions support violation of the law.

The crime rate of a community, in this framework, is determined by the relative strength of the organization favorable to crime and the organization unfavorable to crime within that community. A community in which conventional institutions — families, schools, churches, civic organizations — are strong and criminal organizations are weak will have a low crime rate because the predominant associations available to its members expose them to an excess of anti-criminal definitions. A community in which conventional institutions are weak and criminal organizations are strong will have a high crime rate because the predominant associations expose members to an excess of pro-criminal definitions.

This formulation differs from the social disorganization framework in an important respect. Social disorganization implies that high-crime communities lack organization — that they are characterized by the absence of social structure. Differential social organization implies that high-crime communities are organized, but organized in ways that support criminal behavior as well as (or instead of) conforming behavior. This distinction is analytically significant because it recognizes that criminal subcultures, drug markets, gangs, and criminal networks are themselves forms of social organization — organized structures with their own norms, hierarchies, and socialization processes — rather than symptoms of organizational absence.

Normative Conflict and Group Rates of Crime

Sutherland’s concept of normative conflict provides the theoretical link between differential social organization and individual criminal behavior. In any complex society, multiple normative systems coexist — systems that define different behaviors as acceptable or unacceptable, criminal or legitimate. These normative systems conflict with one another: the conduct norms of one group may prohibit behaviors that the conduct norms of another group permit or even encourage. Criminal law represents the triumph of one set of norms over another — the codification of the dominant group’s conduct norms into legally enforceable rules.

Normative conflict produces differential crime rates because it creates variation in the definitions to which individuals are exposed. In communities where there is broad agreement on the wrongfulness of crime — where conventional and legal norms are aligned and reinforced by strong institutional support — individuals are exposed predominantly to anti-criminal definitions, and crime rates are low. In communities where normative systems conflict — where some groups define certain illegal behaviors as acceptable, necessary, or admirable — individuals are exposed to competing definitions, and those whose associations tip toward pro-criminal definitions are more likely to engage in criminal behavior.

This analysis anticipated the later development of subcultural theory by Albert Cohen (1955) and Richard Cloward and Lloyd Ohlin (1960), who examined how specific subcultures transmit pro-criminal definitions to their members. It also anticipated the cultural criminology tradition, which examines how shared meanings, styles, and expressive practices shape the construction of crime and deviance in particular social environments.

The Link to Social Disorganization

The relationship between differential association and social disorganization theory is one of complementarity rather than competition. Shaw and McKay’s social disorganization theory explains why certain communities have high crime rates — because ecological conditions such as poverty, residential instability, and ethnic heterogeneity weaken the institutional structures that maintain informal social control. Differential association theory explains how these community-level conditions translate into individual criminal behavior — by altering the balance of definitions favorable and unfavorable to crime to which community members are exposed.

In socially disorganized communities, conventional institutions are weak, criminal subcultures are present, and youth are exposed to an abundance of criminal models, definitions, and opportunities through their associations with delinquent peers and criminal adults. The transmission of delinquent traditions from one generation to the next — the process that Shaw and McKay termed cultural transmission — is precisely the mechanism that differential association theory describes at the individual level: older, more experienced offenders transmit techniques, rationalizations, and pro-criminal definitions to younger community members through the intimate personal interactions specified in Sutherland’s propositions.

Matsueda (1988) formalized this multilevel relationship, demonstrating that the structural characteristics of communities (poverty, instability, heterogeneity) affect individual delinquency through their effects on the learning processes specified by differential association theory — specifically, through their effects on the density of delinquent peer associations and the prevalence of pro-criminal definitions within the community. This analysis confirmed that differential association operates as the micro-level mechanism through which macro-level social conditions produce individual criminal behavior.

Empirical Evidence

Early Tests: Short, Reiss, and the Operationalization Challenge

The empirical testing of differential association theory has been complicated by the difficulty of operationalizing its core concepts — particularly the “excess of definitions favorable to violation of law” that constitutes the principle of differential association. How can researchers measure the balance of pro-criminal and anti-criminal definitions to which an individual has been exposed across all of their intimate personal associations, weighted by the frequency, duration, priority, and intensity of each association? The measurement challenge is formidable, and early critics argued that the theory was essentially untestable because its key variables could not be adequately operationalized.

James Short (1957) conducted one of the earliest empirical tests by surveying 126 male and 50 female trainees in a state training school, using a self-report questionnaire that asked respondents about their exposure to definitions favorable and unfavorable to criminal behavior. Short found a significant positive relationship between the reported frequency of exposure to delinquent behavior patterns and self-reported delinquency, and an even stronger relationship when the measures were weighted by intensity (the closeness of the relationship). Albert Reiss and A. Lewis Rhodes (1964) examined the friendship patterns of delinquent and nondelinquent boys and found that delinquency was significantly associated with having close friends who were also delinquent — consistent with the theory’s prediction.

These early studies provided preliminary support for the theory but were limited by the difficulty of measuring the full range of associations and definitions specified by Sutherland. Cross-sectional surveys could capture only a snapshot of the individual’s associational patterns at one point in time, rather than the cumulative balance of associations over the lifespan that the theory describes. The measurement challenge would preoccupy researchers for decades and would eventually motivate Akers’s reformulation, which replaced the difficult-to-measure “definitions” with the more operationally precise concepts of reinforcement and imitation drawn from behavioral psychology.

Matsueda’s Structural Equation Models

Ross Matsueda (1982, 1988) conducted the most sophisticated tests of differential association theory, using structural equation modeling to examine the causal relationships among the theory’s key variables. Matsueda (1982) tested a model in which the structural characteristics of families and communities affect delinquency through their effects on delinquent peer associations and pro-delinquent definitions. Using data from the Richmond Youth Survey, he found that the effect of delinquent peers on delinquency was substantially mediated by the individual’s own definitions favorable to crime — consistent with the theory’s prediction that associations affect behavior through the learning of definitions.

Matsueda’s (1988) follow-up study, “The Current State of Differential Association Theory,” provided a more complete test by distinguishing between the theory’s individual-level and macro-level propositions. At the individual level, he confirmed that delinquent associations predict delinquent definitions, which in turn predict delinquent behavior. At the macro-level, he found that community characteristics (poverty, residential instability, proportion nonwhite) affected individual delinquency through their effects on the availability of delinquent associations — supporting the differential social organization proposition.

These studies were important because they provided the first rigorous causal modeling of the relationships specified by the theory, moving beyond simple bivariate correlations to estimate the full network of causal paths. Matsueda’s finding that the effect of delinquent peers on delinquency is mediated by definitions was particularly significant because it distinguished differential association from a simple “bad companions” theory: it is not merely the presence of delinquent peers that produces criminal behavior but the learning of definitions favorable to crime that occurs through those associations.

Bruinsma’s Reconsidered Test

Gerben Bruinsma (1992) conducted an empirical test of an expanded version of Sutherland’s theory based on Karl-Dieter Opp’s (1974) formal reformulation. Using survey data from 1,196 boys and girls aged 12 to 17, Bruinsma tested a model that included variables representing the frequency of contacts with deviant behavior patterns, the development of positive definitions, the frequency of communication about criminal techniques, and the opportunities for committing crime. The model explained 51 percent of the variance in criminal behavior — a substantial effect size comparable to those reported in tests of Akers’s social learning theory.

Bruinsma’s test also showed that the impact of the frequency of contacts with deviant behavior patterns on the development of positive definitions and on the frequency of communication about relevant techniques was substantial, and that these effects were strongest when the individual had close identification with deviant others. This finding was consistent with Sutherland’s proposition that the intensity of associations — the emotional significance of the relationship — moderates the learning process. The more closely the individual identifies with deviant associates, the more strongly those associations influence the development of pro-criminal definitions and the acquisition of criminal techniques.

The Bruinsma study was notable because it tested Sutherland’s theory in its own terms rather than as a component of Akers’s broader social learning framework, demonstrating that the core propositions of differential association retain substantial explanatory power even without the additional mechanisms (reinforcement, imitation) that Akers incorporated.

The Peer Effect Literature

A large body of research on peer effects has examined the relationship between delinquent peer association and individual delinquency — a relationship that is consistent with differential association theory but also interpretable through other theoretical frameworks (including social control theory and selection effects models). Studies have consistently found that association with delinquent peers is one of the strongest and most consistent predictors of individual delinquency, rivaling or exceeding the predictive power of variables from competing theories (Warr, 2002).

The central debate in this literature concerns the causal direction of the peer-delinquency relationship. Differential association theory predicts that delinquent peers socialize individuals into crime through the transmission of pro-criminal definitions. Social control theory predicts that weak social bonds lead to crime, with delinquent peer association being a consequence rather than a cause of delinquency. Longitudinal studies and social network analyses have generally found evidence for both processes — socialization and selection — operating simultaneously, with the relative strength of each varying across studies and samples (Gallupe, McLevey, & Brown, 2018).

Warr (2002) provided one of the most detailed analyses of peer effects in Companions in Crime, demonstrating that the acquisition of delinquent friends typically precedes the onset of delinquent behavior and that changes in friendship networks — acquiring new delinquent friends or losing existing ones — predict changes in individual offending. These findings are more consistent with a socialization model (differential association) than a pure selection model, although Warr acknowledged that both processes are likely at work in most cases.

McGloin and Thomas (2019) extended this line of research using social network analysis to examine how the structure — not just the composition — of peer networks affects individual offending. They found that network density, the proportion of delinquent ties, and the individual’s centrality within a delinquent network all predicted criminal behavior, consistent with differential association’s emphasis on the frequency and intensity of criminal associations. Their work represents a methodological advance that allows researchers to test Sutherland’s modalities with greater precision than traditional survey measures of peer delinquency.

Table 2. Key Empirical Tests of Differential Association Theory

Study	Data/Sample	Method	Key Finding
Short (1957)	176 training school trainees	Self-report survey	Frequency of delinquent contacts predicts delinquency; intensity moderates effect
Reiss & Rhodes (1964)	Delinquent and nondelinquent boys	Friendship network analysis	Delinquency associated with having close delinquent friends
Matsueda (1982)	Richmond Youth Survey	Structural equation modeling	Delinquent peer effect mediated by pro-delinquent definitions
Matsueda (1988)	Richmond Youth Survey	Structural equation modeling	Community structure affects delinquency through associations and definitions
Bruinsma (1992)	1,196 youth aged 12–17	Expanded model test	R² = .51; contacts, definitions, and technique communication all significant
Warr (2002)	Multiple longitudinal datasets	Temporal analysis of peer effects	Acquisition of delinquent friends precedes onset of delinquent behavior

Applications

White-Collar and Corporate Crime

Differential association theory’s application to white-collar and corporate crime was central to Sutherland’s own research agenda and remains one of the theory’s most distinctive contributions. Sutherland (1949) demonstrated that the seventy largest American corporations had accumulated 980 adverse decisions — convictions, civil judgments, and regulatory orders — for violations including restraint of trade, false advertising, patent infringement, and unfair labor practices. These corporations were habitual offenders, and their criminal behavior was learned, organized, and transmitted through the institutional culture of the corporation.

The theory explains corporate crime through the same mechanism it applies to street crime: executives learn definitions favorable to violation of law through their interactions with other executives, industry peers, and organizational cultures that normalize illegal conduct. Cressey’s (1953) study of embezzlers — Other People’s Money — demonstrated that embezzlement was preceded by the development of rationalizations (definitions favorable to violation of law) that made the criminal act psychologically acceptable: “I’m just borrowing the money,” “the company owes me,” “everyone in the industry does it.” These rationalizations functioned precisely as the pro-criminal definitions that Sutherland’s theory specified.

Contemporary research on corporate crime continues to draw on differential association principles. Studies of financial fraud, environmental violations, and pharmaceutical misconduct have documented how organizational cultures transmit definitions favorable to law violation through mentorship, peer pressure, and institutional reward systems that prioritize profit over compliance. The theory’s ability to explain crime across the full class spectrum — from street crime to suite crime — remains one of its greatest strengths.

Juvenile Delinquency and Gang Membership

The application of differential association to juvenile delinquency was implicit in Sutherland’s original formulation and has been the focus of the majority of the theory’s empirical tests. The theory predicts that youth who are exposed to delinquent peers, delinquent family members, or delinquent neighborhood cultures will acquire definitions favorable to crime and will be more likely to engage in criminal behavior than youth whose associations are predominantly conventional.

Research on gang membership provides strong support for this application. Youth who join gangs are exposed to an intense concentration of pro-criminal definitions, criminal techniques, and reinforcement for criminal behavior within the gang context. Studies have consistently found that gang membership significantly increases individual criminal behavior — an effect that is partially explained by the learning processes that differential association theory describes. Thornberry, Krohn, Lizotte, Smith, and Tobin (2003) found that the onset of gang membership was followed by significant increases in delinquency and that leaving the gang was followed by significant decreases — a pattern consistent with the proposition that criminal behavior is learned through association and diminishes when criminal associations are severed.

The gang context also illustrates the cultural transmission process that links differential association to social disorganization theory. In neighborhoods where gangs have been established for generations, pro-criminal definitions are transmitted from older to younger cohorts through mentorship, initiation rituals, and the modeling of criminal behavior by respected gang leaders. New recruits learn not only the techniques of criminal activity but also the rationalizations, values, and subcultural norms that sustain it.

Cybercrime and Online Learning Environments

Contemporary research has extended differential association theory to the online environment, examining how criminal behavior is learned through digital associations. The Internet has created new contexts for differential association that Sutherland could not have anticipated: hacker forums, dark web marketplaces, extremist social media communities, and online gaming environments all provide spaces in which individuals can interact with others who hold and transmit definitions favorable to specific forms of criminal conduct.

Holt, Burruss, and Bossler (2010) found that social learning variables — including measures of differential association — predicted cyber-deviance, and that the effects were partially mediated by the learning of pro-deviant definitions. Subsequent research has examined whether online peers exert the same influence as offline peers, with findings suggesting that offline associations generally exert stronger effects but that online associations also make significant independent contributions to criminal learning (Dearden & Parti, 2021).

The extension of differential association to digital environments raises important questions about Sutherland’s third proposition — that the principal learning occurs within intimate personal groups. Online communities may lack the face-to-face intimacy that Sutherland emphasized, but they can generate intense identification, emotional attachment, and normative influence that function similarly to the intimate personal groups of the physical world. The adaptation of differential association theory to the digital age demonstrates the theory’s continuing relevance while also highlighting the need to update its assumptions about the nature of social interaction and intimate association.

Criticisms and Limitations

Operationalization and Testability

The most persistent criticism of differential association theory concerns the difficulty of operationalizing its core concepts — particularly the “excess of definitions favorable to violation of law” that constitutes the principle of differential association. How can researchers measure the cumulative balance of pro-criminal and anti-criminal definitions that an individual has acquired over a lifetime of social interaction, weighted by the frequency, duration, priority, and intensity of each association? The measurement challenge is formidable, and no study has yet fully operationalized the theory’s key variables as Sutherland specified them.

Critics argue that this difficulty makes the theory effectively untestable in its complete form. Researchers can measure proxies — the number of delinquent friends, the frequency of contact with deviant peers, self-reported attitudes toward crime — but these proxies capture only a fraction of the theoretical construct. The four modalities (frequency, duration, priority, intensity) have rarely been measured simultaneously, and the cumulative, lifelong balance of definitions that the theory describes is beyond the reach of cross-sectional survey methods.

Defenders of the theory respond that the operationalization challenge is a practical limitation rather than a theoretical flaw — that the difficulty of measuring a concept does not mean the concept is meaningless or that the theory is wrong. Matsueda’s (1982, 1988) structural equation models and Bruinsma’s (1992) expanded test demonstrate that the theory can be tested with available data and that the results are consistent with its predictions. The operationalization challenge has also served as a productive stimulus for theoretical development, motivating Akers’s reformulation of the theory in terms of more operationally precise behavioral concepts.

The Tautology Problem

A second criticism concerns the potential tautology in the theory’s central proposition. Proposition 6 states that a person becomes criminal when definitions favorable to violation of law exceed definitions unfavorable to violation of law. Critics argue that this proposition is tautological — essentially stating that a person commits crime because the person favors crime — unless “definitions” are measured independently of the criminal behavior they are supposed to explain. If definitions favorable to crime are measured by asking individuals whether they approve of criminal behavior, and criminal behavior is measured by asking whether they have committed crimes, then the correlation between the two may reflect nothing more than consistency in self-reporting.

Matsueda (1988) addressed this concern by demonstrating that the effect of definitions on delinquency persists even when definitions are measured at a prior time point and delinquency is measured subsequently — providing evidence that the relationship is causal rather than tautological. However, the concern remains relevant for cross-sectional studies that measure definitions and behavior simultaneously, and the degree to which the tautology critique undermines the theory’s explanatory power continues to be debated.

The tautology problem is not unique to differential association theory — similar concerns apply to social control theory (do weak bonds cause crime, or does crime weaken bonds?) and to strain theory (do criminal definitions result from strain, or does criminal behavior produce the strain?). The general solution — longitudinal designs that establish temporal ordering — has been applied with generally supportive results, but the reciprocal nature of the relationships between definitions, associations, and behavior means that the causal dynamics are more complex than simple unidirectional models can capture.

Individual Variation and the Role of Choice

A third criticism concerns the theory’s apparent denial of individual agency. If criminal behavior is determined by the balance of definitions to which an individual is exposed through social associations, then individuals appear to be passive products of their social environments — learning machines that absorb whatever definitions their associations transmit, without independent judgment, moral reasoning, or personal choice. Critics argue that the theory cannot explain why individuals exposed to the same associational patterns sometimes respond differently — why one sibling becomes delinquent while another, raised in the same household and exposed to the same associations, remains law-abiding.

Sutherland was aware of this criticism and addressed it in part through proposition 7’s specification that associations vary in frequency, duration, priority, and intensity — variations that can produce different behavioral outcomes even among individuals in the same household. However, the theory does not provide a satisfactory account of the internal cognitive processes through which individuals interpret, evaluate, and selectively respond to the definitions they encounter. This gap was partially addressed by Akers’s incorporation of cognitive learning processes from Bandura’s social cognitive theory, but the question of how individual agency operates within the framework of differential association remains an open theoretical issue.

The theory’s treatment of individual variation is closely related to its relationship to psychological theories of crime. Sutherland deliberately excluded psychological variables from his theory, arguing that they were unnecessary for a sociological explanation of crime. Critics counter that the exclusion of individual differences in temperament, intelligence, impulsivity, and moral reasoning limits the theory’s explanatory power and leaves it unable to account for within-group variation in criminal behavior among individuals with similar associational patterns.

Legacy and Theoretical Influence

Social Learning Theory: Akers’s Extension

The most important theoretical legacy of differential association is Ronald Akers’s social learning theory, which reformulated Sutherland’s propositions by incorporating the mechanisms of operant conditioning (from B. F. Skinner) and observational learning (from Albert Bandura) that Sutherland had deliberately left unspecified. Burgess and Akers (1966) translated each of Sutherland’s propositions into the language of behavioral psychology, specifying how differential association operates through the concrete mechanisms of reinforcement, punishment, and modeling.

Akers (1998) emphasized that social learning theory is not a rival of differential association but a broader framework that modifies, builds on, and integrates Sutherland’s propositions. Support for differential association provides support for social learning theory, because the core propositions are shared. The relationship between the two theories is one of specification: social learning theory retains all of Sutherland’s propositions while adding the behavioral mechanisms that make the learning process empirically more precise and practically more tractable.

The success of social learning theory — which has become one of the most extensively tested and broadly supported theories in criminology — represents the fulfillment of Sutherland’s theoretical vision through the incorporation of psychological mechanisms that he chose not to specify. The explanatory power that social learning theory has demonstrated across substance use, violence, sexual coercion, white-collar crime, and cybercrime is, in an important sense, evidence for the enduring validity of Sutherland’s core insight: that criminal behavior is learned through social interaction.

Subcultural Theory and Cultural Transmission

Differential association theory also directly influenced the development of subcultural theories of crime. Cohen’s (1955) theory of delinquent subcultures, Cloward and Ohlin’s (1960) differential opportunity theory, and Wolfgang and Ferracuti’s (1967) subculture of violence thesis all drew on Sutherland’s proposition that criminal behavior is learned through exposure to pro-criminal definitions transmitted within specific social groups. The concept of cultural transmission — the intergenerational passing of delinquent traditions within socially disorganized neighborhoods — is essentially the macro-level manifestation of the differential association process operating across cohorts.

These subcultural theories elaborated on aspects of differential association that Sutherland left undeveloped — particularly the question of how and why pro-criminal subcultures originate in the first place. Cohen’s status frustration thesis and Cloward and Ohlin’s differential opportunity framework offered structural explanations for the emergence of delinquent subcultures that complemented Sutherland’s account of how criminal definitions are transmitted within them. Together, differential association and subcultural theory provide a multilevel explanation of crime that links structural conditions to cultural processes to individual behavior.

The cultural transmission tradition has continued to influence contemporary research, including studies of gang culture, prison culture, and the online transmission of extremist ideologies. In each of these contexts, the core mechanism is the one Sutherland described: individuals learn criminal or deviant orientations through intimate interaction with others who model, reinforce, and rationalize such behavior.

Differential Association in Integrated Theories

Differential association has been incorporated into several integrated theories that combine propositions from multiple theoretical traditions. Elliott, Ageton, and Canter (1979) developed an integrated model that combines strain, social control, and social learning (differential association) variables into a sequential causal model: strain weakens social bonds, which increases exposure to delinquent peers, which increases delinquent definitions, which increases delinquent behavior. Thornberry’s (1987) interactional theory proposed reciprocal causal relationships among social bonds, delinquent associations, and delinquent behavior, treating differential association as one element of a dynamic developmental process.

These integrated frameworks testify to the enduring relevance of differential association’s core concepts. Even theorists who regard the theory as incomplete or insufficiently specified recognize that peer association and the learning of pro-criminal definitions are among the most powerful and consistent predictors of criminal behavior — variables that no adequate theory of crime can afford to exclude.

The integration of differential association with developmental and life-course perspectives represents one of the most productive contemporary extensions. Research has demonstrated that the influence of delinquent peers varies across the life course — peaking during adolescence, when peer influence is strongest, and declining during adulthood as individuals transition into conventional roles (marriage, employment, parenthood) that restructure their associational patterns.

Policy Implications

Restructuring Associations

The most direct policy implication of differential association theory is the need to alter the associational patterns that expose individuals to pro-criminal definitions. Programs that remove youth from delinquent peer networks and place them in environments dominated by prosocial associations — residential programs, mentoring relationships, structured after-school activities — seek to shift the balance of definitions toward conformity. Functional family therapy and multisystemic therapy target the family environment in which early associations are formed, working to strengthen parental bonds and reduce exposure to delinquent family members and peers.

However, the theory also cautions against interventions that inadvertently concentrate high-risk youth in settings where they are exposed to an intensified excess of pro-criminal definitions. Group-based interventions that bring delinquent youth together — such as boot camps, scared-straight programs, and some group therapy formats — may increase rather than decrease criminal behavior by strengthening delinquent associations and reinforcing pro-criminal definitions. The “peer contagion” effect documented in research on group interventions for delinquent youth (Dishion, McCord, & Poulin, 1999) is precisely what differential association theory would predict.

The theory also supports community-level interventions that restructure the associational opportunities available to residents of high-crime neighborhoods — creating prosocial gathering spaces, youth development programs, and community organizations that provide alternatives to the criminal associations that predominate in disorganized communities.

Altering Definitions

A second policy strategy targets the definitions that individuals hold regarding criminal behavior. Cognitive-behavioral interventions — the most effective approach to offender treatment according to the “what works” literature — directly address pro-criminal definitions (antisocial attitudes, criminal thinking patterns, rationalizations for offending) by teaching individuals to recognize, challenge, and replace them with prosocial alternatives. These interventions are grounded in the differential association insight that criminal behavior is sustained by the definitions that motivate and justify it.

School-based prevention programs that teach resistance skills, moral reasoning, and prosocial decision-making seek to strengthen anti-criminal definitions before pro-criminal ones take hold. Media literacy programs that help young people critically evaluate media representations of crime and violence address the indirect transmission of pro-criminal definitions through mass communication — a pathway that Sutherland underestimated but that has become increasingly significant in the age of social media and digital content.

Community-based messaging campaigns that communicate anti-criminal norms — such as anti-violence campaigns, community standards initiatives, and positive youth development programs — seek to shift the normative environment at the community level, increasing the prevalence of anti-criminal definitions and reducing the normative support for criminal behavior.

Community-Level Prevention

At the macro level, differential association theory supports community-level interventions that address the differential social organization that produces high crime rates. Investments in conventional institutions — schools, churches, community centers, civic organizations — strengthen the organizational infrastructure that generates anti-criminal definitions and provides prosocial associational opportunities. Economic development programs that create legitimate employment reduce the relative attractiveness of criminal alternatives and provide institutional contexts in which conventional definitions are reinforced.

The theory also supports efforts to reduce the concentration of disadvantage that produces the normative conflict and criminal subcultures described by Sutherland. Residential mobility programs (such as the Moving to Opportunity experiment), mixed-income housing developments, and neighborhood revitalization initiatives that diversify the social composition of disadvantaged communities can alter the associational landscape by introducing conventional role models and reducing the density of pro-criminal associations.

These community-level strategies reflect Sutherland’s macro-level proposition that crime rates are determined by the relative strength of the organization favorable to crime and the organization unfavorable to crime within a community. By strengthening the organization unfavorable to crime — through institutional investment, economic opportunity, and residential diversification — communities can shift the balance of definitions toward conformity and reduce the production of criminal behavior through the social learning processes that Sutherland described.

Conclusion

Differential association theory transformed criminology by establishing the proposition that criminal behavior is a product of normal social learning rather than individual pathology. Sutherland’s nine propositions — that crime is learned through communication in intimate personal groups, that the learning includes techniques and definitions, and that an excess of definitions favorable to violation of law produces criminal behavior — provided a general theory that applied across all social classes, offense types, and demographic groups. The theory’s ability to explain white-collar crime alongside street crime, juvenile delinquency alongside corporate fraud, demonstrated that the causes of crime lie not in the defective biology or psychology of individual offenders but in the social relationships and cultural environments that shape their behavioral orientations.

The theory’s limitations — particularly the difficulty of operationalizing its core concepts and its deliberate vagueness about the mechanisms of learning — stimulated the development of Akers’s social learning theory, which specified the behavioral mechanisms through which differential association operates. The success of social learning theory confirms rather than replaces Sutherland’s framework: the core insight that criminal behavior is learned through social interaction remains the foundation on which the most extensively tested and broadly supported theory in contemporary criminology is built.

The future of differential association theory lies in its continued application to emerging forms of crime — including cybercrime, online radicalization, and corporate fraud in globalized economies — and in the further integration of its propositions with developmental, contextual, and cognitive perspectives that can account for the individual-level variation and the dynamic, reciprocal processes that Sutherland’s original formulation did not fully address. Nearly nine decades after its initial presentation, differential association remains one of the most consequential ideas in the history of criminological thought.

References

Akers, R. L. (1998). Social learning and social structure: A general theory of crime and deviance. Northeastern University Press.
Bruinsma, G. J. N. (1992). Differential association theory reconsidered: An extension and its empirical test. Journal of Quantitative Criminology, 8(1), 29–49. https://doi.org/10.1007/BF01062758
Burgess, R. L., & Akers, R. L. (1966). A differential association-reinforcement theory of criminal behavior. Social Problems, 14(2), 128–147. https://doi.org/10.2307/798612
Cloward, R. A., & Ohlin, L. E. (1960). Delinquency and opportunity: A theory of delinquent gangs. Free Press.
Cohen, A. K. (1955). Delinquent boys: The culture of the gang. Free Press.
Cressey, D. R. (1953). Other people’s money. Free Press.
Cressey, D. R. (1960). Epidemiology and individual conduct: A case from criminology. Pacific Sociological Review, 3(2), 47–58. https://doi.org/10.2307/1388319
Dearden, T. E., & Parti, K. (2021). Cybercrime, online peer association, and social learning. American Journal of Criminal Justice, 46(3), 442–460. https://doi.org/10.1007/s12103-020-09582-8
De Fleur, M. L., & Quinney, R. (1966). A reformulation of Sutherland’s differential association theory and a strategy for empirical verification. Journal of Research in Crime and Delinquency, 3(1), 1–22. https://doi.org/10.1177/002242786600300101
Dishion, T. J., McCord, J., & Poulin, F. (1999). When interventions harm: Peer groups and problem behavior. American Psychologist, 54(9), 755–764. https://doi.org/10.1037/0003-066X.54.9.755
Elliott, D. S., Ageton, S. S., & Canter, R. J. (1979). An integrated theoretical perspective on delinquent behavior. Journal of Research in Crime and Delinquency, 16(1), 3–27. https://doi.org/10.1177/002242787901600102
Gallupe, O., McLevey, J., & Brown, S. (2018). Selection and influence: A meta-analysis of the association between peer and personal offending. Journal of Quantitative Criminology, 35(2), 313–335. https://doi.org/10.1007/s10940-018-9384-y
Holt, T. J., Burruss, G. W., & Bossler, A. M. (2010). Social learning and cyber-deviance: Examining the importance of a full social learning model in the virtual world. Journal of Crime and Justice, 33(2), 31–61. https://doi.org/10.1080/0735648X.2010.9721287
Kornhauser, R. R. (1978). Social sources of delinquency: An appraisal of analytic models. University of Chicago Press.
Matsueda, R. L. (1982). Testing control theory and differential association: A causal modeling approach. American Sociological Review, 47(4), 489–504. https://doi.org/10.2307/2095194
Matsueda, R. L. (1988). The current state of differential association theory. Crime and Delinquency, 34(3), 277–306. https://doi.org/10.1177/0011128788034003005
McGloin, J. M., & Thomas, K. J. (2019). Peer influence and delinquency. Annual Review of Criminology, 2, 241–264. https://doi.org/10.1146/annurev-criminol-011518-024551
Opp, K.-D. (1974). Abweichendes Verhalten und Gesellschaftsstruktur. Luchterhand.
Reiss, A. J., & Rhodes, A. L. (1964). An empirical test of differential association theory. Journal of Research in Crime and Delinquency, 1(1), 5–18. https://doi.org/10.1177/002242786400100102
Shaw, C. R., & McKay, H. D. (1942). Juvenile delinquency and urban areas. University of Chicago Press.
Short, J. F. (1957). Differential association and delinquency. Social Problems, 4(3), 233–239. https://doi.org/10.2307/798904
Sutherland, E. H. (1939). Principles of criminology (3rd ed.). Lippincott.
Sutherland, E. H. (1947). Principles of criminology (4th ed.). Lippincott.
Sutherland, E. H. (1949). White collar crime. Dryden Press.
Sutherland, E. H., Cressey, D. R., & Luckenbill, D. F. (1992). Principles of criminology (11th ed.). General Hall.
Thomas, W. I., & Znaniecki, F. (1918–1920). The Polish peasant in Europe and America (5 vols.). University of Chicago Press.
Thornberry, T. P. (1987). Toward an interactional theory of delinquency. Criminology, 25(4), 863–892. https://doi.org/10.1111/j.1745-9125.1987.tb00823.x
Thornberry, T. P., Krohn, M. D., Lizotte, A. J., Smith, C. A., & Tobin, K. (2003). Gangs and delinquency in developmental perspective. Cambridge University Press.
Warr, M. (2002). Companions in crime: The social aspects of criminal conduct. Cambridge University Press.
Wolfgang, M. E., & Ferracuti, F. (1967). The subculture of violence: Towards an integrated theory in criminology. Tavistock.