Social Learning Theory in Criminology Research Paper

This sample Social Learning Theory in Criminology Research Paper is published for educational and informational purposes only. If you need help writing your assignment, please use our research paper writing service and buy a paper on any topic at affordable price. Also check our tips on how to write a research paper, see the lists of criminal justice research paper topics, and browse research paper examples.

In their differential association-reinforcement theory, Burgess and Akers (1966) reformulated, and restated, Edwin Sutherland’s principles of differential association (1947) to include the learning principles of modern behavioral learning. Sutherland’s theory had been roundly criticized on two main grounds: (1) its failure to make explicit the precise mechanisms of learning and (2) difficulties operationalizing and empirically testing the main ideas of the theory leading some to assert that the theory was unfalsifiable (Nettler 1984). Akers (1973, 1998) subsequently revised differential association-reinforcement theory into a social learning theory of crime and deviance that included four key constructs: differential association, definitions, differential reinforcements, and imitation. In the most general terms, Akers’ social learning theory asserts that both conforming and deviant behavior are learned in the same way, that one’s likelihood of engaging in antisocial behavior is influenced by his/her prior and anticipated consequences of behavior, or that the probability of engaging in deviant behavior is contingent on the prior and anticipated future reinforcements and punishments one has experience regarding deviance.

Social learning theory remains a dominant perspective in criminology, and the empirical literature testing this theory is arguably one of the largest and most persuasive in the field (see Warr 2002). This research paper provides a general overview of the assumptions, constructs, and propositions that are central to social learning theory, summarizes the empirical literature testing this perspective, addresses some of the key criticisms, and concludes with new directions for the theory.


In 1965, C. Ray Jeffrey recommended that Sutherland’s differential association theory be replaced with a single statement drawing on the principles of modern behavioral learning. Specifically, Jeffrey (1965, p. 295) noted that criminal behavior, like all other human behavior, is governed by the principles of differential reinforcement such that it “occurs in an environment in which in the past the actor has been reinforced behaving in this manner, and the aversive consequences attached to the behavior have been of such a nature that they do not control or prevent the response.” The importance of Jeffrey’s recommendation cannot be overstated, as he correctly pointed out that criminologists from the learning paradigm had neglected to incorporate the significant advancements in behavioral psychology into the understanding of crime since Sutherland’s original contributions.

Around the same time, Burgess and Akers were in the process of integrating these modern learning principles into the differential association perspective. They believed that all of the explanatory variables and processes in differential association theory could be subsumed under a more general theory of human behavior that is based on behavioral learning principles. Burgess and Akers accomplished this by revising each of Sutherland’s nine statements to form seven new propositions that invoke language from behavioral psychology and modern learning. For instance, the first and eighth propositions in Sutherland’s differential association theory which read “Criminal behavior is learned” (p. 6) and “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” (p. 7), respectively, were combined into a single proposition that read “Criminal behavior is learned according to the principles of operant conditioning” (Burgess and Akers 1966, p. 132). Further, the key proposition of Sutherland’s theory which read “A person becomes delinquent because of an excess of definitions favorable to violation of law over definitions unfavorable to violations of law” was similarly revised in a manner consistent with behavioral learning, such that “Criminal behavior is a function of norms which are discriminative for criminal behavior, the learning of which takes place when such behavior is more highly reinforced than non-criminal behavior” (p. 143).

The initial formulation of social learning theory in criminology was heavily influenced by the behavioral learning paradigm developing in psychology (Bandura 1969). In this seminal work, Burgess and Akers discussed the importance of things like discriminative stimuli (signals that elicit a particular behavioral response), saturation (when the potency of a reinforcer is maximized and no longer affects behavior), and extinction (the reduction or loss of a behavioral response when reinforcement is withheld) when describing the learning process. The complexity inherent in this perspective, however, led many criminologists to question the utility and testability of this learning perspective. These concerns, along with the movement towards a variable-based criminology (Hirschi 1969), led Akers to revise social learning theory into a more tenable set of constructs and propositions.

Akers’ Revised Social Learning Theory

Social learning theory is developed to explain not just the acquisition of criminal behavior but also the maintenance and modification of criminal behavior (Akers 1998). Put simply, the process of social learning offers not just explanations on the motivation to initially engage in criminal behavior but also explanations on how certain variables work to maintain criminal behavior throughout time and how these variables also can maintain and promote conformity over time. Akers (1998, p. 50) argues that “The basic assumption in social learning theory is that the same learning process… produces both conformity and deviant behavior. The differences lie in the direction.. .. [of] the balance of influences on behavior.” Thus, individuals are at an increased likelihood of engaging in deviant behavior when they differentially associate with others who commit criminal behavior themselves, who define deviance as desirable, and who provide (or have provided) greater reinforcement than punishment for that deviance. Conversely, individuals are more likely to be prosocial to the extent that they differentially associate with individuals who engage in and espouse norms conducive to conformity.

When developing his revised social learning theory, Akers asserts that the social learning process can be principally captured by four distinct, yet interdependent, constructs: differential association, definitions, differential reinforcement, and imitation. The subsequent sections of this research paper will elaborate on these four constructs and discuss how they influence the probability of engaging in criminal and deviant behavior.

Differential Association

Social learning theory begins with the assumption that all learning takes place in a context of structural interactions with others. The most important social context in which mechanisms of learning operate is through the differential associations individuals have. This construct draws heavily on Sutherland’s theory, which emphasizes the importance of “intimate personal groups” in the learning process.

Akers asserts that differential association can influence delinquency both through behavioral/ interactional dimensions and normative dimensions. The former refers simply to the direct association and interactions individuals have with others and whether the behavioral patterns that these associations expose to an individual are criminal or conforming. In other words, it refers to the behavioral nature of one’s associates (e.g., deviant vs. non-deviant). The latter, however, are the different patterns of norms and values that an individual is exposed to through these associations. Sutherland’s discussion of differential association was primarily concerned about the transmission of definitions that occurred in these differential interactions, but Akers essentially allows the differential association construct to incorporate a much broader range of deviant influences present when interacting with others, including the transmission of deviant definitions, providing reactions that are favorable to delinquent conduct that alters the perceived rewards/ costs of crime and exposing individuals to deviant models from which to imitate.

Like Sutherland, Akers’ social learning theory does not view all interactions as being equally influential in affecting human behavior. Instead, Akers retains the four modalities discussed by Sutherland, asserting that associations, and their subsequent influence on deviant behavior, vary in terms of frequency, duration, priority, and intensity. Frequency is how often an individual interacts with a group or individual. Duration concerns the length and relative amount of time spent in different associations. Priority refers to when in time the association began, in other words, which associations came earliest in time. Finally, intensity has to do with the “significance, saliency, or importance of the association to the individual” (Akers 1998, p. 64). Drawing on these four modalities, Akers (like Sutherland) argues that associations will have a greater impact on shaping an individual’s behavior if they occur more frequently, if they have been interacting for longer periods of time, if they began earlier in the individuals life, and if they are more meaningful relationships to the individual.

This construct of differential association subsumes many of the major elements of Sutherland’s differential association theory. Sutherland’s recognition that the interactions that individuals have when associating with others is central to the learning process is acknowledged by Akers, and he continues to stress the importance that different modalities have on influencing how these associations shape behavior. In fact, given Akers’ assertion that most learning is shaped by social reinforcement contingencies – rather than nonsocial ones (e.g., intrinsic rewards/ punishments) – differential association is a necessary condition in the social learning perspective.


Sutherland’s emphasis on the transmission and acquisition of definitions favorable to delinquency when explaining deviant behavior made it difficult to distinguish that construct from differential associations. However, Akers explicitly separates these two constructs in his social learning theory. Despite this, Akers’ characterization of definitions is nearly identical to that described by Sutherland. Definitions can generally be thought of as the normative beliefs, attitudes, or meanings that individuals attach to any given behavior. These can include the motives, drives, rationalizations, justifications, and attitudes that define the circumstances in which the commission of an act is right or wrong, desirable or undesirable, justified or unjustified (Akers 1998, p. 77). In social learning theory, definitions can be both general and specific. General definitions refer to the broad religious and moral beliefs and norms one holds that favor either criminal or conforming behavior. Specific definitions refer to specific acts and situations in which behavior is, at minimum, justified. The distinction between general and specific definitions is of theoretical importance. Individuals may generally hold the belief that “breaking the law is wrong” but at the same time believe that “smoking marijuana is OK.” Similarly, individuals may hold the general definition that “thou shall not steal” while simultaneous believing “it is morally justifiable to steal when hungry.”

Definitions in Akers’ theory do not require deviant behavior, as some scholars have wrongly interpreted in Sutherland’s theory (Kornhauser 1978). Akers (1998) argues that some definitions favorable to delinquency can be so strongly held that they require delinquency. More common, however, is that individuals hold definitions that rationalize or justify delinquency in certain situations. In this way, definitions in Akers’ theory are similar to the belief construct in Hirschi’s (1969) social control theory, in that the definitions themselves do not directly provide motivation for delinquency but rather are conventional beliefs that are weakly held and can allow for other motivating factors (e.g., differential reinforcement) to have important and direct effects on antisocial behavior. That is to say, their moral reservations against delinquency are weak enough that it can be bypassed in some situations and thus allow for delinquent behavior when the right circumstances or reinforcement contingencies are present (Warr 2002).

Differential Reinforcement

Arguably the construct of greatest importance in Akers’ social learning theory is differential reinforcement. Though Sutherland stated that criminal behavior is learned like all forms of human behavior, he spent little time discussing the learning process. Drawing on research in psychology, Akers argues that the process of learning antisocial behavior is best captured through the construct of differential reinforcement. “Differential reinforcement refers to the balance of anticipated or actual rewards and punishments that follow or are consequences of behavior” (Akers 1998, p. 67). If an individual believes that they are likely to be rewarded for engaging in deviance and/or punished for deciding upon conforming behavior, then that individual is likely to engage in deviance. Moreover, when faced with a similar situation in the future, that individual is likely to draw on their prior experiences and view that deviant behavioral choice as more reinforcing and the conforming option more punishing. This ultimately leads to what is arguably the most important proposition in Akers’ theory: “Whether individuals will refrain from or initiate, continue committing, or desist from criminal and deviant acts depends on the relative frequency, amount, and probability of past, present, and anticipated rewards and punishments perceived to be attached to the behavior” (Akers 1998, p. 66).

From the social learning perspective, human behavior is governed by both reinforcements and punishments, whereby the former increases the probability that a behavior occurs and the latter decreases it. Both reinforcements and punishments can be either positive or negative, which references the fact that some consequences influence behavior by adding stimuli (positive) while others influence it by removing stimuli (negative). Accordingly, positive reinforcements increase the likelihood of engaging in or repeating deviance by providing a positive outcome or reaction to it (e.g., friends’ approval, money, or intrinsic feelings of pleasure). Negative reinforcements increase the likelihood of deviance by removing an aversive or unpleasant feeling or event (e.g., eliminating the ridicule of classmates). Meanwhile, positive punishments decrease the likelihood of criminal behavior by providing a painful or unpleasant consequence associated with a behavior (e.g., an arrest), and negative punishments affect behavior by removing pleasant consequences or rewards (e.g., parental respect). These reinforcements and punishments influence the likelihood of engaging in deviance by altering an individual’s perceived anticipated consequences when faced with similar situations in the future, which in turn affect one’s decision-making calculus at the time of the offense (Akers 1990, 1998).

Just as with differential associations, reinforcement and punishment contingence are not all weighted equally but are influenced by the amount, frequency, and probability of their occurrence. The greater the amount or value placed on the reinforcement contingencies, the more frequently it is reinforced, and the greater the likelihood of experiencing the reinforcement all increase the likelihood of engaging in deviant behavior. Thus, for instance, if an individual places great value on the relationship she shares with her friend who frequently rewards deviant behavior, and she is certain that her friend will reward her for stealing, that individual will have a high probability of engaging in deviant behavior. However, the reinforcements and punishments that are provided by strangers for whom the individual has no personal relationship with will likely have little effect both on the present and on future perceptions of rewards and punishments.

Though the examples provided above have focused largely on social reinforcements and punishments (i.e., peers and parents), Wood et al. (1997) note that reinforcement contingencies can be nonsocial. For instance, the physiological effects of drugs and alcohol can be rewarding to some, which in turn increases the likelihood of repeated use. Moreover, some individuals may view deviant behavior itself as intrinsically rewarding. This would occur if the behavior itself, regardless of social reactions, is viewed as “fun” or “thrilling.” Wood et al. provide support for this premise: positive intrinsic sensations from crime are significantly related to the maintenance of criminal behavior. Interestingly, Akers vests little in the idea that nonsocial factors are inherently reinforcing. For instance, he argues that individuals learn that the effects of drugs are enjoyable or unpleasant through social reinforcement. Similarly, Akers also argues that some individuals are prone to be risk-seeking but that they learn to become risk-seeking, and learn to see delinquency as “fun,” through social interactions with others. In other words, individuals may experience physiological and intrinsic rewards from drug use and crime but primarily only do so if they are predisposed by a social learning history to do so.

Akers believes that the primary source of learning occurs through social reinforcements and punishments. When individuals value the relationships they share with others, the prior, anticipated, and actual reactions that they attach to deviant behavior are likely to be highly influential in affecting the likelihood that an individual offends. Importantly, this is not the same as having peers pressuring individuals to offend, but rather this process refers to the influence that these friends have by altering the perceived consequences of different behavioral choices. Social learning theory recognizes that humans are social beings who are concerned with the reactions and opinions of others and that our perceptions on what the anticipated consequences of behavior entail are necessarily a function of prior experiences in similar situations. Even tangible and material items, such as money and wealth, are valued by individuals because of the social rewards and prestige that they provide. Put simply, Akers argues that tangible and material rewards from crime derive their power of influence because they are symbolic of social rewards (i.e., prestige and status) and that their value (and thus their influence on behavior) is of little importance when the social rewards are disassociated from them.

In sum, the inclusion of differential reinforcement into the learning paradigm of criminology filled the important gap as to how criminal behavior is learned (i.e., a focus on the “reception” and not the “transmission” process). Though this contribution by Akers is arguably the most important aspect in social learning theory, it also remains one of the most misunderstood. The idea that individuals are affected by reinforcement contingencies does not mean that humans are passively conditioned by their environment, simply reacting to stimuli like Pavlovian dogs. On the contrary, social learning theory views that individuals are mostly actively engaged and attuned to the anticipated consequences of behavioral actions (Akers 1990), which presupposes that individuals employ some cognition before engaging in delinquency. If an individual determines that the relative balance of the consequences of deviance provides greater reinforcements than punishments, then that individual is at an increased likelihood of engaging in deviant behavior. Moreover, social learning theory argues that the actual consequence attached to the behavior will affect their future likelihood of offending: if an individual engages in antisocial behavior and they are rewarded for it, they are at an increased likelihood of repeating the action when faced with a similar situation in the future. Conversely, if individuals engage in deviance but experience a relatively greater level of negative consequences, they are likely to avoid deviant behavior in the future. Simply put, social learning theory argues that an individual’s prior experiences influence how they interpret anticipated consequences in the future and that these anticipated consequences are influential when deciding on a behavioral action choice (Akers 1998).


Imitation refers to the modeling of behavior after similar behavior is observed in others. The classic example of imitation derives from Bandura’s (1961) Bobo doll experiments, where the behavior of children was observed after watching a model aggress on an inflatable doll. As is evident in these experiments, not all behavior that is observed is imitated by others. Of particular importance is the observed consequence of the behavior: if one observes an individual engage in a behavior and that individual is subsequently rewarded for the behavior, it is more likely that that behavior will be modeled. This is known as vicarious reinforcement. Conversely, if one observes behavior that is subsequently punished, it is less likely that the behavior will be modeled – what is known as vicarious punishment.

Another factor that influences whether an observed behavior is modeled is the characteristic of the model. In this way, imitation shares some of the modalities as the other constructs in social learning theory. The most salient models are those persons “with whom one is in direct contact in primary groups” (Akers 1998, pp. 76–77). This most obviously means that individuals are more likely to imitate the behavior of close, personal associates than that of strangers. But this statement also implies that individuals are more likely to model behavior observed in direct interaction than behavior observed through media forums. Akers (1998) has argued that technological and social changes have increased the relative importance of modeling behavior on television, movies, and video games. More specifically, given that the relative time spent watching television and movies and playing video games has increased over the last several decades, the modalities of learning have also increased (frequency, duration, priority, and intensity), making media models more salient.

Intuitively, the imitation of others’ behavior is most influential in the acquisition and performance of novel behavior but plays little role in the maintenance of deviant behavior. The continuation of the behavior is more affected by direct reinforcements and punishments. For example, an individual may learn initially that violence is rewarded when one’s status is challenged, but whether they continue to behave in that matter will be contingent on their own experiences with such behavior. However, imitation may still play a role in the acquisition of criminal skills that can make crime easier and less risky.

Though these constructs are treated and discussed independently, there is considerable interdependency among them. Indeed, social learning is itself a process that incorporates all of these constructs in a complex and recursive manner. Initial deviance is influenced by the balance of definitions one has been socialized to hold, by the modeling of others’ behavior, and/or by the anticipated rewards and punishments that derive from a criminal event. Immediately after this event, the actual consequences of the action play an important role – if the individual experiences some reward, then he/she is likely to believe there will be positive consequences in the future. Further, if one does experience reinforcement, this will likely alter the definitions they hold and, in turn, could potentially alter the individual’s criminal and non-criminal associates (i.e., self-selection). Indeed, social learning theory does not claim that associations are random but predicts that individuals do self-select into peer networks. However, social learning theory does predict that more often than not delinquent associations precede delinquency.

Research On Social Learning Theory

Few theories have been tested empirically to the extent that social learning theory has. These tests most often assess the relationship and direction between variables from the social learning perspective – differential association, definitions, differential reinforcement, and imitation – and some deviance outcome measure (i.e., general delinquency, drug use, violence). The overwhelming majority of these studies have found strong support for social learning predictions, with very few studies finding negative support for the theory.

One of the earliest and most seminal tests of the theory was conducted by Akers and his colleagues in 1979. Using substance use as an outcome, the results found that differential association (i.e., delinquent friends), definitions (how right/wrong substance use is), and differential reinforcement (how parents and friends would react to respondent drug use) were all strongly and statistically related to substance use, providing support for the social learning predictions. Imitation, however, was only weakly and inconsistently related to substance use, though Akers et al. (1979) note that this finding is not unexpected, as imitation is only predicted to have strong effects on the initiation of delinquent behavior. Though the research findings were strongly supportive of the theory, the clear operationalization of the theory into testable measures and predictions within the article is an equally important contribution. Since this work, many empirical studies have assessed the empirical validity of social learning, with most supportive of the perspective (Pratt et al. 2010).

Another common manner in which research has inferred the empirical validity of social learning theory is through an assessment of the relationship between delinquent peers and offending. As would be predicted by social learning theory, the relationship between peers and delinquency is one of the strongest and most consistent findings in all criminological research (Warr 2002). More frequent, longer-term, and closer associations with peers who engage in delinquent behavior is consistently related to the delinquency or respondents, while associating with prosocial peers is consistently related to conformity. This strong relationship has generally held using various samples and methodological specifications. Ultimately, these findings have led some scholars to conclude that delinquent peers are one of, if not, the most important causes of adolescent delinquency (Warr 2002) and provide considerable support for both the differential association and social learning perspectives.

Perhaps the most telling support for social learning theory comes in two forms: its relative superiority when faced with theoretical competition and its ability to explain known correlates of offending. Regarding the former, many empirical works in the field of criminology have sought to pit variables from one theoretical perspective against those from others. Most generally, this involves putting learning variables in statistical models and comparing the size of the effects against variables from control and strain theories. When this is done, the variables from learning theory tend to explain more of the variance in delinquency than variables from other perspectives (Warr 2002). Learning variables also do an impressive job explaining some of the well-known correlates of offending. For instance, Warr (1993, 1998) has demonstrated that associating with delinquent peers can explain both the age and marriage effect. Similarly, Mears et al. (1998) have shown that learning theory also explains a substantial portion of both race and gender differences in crime, respectively.

The empirical support for social learning theory has most recently been summarized in a meta-analysis conducted by Pratt and colleagues (2010). This study reviewed the published extant tests of the perspective to provide a synthesized estimate of the effect size of the social learning theory constructs. In reviewing these studies, Pratt et al. conclude that the empirical support for social learning theory is stronger than many other criminological perspectives and is at least as comparable as the most strongly supported theories (see Pratt and Cullen 2000). Interestingly, however, the different constructs of social learning theory appear to have varying levels of explanatory power. The constructs most strongly related to deviance are differential associations and definitions, the two constructs already captured in Sutherland’s theory. Differential reinforcements and imitation are only modestly related to delinquency. If indeed differential reinforcement is the most important construct in the social learning perspective, these findings raise some questions about Akers’ perspective and provide researchers with an important avenue to further assess the interdependence of the four constructs of social learning theory.

An extension of social learning occurred in 1998 when Akers (1998) suggested that an integrated theory (social structure and social learning) could explain structural variations in crime rates across communities. Although this premise has yet to receive much empirical attention, the basic argument is that social structure is linked to individual behavior – and in turn, crime rates – through its effects on social learning variables, namely, differential association, differential reinforcement, definitions, and imitation. Structural characteristics provide the contexts within which social learning variables operate: it causes individuals to differentially associate with delinquents and exposes individuals to a greater ratio of deviant reinforcement contingencies and to definitions favorable to delinquency, as well as a greater number of delinquent models to imitate.

Four dimensions make up the SSSL perspective. Differential social organization refers to the distribution of structural correlates of crime (age, race, gender, SES) in the community or society that affect rates of delinquency. Differential location in the social structure refers to one’s own social and demographic location in the social structure that affects social learning variables (e.g., younger individuals are more likely to be exposed to other younger individuals who, in turn, are more crime prone). Theoretically defined criminogenic aspects of social structure draws on anomie, social disorganization, and conflict theories which, again, Akers argues affects one’s exposure to criminal associates, definitions, models, and reinforcements that can promote deviant behavior. Finally, differential social location in primary, secondary, and reference groups refers to the more immediate personal networks that filter the larger social environment to either promote or discourage criminal behavior. In sum, the general premise of SSSL is that social learning variables fully mediate social structural characteristics correlated with high rates of crime. To date, however, this perspective has not fully been tested by researchers, and the ability of social learning theory to explain differences in crime rates remains in question.


Pratt and colleagues’ (2010) finding that differential reinforcement and imitation had relatively smaller effects in predicting delinquency when compared to differential association and definitions can be seen as challenging to Akers’ contributions to the learning paradigm, but several other criticisms also plague the theory. Most notable, several scholars have challenged the supposed causal relationship between peers and delinquency that is hypothesized by learning theorists. The argument is that the strong correlation between delinquent peers and delinquency is not evidence that differential associations cause delinquency but instead reflects the tendency for “birds of a feather to flock together” (Gottfredson and Hirschi 1990). This “selection” argument has more recently been championed most fervently by scholars from the control perspective who assert that both delinquency and the association with delinquent peers are both outcomes of similar causal factor (i.e., weak attachments to society or low self-control). The criticism is that delinquent associations are not random, but instead individuals self-select into friendship networks based on delinquent tendencies. Thus, proponents of the selection perspective argue that the relationship between peers and delinquency is spurious, as delinquent behavior is thought to actually precede delinquent peer associations.

Though often seen as one of the greatest challenges to social learning theory, this criticism falls short for both theoretical and empirical reasons. First, social learning theory does not predict that peer associations are randomly constructed but instead explicitly acknowledges that individuals self-select into peer groups based on their prior learning. Unlike control theories, however, social learning theory predicts that once this selection does occur, these associates are likely to have a continued influence on delinquent behavior through differential reinforcement and other learning mechanisms. The empirical research is supportive of the learning predictions: even when statistically controlling for competing variables that are thought to render the peerdelinquency relationship spurious (Gottfredson and Hirschi 1990), the association between peers and delinquency remains (Matsueda and Anderson 1998). Associations with delinquent peers most often precede delinquent behavior (Elliott and Menard 1996), which is supportive of the predictions of social learning theory. Still, this selection versus socialization debate is far from settled (see Weerman 2011). Future research using experimental and quasiexperimental designs will likely shed more light on this issue.

Another criticism of social learning theory concerns the types of deviant behaviors in which the theory is able to explain. Many tests of social learning theory have assessed whether Akers’ theory can explain variation in substance use and other forms of minor delinquency. Much less research has examined how well social learning theory explains more serious forms of delinquent behavior, and the research that does finds that learning theory does a poorer job explaining these more serious forms of delinquency. The reason behind this remains relatively unexplored, but one possibility is that individuals who engage in more serious forms of violence have such high levels of criminal propensity (i.e., high in impulsivity) that they do not require learning mechanisms to engage in delinquent behavior. In any case, the finding that social learning has more trouble explaining more serious forms of delinquency is problematic for the theory, as Akers’ perspective is designed to explain all forms of crime and deviance.

Finally, following Kornhauser’s critique (1978, p. 34) of differential association/social learning theory assumptions that “man has no nature, socialization is perfectly successful, and cultural variability is unlimited,” social learning theory has been interpreted by some as a theoretical perspective unable to explain individual differences and only applicable to group differences in crime where subcultural definitions adhere to deviance (Akers 1996, p. 230). A corollary is that the theory focuses entirely on the development of novel criminal behavior that, once learned, requires deviance. In essence, critics see little to differentiate differential association and social learning from cultural deviance theories. Akers (1996), in an exchange with Travis Hirschi on this subject (Hirschi 1996), suggests that these criticisms stem from ambiguities associate with and misrepresentations of Sutherland’s statement of the differential association – problems that Akers (1996, p. 238) believes have been rectified in his specification of social learning principles by allowing for variations in both deviant definitions held by individuals and the social reinforcement they receive. “These deviations can develop within the same normative or cultural system. They do not require the existence or participation in an organized deviant subculture.”

New Directions

Despite these criticisms, social learning theory remains one of the preeminent theories of crime, and the considerable empirical support for this perspective seems to justify this prominence. Nevertheless, there are some important questions deriving from social learning theory that have been left unexplored.

Saturation Effects and Discriminative Stimuli. In describing a social learning theory of crime and deviance, Akers (1998; see also Burgess and Akers 1966) invokes language from psychology that describes how learning differs across time and situations. Most notably are the concepts of saturation and discriminative stimuli. Saturation occurs when the potency of a reinforcer has been maximized, and that reinforcer no longer has an effect on behavior. This has clear implications for the structure of peer networks: some individuals may no longer be influenced by associating with delinquent associates, as their prior learning and already existing associates may have already hit a saturation point. Put simply, the relationship between delinquent associates and delinquency may be nonlinear, increasing at a decreasing rate. Zimmerman and Vasquez (2011) have assessed this saturation hypothesis at a neighborhood level. Their results provide support for the saturation effect: delinquent peers have a strong effect for individuals residing in non-disadvantaged neighborhoods but a diminishing effect in disadvantaged ones. Still, there is considerable research left unexplored on the saturation effect at the more proximal friendship network level and over the life course. How does the inclusion of a delinquent peer change the delinquency of an individual in a relatively nondelinquent network versus a relatively highly delinquent network? Further, do peers have the same effect over the life course? Or do individuals become saturated from delinquent peers leading to a declining peer effect as one ages?

The discriminative stimuli effect has been relatively neglected in criminological research (however, see McGloin et al. 2011), despite the direct implications it has for theory. Discriminant stimuli refer to stimuli that influence how an individual will act in a given situation. Discriminant stimuli are most important when they signal which situations and circumstances will yield reinforcement. To be clear, it is highly unlikely that individuals view that their friends will reward violence, theft, or other forms of deviance in all situations. Rather, it is more common that an individual’s delinquent friends will reward it in some situations and under some circumstances. Discriminative stimuli serve as the cues as to when these situations and circumstances occur. The same logic applies in understanding one’s definitions favorable/unfavorable to delinquency. As noted above, most delinquents do not hold the universal values that “deviance is right” but rather hold definitions that justify delinquency in some situations (Akers 1998). Unfortunately, most measures tapping into peer reactions and individual definitions use global statements on the rightness and wrongness of antisocial behavior. Considerable theoretical insight can be gained by beginning to incorporate this discriminative stimuli logic into learning measures and assessing what situations and circumstances allow individuals to justify deviant behavior. The overall point is that there remains important elements from the more general behavioral learning paradigm that have not been explored in criminological research that can shed some important light on the understanding of crime.

Differential Susceptibility: How and When Learning Variables Matters. Just as learning variables may not have ubiquitous effects across all situations, there may also be variability among individuals in this regard. For generations the primary question surrounding social learning theory has been the central question of causality: do peers and other associates (e.g., family) hold causal role in the facilitation of delinquency? As noted above, the overwhelming evidence supports the predictions from social learning theory. At this point, it seems worthwhile to move beyond simple assessments of if learning variables influence delinquency and begin to assess if they matter differently for different people.

Take, for instance, the relationship between delinquent peers and delinquency. Akers suggests that delinquent peers influence delinquent behavior by providing positive reactions to such behavior. He goes on to specifically acknowledge that the influence of such reactions is necessarily dependent on the ability of individuals to consider the longer-term consequences of behavior; in other words, it requires individuals to reflectively consider how their peers would react before deciding to offend (Akers 1990). If this is the case, then some individuals may not be susceptible to the risk of associating with delinquent peers. In particular, individuals high in impulsivity, who are characterized by their tendency to make decisions based almost entirely on situational stimuli, may not need or consider the reactions of their peers when deciding to offend. Instead, these individuals will be most susceptible to the situational risk of unstructured socializing with peers (Osgood et al. 1996). This hypothesis was supported by Thomas and McGloin (2013) using three different large-scale data sets: delinquent peers have their strongest effects on individuals low in impulsivity but only weak (and nonsignificant) effects on individuals high in impulsivity. Further, the situational risk of unstructured socializing with peers has its strongest effects on this high in impulsivity, but a weak (and nonsignificant) effect on this low in impulsivity. Thus, some individuals may be particularly susceptible to the reactions of their peers, while others may not need such reactions to motivate them to deviate. The larger point is that researchers should begin to think more critically about the mechanisms by which learning variables facilitate delinquency. These more nuanced considerations can shed light on important considerations as to how and when peers and other associates facilitate delinquent behavior. Do all individuals rely on delinquent associates to deviate? How and when do individuals consider their own definitions before deciding to offend? Indeed, the question as to if social learning variables are related to crime and deviance appears to have been answered, and researchers should begin to assess whether these variables matter differently for different people.

Objective versus Perceptual Measures. An area of growing interest among criminologists deals with the most appropriate way to measure variables of theoretical interest, and social learning theory has been particularly in the crosshairs of this discussion. Several scholars have expressed concerns regarding the adequacy of traditional perceptual measures of delinquent peers, whereby respondents are asked to report on the behavior of their friends. Most critically, critics have suggested that perceptual measures have had a tendency to overestimate the delinquent peer effect, because individuals are not accurately reporting the behavior of their friends, but instead are merely projecting their own behavior to them (Gottfredson and Hirschi 1990). The limited research that has been conducted is supportive of this claim: when objective measures of peer deviance are used, the relationship between peers and delinquency is reduced, though still substantively large. These findings have led many scholars of crime to encourage the use of objective measures over the use of perceptual ones while outright rejecting the utility of the latter measures because they reflect an individual’s misperceptions of their friends’ behavior and not the reality of it.

It is interesting that this movement has occurred during the same time period in which rational choice and deterrence theorists have advocated the use of perceptual measures over objective ones when assessing criminal decision making. After decades of empirical work challenged the core propositions of rational choice theory when using objective measures, scholars began to realize that sanction threats can only influence behavior if they are perceived by the individual. The crux of their argument was that these choice theories are perceptual in nature and, thus, require scholars to assess one’s perceptions of their beliefs of risk rather than utilizing the objective reality of that risk. The positive research using these perceptual measures has led to some resurgence in the rational choice perspective (Nagin 1998). These positive findings has led Nagin (1998) to argue that scholars interested in the choice perspective should move beyond assessing if risk perceptions influence behavior and focus on how individuals actually formulate their perceptions of risk, in other words, what information (or “signals”) go into the development of subjective risk perceptions.

Interestingly, social learning theory is also an inherently perceptual theory of crime. The core proposition of the theory is that the perceived reinforcements and punishments associated with behavior influence the present and future likelihood of offending. Given this strong emphasis on perceived reinforcements and punishments, and the robust relationship between these perceptions and offending, one future direction for social learning scholars is to heed Nagin’s (1998) call and end the discussion on the superiority of objective measures over perceptual ones and begin to consider what factors go into one’s perceptions of social reinforcements and punishments. To be clear, a recent study conducted by Young et al. (2011) found that individuals have a tendency to misperceive their own friends’ delinquent behavior. Of course, this finding could be used as more ammunition for the objective-measure-superiority debate, but it also can be used to ask more interesting questions regarding individual perceptions of reinforcements: what causes individuals to misperceive their friends’ behavior? Or more generally, how do individuals formulate their subjective perceptions of social reinforcements? The answer to these questions has considerably more theoretical utility and can shed important light on the learning process, in general.


Social learning theory begins with the assumption that criminal behavior, like all human behavior, is learned in the interaction with others. This theory went beyond Sutherland’s perspective by more clearly specifying what this learning process entailed. Simply, social learning theory asserts that the probability an individual commits deviant acts depends on the relative frequency, amount, and probability of past, present, and anticipated rewards and punishments perceived to be attached to the behavior. The empirical literature supportive of the theory is one of the most impressive in the field of criminology. Nevertheless, considerable avenues of ongoing and future research can lead to more theoretical clarity.


  1. Akers RL (1973) Deviant behavior: a social learning approach. Wadsworth, Belmont
  2. Akers RL (1990) Rational choice, deterrence, and social learning theory: the path not taken. J Crim Law Criminol 81:653–676
  3. Akers RL (1996) Is differential association/Social learning cultural deviance theory? Criminology 34:229–247
  4. Akers RL (1998) Social learning and social structure: a general theory of crime and deviance. Northeastern University Press, Boston
  5. Akers RL, Krohn MD, Kaduce LL, Radosevich M (1979) Social learning and deviant behavior: a specific test of a general theory. Am Sociol Rev 44:635–655
  6. Bandura A (1969) Principles of behavior modification. Holt Rinehart and Winston, New York
  7. Bandura A, Ross D, Ross SA (1961) Transmission of aggression through imitation of aggressive models. J Abnorm Soc Psychol 63:575–582
  8. Burgess RL, Akers RL (1966) A differential association reinforcement theory of criminal behavior. Soc Probl 14:128–147
  9. Elliott DS, Menard S (1996) Delinquent friends and delinquent behavior: temporal and developmental patterns. In: Hawkins JD (ed) Delinquency and crime: current theories. Cambridge University Press, New York, pp 28–67
  10. Gottfredson M, Hirschi T (1990) A general theory of crime. Stanford University Press, Palo Alto
  11. Hirschi T (1969) Causes of delinquency. University of California Press, Berkeley
  12. Hirschi T (1996) Theory without ideas: reply to Akers. Criminology 34:249–256
  13. Jeffrey CR (1965) Criminal behavior and learning theory. J Crim Law Criminol Police Sci 56:294–300
  14. Kornhauser R (1978) Social sources of delinquency. University of Chicago Press, Chicago
  15. Matsueda RL, Anderson K (1998) The dynamics of delinquent peers and delinquent behavior. Criminology 36:269–308
  16. McGloin JM, Schreck CJ, Stewart EA, Ousey GC (2011) Predicting the violent offender: the discriminant validity of the subculture of violence. Criminology 49:767–794
  17. Mears DP, Ploeger M, Warr M (1998) Explaining the gender gap in delinquency: peer influence and moral evaluations of behavior. J Res Crime Delinq 35:251–266
  18. Nagin DS (1998) Criminal deterrence research at the outset of the twenty-first century. Crime and justice: a review of research, vol 23. University of Chicago Press, Chicago, IL
  19. Nettler G (1984) Explaining crime, 2nd edn. McGraw Hill, Boston
  20. Osgood DW, Wilson JK, O’Malley PM, Bachman JG, Johnson LD (1996) Routine activities and individual deviant behavior. Am Sociol Rev 61:635–655
  21. Pratt TC, Cullen FT (2000) The empirical status of Gottfredson and Hirschi’s general theory of crime: A meta-analysis. Criminology 38:931–964
  22. Pratt TC, Cullen FT, Sellers CS, Winfree LT, Madensen T, Daigle LE, Fearn N (2010) The empirical status of social learning theory: a meta-analysis. Justice Q 27:765–802
  23. Sutherland EH (1947) Principles of criminology, 4th edn. Lippincott, Philadelphia
  24. Thomas KJ, McGloin JM (2013) A dual-systems approach to understanding differential susceptibility to different mechanisms of peer influence. Criminology (Forthcoming)
  25. Warr M (1993) Age, peers, and delinquency. Criminology 31:17–40
  26. Warr M (1998) Life-course transitions and desistance from crime. Criminology 36:183–216
  27. Warr M (2002) Companions in crime: the social aspect of criminal conduct. Cambridge University Press, Cambridge
  28. Weerman FM (2011) Delinquent peers in context A longitudinal network analysis of selection and influence effects. Criminology 49:253–286
  29. Wood PB, Gove WR, Wilson JA, Cochran JK (1997) Nonsocial reinforcement and habitual criminal conduct: an extension of learning theory. Criminology 35:335–366
  30. Young JTN, Barnes JC, Meldrum RC, Weerman FM (2011) Assessing and explaining misperceptions of peer delinquency. Criminology 49(2):599–630
  31. Zimmerman GM, Vasquez BE (2011) Decomposing the peer effect on adolescent substance use: Mediation, nonlinearity, and differential nonlinearity. Criminology 49:1235–1273

See also:

Free research papers are not written to satisfy your specific instructions. You can use our professional writing services to buy a custom research paper on any topic and get your high quality paper at affordable price.


Always on-time


100% Confidentiality
Special offer! Get discount 10% for the first order. Promo code: cd1a428655