Social Cognitive Learning Theory (Albert Bandura – 1977)

Bandura described his theory of social learning as being developed in a context in which “the prevailing analysis of learning focused almost entirely on learning through the effects of one’s actions [with] the explanatory mechanisms [cast] in terms of peripheral association of environmental stimuli to responses” (Bandura, 2006a, p. 55). He viewed this type of behavioristic theorizing as “discordant with the obvious social reality that much of what we learn is through the power of social modeling” (Bandura, 2005, p. 10; Bandura, 2006a, p. 55)—noting the absurdity of a culture in which language, customs, practices, occupational competencies, educational practices, religious practices, etc., would be “gradually shaped in each member by rewarding or punishing consequences of their trial-and-error performances” (Bandura, 2005, p. 10; Bandura, 2006a, p. 55). Instead, he posited that in reality the “tedious and potentially hazard process” (p. 11) of trial-and-error learning, which could prove fatal in some cases,[1] are “shortcut by social modeling” (p. 11). Although previous research had been conducted and published in this area, Bandura described his program of research as the first to position social modeling in a context apart from the behaviorist paradigm:

The foremost proponents of behaviorism, Watson (1908) and Thorndike (1898), dismissed the existence of observational learning because, in their view, learning required performance of responses. The notion of learning by observation was too divergent to be given serious consideration. This was a durable legacy. Despite the centrality and pervasiveness of social modeling in everyday life, there was no research to speak of on modeling processes until the publication of Social Learning and Imitation by Miller and Dollard in 1941. They recognized modeling phenomena but construed them as a special case of discrimination learning. A model provides a social cue, the observer performs a matching response, and its reinforcement strengthens the tendency to behave imitatively.

I found this conception wanting on the determinants, mechanisms, and scope of observational learning. We launched a program of research on observational learning as it typically occurs in the absence of reinforced performance. We tested the determinants of observational learning and the mechanisms through which it works.

In a chapter entitled “Vicarious Processes: A Case of No-Trial Learning” (Bandura 1965), I presented the findings of our studies showing that observational learning requires neither response enactment nor reinforcement. (Bandura, 2005)

Bandura’s program of research on observational learning stemmed from his roots in psychotherapy. After receiving his PhD from the University of Iowa in 1952 his first decade of publishing focused on topics including: “‘Primary’ and ‘Secondary’ Suggestibility” (1953); “The Rorshach White Space Response and ‘Oppositional Behavior'” (1954a); “The Rorschach White Space Response and Perceptual Reversal” (1954b); “Psychotherapists’ Anxiety Level, Self-Insight, and Psychotherapeutic Competence” (1956); “Review of Case Studies in Childhood Emotional Disabilities (Vol. 2) by G. Gardner” (Bandura, 1957); “Child-Rearing Patterns Associated With Adolescent Aggressive Disorders” (Bandura, 1958a); “Dependency Conflicts in Aggressive Delinquents” (Bandura, 1958b); “Adolescent Aggression” (Bandura, 1959); and “Psychotherapists’ Approach-Avoidance Reactions to Patients’ Expression of Hostility” (Bandura, Lipsher, & Miller, 1960).

His transition toward a study of observational and social learning began with the publication of  “Psychotherapy as a Learning Process” (Bandura, 1961). At the time this article was published a view of psychotherapy as a learning process was not new, or particularly unique. In fact, Bandura opened by stating that “it is customary [italics added] to conceptualize psychotherapy as a learning process” (p. 143). He went on to say, however, that few therapists who espouse this view “accept the full implications of this position” (p. 143).  In the article he reviewed various principles of contemporary learning theory that might fruitfully be applied to psychotherapy, including counterconditioning, extinction, discrimination learning, reward, punishment, and social imitation. More importantly, however, he framed the central question that he himself would later answer with what is now known as social cognitive theory: “Can human behavior be modified through psychological means and if so, what are the learning mechanisms that mediate behavior change?” (p. 143).

In early research leading up to the definition of social cognitive theory, Bandura provided empirical evidence that children do, in fact, imitate adult models (Bandura & Huston, 1961). Building on his earlier studies of aggression in adults and children, he conducted a series of experiments in which he demonstrated the transmission of aggression from adults to children through delayed imitation of aggressive adult models (Bandura & Ross, 1961). These experiments are now quite famously known as the “Bobo doll” experiments. In similar experiments he showed that imitation of adult behavior could just as effectively be mediated by film, and that there were no significant differences in the effective transmission of aggression via film model as compared to transmission via live model (Bandura, Ross, & Ross, 1963). In other words, violent acts could be taught simply by carrying them out in front of the child, and whether they were presented by a live model or by film made no difference. These findings were unsettling to an already apprehensive public and contributed to growing concern regarding the effects of televised violence on children (Bandura, 2006a, p. 57). They also drew significant opposition to Bandura’s work from the broadcast industry—to the point of airing a dramatized “blistering cross-examination concerning [his] modeling studies” (p. 58) in which the character playing his role did not fare well.

In another study, Bandura and McDonald (1963) found that even the moral judgments of  children “are readily modifiable, particularly through the utilization of adult modeling cues” (p. 280). Three conditions were compared for effectiveness: (a) model plus reinforcement, (b) model without reinforcement, and (c) reinforcement with no model. These three groups were described as follows:

One group of children observed adult models who expressed moral judgments counter to the group’s orientation and the children were positively reinforced for adopting the models’ evaluative responses. A second group observed the models but the children received no reinforcement for matching the models’ behavior. The third group had no exposure to the models but each child was reinforced whenever he expressed moral judgments that ran counter to his dominant evaluative tendencies. Thus the experimental design permitted the test of the relative efficacy of social reinforcement, the behavior of models, and these two factors combined in shaping children’s moral judgments. (pp. 275-276)

From the results of this study Bandura and McDonald concluded that “the provision of models alone…was as effective in altering children’s moral judgments as was the experimental condition with social reinforcement” (p. 279). They also noted that operant conditioning procedures appeared to be “particularly inefficient when there are strong dominant response tendencies and the desired alternative responses are only weakly developed or absent” (p. 281). This they attributed to the relatively infrequent display of the desired responses resulting in “little opportunity to influence them through reinforcement” (p. 281).

In another study Bandura and Mischel (1965) tested “the relative efficacy of live and symbolic models for modifying delay-of-reward behavior [in which] groups of children with marked preferences for either immediate but less valued rewards, or more valuable delay reinforcers, were assigned randomly to one of three experimental conditions” (p. 698, abstract). Children in the first group “observed live models who exhibited delay behavior that was counter to the children’s pattern” (p. 698, abstract). Children in the second group were “presented essentially the same modeling cues except in symbolic verbal form” (p. 698, abstract) via written recording of “the model’s choices and accompanying philosophy-of-life commentaries” (p. 701). Children in the third group “had no exposure to any models” (p. 698, abstract). Results of the study provided support for the “influential role of modeling variables in the social transmission of self-controlling responses” (p. 703):

Children who had shown a predominantly delayed-reward pattern displayed an increased preference for immediate and less valued rewards as a function of models favoring immediate gratification; conversely, subjects who had exhibited a marked disposition toward immediate rewards displayed an enduring increased willingness to wait for more highly valued delayed reinforcers following exposure to models displaying high-delay behavior. (Bandura & Mischel, 1965, p. 703)

In a milestone culmination of his early research, Bandura (1965) argued for a repositioning of the study of vicarious and observational learning outside of the traditional operant conditioning paradigm, in which responses must be performed and reinforced in order for learning to occur:

Research and theoretical interpretations of learning processes have focused almost exclusively on a single mode of response acquisition which is exemplified by the operant or instrumental conditioning paradigm. In this procedure the organism is impelled, in one way or another, to perform responses under specific stimulus conditions and, through differential reinforcement of spontaneously emitted variations in behavior, new response patterns are developed or existing repertoires are brought under new discriminative stimulus control. It is generally assumed that the principles governing the latter mode of response acquisition account also for social learning phenomena occurring under naturalistic conditions. (Bandura, 1965, pp. 1-2)

Bandura argued that one of the biggest problems with the prevailing paradigm was that “apart from other questions of efficiency…and survival, it is doubtful if many classes of responses would ever be acquired if social training proceeded solely by the method of approximations through differential reinforcement of emitted responses” (Bandura, 1965, p. 2). He further argued that vicarious learning is a means of short-circuiting this method, and humorously employed a play on terminology to underscore the fallacy of differential reinforcement as a suitable means for acquiring certain types of skills in hazardous situations:[2]

It is evident from informal observation that vicarious-learning experiences and response-guidance procedures involving both symbolic and live models are utilized extensively in social learning to short-circuit the acquisition process, and to prevent one-trial extinction of the organism in potentially hazardous situations. (Bandura, 1965, p. 2)

In another clever terminology play, as if to one up Guthrie, Bandura entitled his monograph Vicarious Learning: A Case of No-Trial Learning. By “no-trial” he meant to emphasize that learning could occur without the need for even one single reinforced response:

For the purposes of the present discussion, a vicarious learning event is defined as one in which new responses are acquired or the characteristics of existing response repertoires are modified as a function of observing the behavior of others and its reinforcing consequences, without the modeled responses being overtly performed by the viewer during the exposure period. In demonstrating vicarious learning phenomena, it is therefore necessary to employ a nonresponse acquisition procedure in which a subject simply observes a model’s behavior, but otherwise performs no overt responses, nor is administered any reinforcing stimuli during the period of acquisition. Any learning that occurs under these limiting conditions is purely on an observational or covert basis. This mode of response acquisition is accordingly designated as no-trial learning, since the observer does not engage in any over responding trials although, as will be shown later, he may require multiple observational trials in order to reproduce the stimuli accurately.[3] (Bandura, 1965, p. 3)

It is important to note that in promoting this change Bandura did not advocate the wholesale abandonment of principles of learning in the behaviorist tradition altogether. For example, although he argued that learning can take place through observation without the observer performing any responses to be reinforced, he was not throwing out the principle of reinforcement. What he was doing was presenting a supplementary view of reinforcement based on evidence from a number of empirical studies which suggested that reinforcement, both of the classical conditioning and the operant conditioning varieties, may occur vicariously. Thus, reinforcement could be provided not only by rewarding or punishing the learner directly, but by rewarding or punishing someone else in their presence, with whom they could identify.

In addition to the argument made in favor of broadening the general view of learning to account for “significant vicarious phenomenon…evident in the delayed reproduction of modeling behavior originally learned by observers under a nonresponse acquisition procedure” (Bandura, 1965, p. 47) Bandura’s publication of Vicarious Learning: A Case of No-Trial Learning was important for two other reasons. First, it emphasized “the function of representational processes in observational learning” (p. 47) under which “imaginal and verbal representations of modeling stimuli constitute the enduring learning products of observational experiences” (p. 47). Second, it concluded with a challenge that set a new standard for the variables of learning theory:

The study of social transmission of response patterns is necessitated by the fact that the behavioral repertoires which constitute an enduring part of a culture are to a large extent transmitted on the basis of repeated observation of behavior displayed by social models rather than by memory drums. While the learning process is essentially the same, the characteristics of the social transmitters and other interpersonal variables can greatly affect the rate, level, and types of responses that will be acquired observationally. Moreover, the efficacy of parameters established on the basis of learning in one-person situations may differ in dyadic and group situations (Bandura et al., 1963b). A comprehensive theory of behavior must therefore be based on experimentation involving both social and learning variables. (Bandura, 1965, p. 48)

Twelve years after declaring the need for any comprehensive theory of behavior to include an accounting of both learning and social variables Bandura answered his own challenge with the publication of Social Learning Theory (Bandura, 1977b).[4]  Building on his earlier research, Bandura’s theory of social learning[5] assumes that behavior originates not only from response consequences of direct experience (p. 17), but also through the observation of other people. In fact, the central claim of the theory is that most human behavior is learned in this way:

Learning would be exceedingly laborious, not to mention hazardous, if people had to rely solely on the effects of their own actions to inform them what to do. Fortunately, most human behavior is learned observationally through modeling: from observing others one forms an idea of how new behaviors are performed, and on later occasions this coded information serves as a guide for action. (Bandura, 1977b, p. 22)

According to social learning theory, observational learning is comprised of four component processes: (a) attentional, (b) retention, (c) motor reproduction, and (d) motivational:

1. Attention – “People cannot learn much by observation unless they attend to, and perceive accurately, the significant features of the modeled behavior.” (Bandura, 1977b, p. 24)

2. Retention – “People cannot be much influenced by observation of modeled behavior if they do not remember it.” (Bandura, 1977b, p. 25)

3. Motor reproduction – “The third component of modeling involves converting symbolic representations into appropriate actions…by organizing one’s responses spatially and temporally in accordance with the modeled patterns.” (Bandura, 1977b, p. 27)

4. Motivational process – “People do not enact everything they learn. They are more likely to adopt modeled behavior if it results in outcomes they value than if it has unrewarding or punishing effects…those behaviors that seem to be effective for others are favored over behaviors that are seen to have negative consequences.” (Bandura, 1977b, p. 28)

Social learning theory acknowledges the acquisition of antecedent determiners of behavior. Antecedent determiners are expectancies of “what leads to what” (Bandura, 1977b, p. 58). They can be learned not only through direct experience, but also through symbolic generalization, such as the casting of stereotypes (pp. 63-64), and of course, through vicarious experience (pp. 65-67).

Social learning theory also acknowledges the acquisition of consequent determinants of behavior, or, learning from the result of one’s actions, (p. 96). Consequent determinants are often learned through direct experience, but can also be learned through vicarious reinforcement (p. 117)—that is, through observation of the reinforcement that others receive in answer to their behavior. In fact, it has been shown that “by attending to the pattern of successes and failures of others, observers generally learn faster than do the performers themselves [especially if] the tasks depend more heavily on conceptual [italics added] than on manual [italics added] skills” (p. 122).

Like reinforcement in direct experiential learning, reinforcement in observational learning has the same informative, incentive, and strengthening functions (Bandura, 1977b, p. 17) that it does in learning through direct experience. However, in observational learning it is assumed that reinforcement has an antecedent influence rather than a consequent one. It is facilitative rather than necessary, and is only “one of several factors that can influence what is observed and what goes unnoticed” (p. 37):

It follows from social learning theory that observational learning can be achieved more effectively by informing observers in advance about the benefits of adopting modeled behavior than by waiting until they happen to imitate a model and then rewarding them for it. (Bandura, 1977b, p. 37)

Breaking further from the behaviorist tradition, in Social Learning Theory  Bandura (1977b) introduced an element of cognitive control into the learning process in the forms of (a) “cognitively based motivation” (p. 160), (b) “cognitive representation of contingencies” (p. 165), (c) “[mental] representational guidance of behavior” (p. 170), (d) “problem solving in thought rather than action” (p. 171), (e) “means of distinguishing accurate from inaccurate thinking” (p. 180), and (f) the regulation of “influential cognitions” (p. 187).

Bandura also introduced the concept of triadic reciprocal determinism.[6] Although this was the last idea discussed in Social Learning Theory, it became the first idea presented in most subsequent explanations of the theory. Triadic reciprocal determinism was explained as “a continuous reciprocal interaction between personal, behavioral, and environmental determinants” (Bandura, 1977b, p. 194). Where the behaviorist view was one in which the environment determined behavior, social learning theory included the additional notions that (a) behavior also determines the environment to which a person is subjected to, and (b) that cognitive regulation influences both behavior and how environmental stimuli are perceived.

The same ideas presented in 1977 were again published in 1986 under the name of social cognitive theory (Bandura, 1986). This name change helped to distinguish Bandura’s theory from other similarly named theories with which it was often confused, and to clarify its use in explaining not only “how people acquire cognitive, social, emotional, and behavior competencies but also how they motivate and regulate their behavior and create social systems that organize and structure their lives” (Bandura, 2006a, p. 65). Bandura explained the purpose behind his use of the terms social and cognitive by saying,

In the more fitting appellation as social cognitive theory, the social portion of the title acknowledges the social origins of much human thought and action; the cognitive portion recognizes the influential contribution of cognitive processes to human motivation, affect, and action. (Bandura, 2006a, p. 65)

Along with the name change, many of the transitional links to behavioral learning theory that were leveraged in the initial presentation were dropped—presumably because they were no longer needed in order for the core ideas to be accepted. By 1986 observational and vicarious learning was well known, and well supported by both empirical and anecdotal evidence. The ideas of the theory were further organized and presented a decade later in an entry for the International Encyclopedia of Education (Bandura, 1996). In summary, they are as follows:

1. Model of Causation. Human adaptation is explained in terms of “triadic reciprocal causation” (p. 5513) with continuous interaction between environmental, behavior, and personal (cognitive, biological, and other) factors. “People are both producers and products of their environment” (p. 5513).  “People are neither driven by inner forces nor shaped and controlled by their environment….[but are] contributors to their own development and psychosocial functioning within a network of reciprocally interacting influences” (p. 5513).

2. Symbolizing Capability. “Social cognitive theory assigns a central role to cognitive, vicarious, self-regulatory, and self-reflective processes in human development and functioning” (p. 5513). The remarkable capacity for symbolization enables people to “process and transform transient experiences into cognitive models that server as guides for reasoning and action”(p. 5513). Understanding is gained, and knowledge is expanded by “operating symbolically on the information derived from personal and vicarious experiences” (pp. 5513-5514). This “remarkable flexibility of symbolization enables people to create ideas that transcend their experiences” (p. 5514).

3. Vicarious Capability. “A culture could never transmit its language, mores, social practices, and adaptive competencies if they had to be shaped laboriously in each new member by response consequences” (p. 5514). “The abbreviation of the acquisition process through modeling is vital for survival as well as for human development because natural endowment provides few inborn skills, and errors can be hazardous” (p. 5514). “Humans have evolved an advanced capacity for observational learning….[and] virtually all behavioral, cognitive, and affective learning resulting from direct experience can occur vicariously by observing people’s behavior and its consequences for them” (p. 5514). The four subfunctions that govern observational learning are (a) attentional processes, (b) cognitive representational and memory processes, (c) the behavioral production process, and (d) motivational processes (p. 5514). “Modeling is not a process of behavioral mimicry. Rather, modeling influences convey rules for generative and innovative behavior….[in which] higher-level learning is achieved through abstract modeling in which observers extract the rules governing the modeled judgments and action” (p. 5514). “Modeling influences can strengthen or weaken restraints over behavior patterns that have been previously learned.…by the information it conveys about the probable rewarding or punishing consequences of modeled courses of action” (p. 5515). “People are easily aroused by the emotional expressions of others….[and] learn to fear the things that frighten others, to dislike what repulsed them, and to like what gratifies them” (p. 5515). “In sum, modeling influences serve divers functions as tutors, inhibitors, disinhibitors, social prompters, emotion arousers, and shapers of values and conceptions of reality” (p. 5515).

4. Forethought Capability. “Most human behavior, being purposive, is regulated by forethought” (p. 5515). “Future events cannot be causes of current motivation and action because they have no actual existence. However, by being represented cognitively in the present, foreseeable future events are converted into current motivators and regulators of behavior” (p. 5515).

5. Self-regulatory Capability. “People are not simply knowers and performers. The are also self-reactors with a capacity for self-direction” (p. 5515). “Human self-regulation relies on discrepancy production, as well as discrepancy reduction. People motivate and guide their actions by setting themselves challenging goals and then mobilizing their skills and effort to reach them” (p. 5516). “Because internalized controls can be selectively activated and disengaged, marked changes in moral conduct can be achieved without altering people’s personality structures, moral principles or self-evaluative systems” (p. 5516).

6. Self-reflective Capability. “Effective cognitive functioning requires ways of distinguishing between accurate and faulty thinking….Four different modes of thought verification can be distinguished: enactive, vicarious, persuasory, and logical forms” (p. 5516). “Enactive verification relies on the adequacy of the fit between thought and the results of one’s actions” (p. 5516). “In the vicarious mode…observing other people’s behavior and its effects serves as a way of checking the correctness of one’s own thinking about what leads to what” (p. 5516). “The persuasory mode of thought verification relies on comparing one’s thoughts to the judgments of other [sic]” (p. 5516). “In the course of development, people acquire rules of inference….that [enable] them to detect certain errors in thought by logical verification” (p. 5516). “People’s beliefs in their efficacy influence how they think, feel, act and motivate themselves” (p. 5517). These beliefs come from: “a) performance mastery experiences; (b) vicarious experiences for judging capabilities in comparison with performances of others; (c) verbal persuasion and related types of social influences;…. and (d) physiological states and reactions” (p. 5517).

7. Characteristics of Human Nature. “Human nature is characterized by a vast potentiality that can be developed by direct and vicarious experience into a variety of forms within biological limits” (p. 5517).

Based on a comparative analysis of several presentations of social cognitive theory (Bandura, 1977b; Bandura, 1989; Bandura, 1994b; Bandura, 1996; Bandura, 1999a; Bandura, 2009b) it appears that unlike some of the other theories reviewed in the present study, Bandura’s theory of social learning has been refined over time, but not substantially revised or reworked since its initial articulation. While originally cast with a behavioral orientation (Bandura, 1977b), it took on a more cognitive feel in the 1980s and early-to-mid 1990s (Bandura, 1989; Bandura, 1994b; Bandura, 1996). In more recent years it has undergone another shift toward an agentic perspective of social cognitive theory (Bandura, 2001; Bandura, 2008b; Bandura, 2008c) and toward a psychology of human agency (Bandura, 2006b). Throughout these changes, however, the core ideas and basic assumptions of the theory as originally presented remain intact. Bandura’s further expansion of the theory into the realms of self efficacy and agency will not be reviewed here, as they have already been addressed in our summary of human learning theories in a  previous section.

[1] Bandura (1965, p. 2) gives the following examples:


In laboratory investigations experimenters arrange comparatively benign environments in which errors will not produce fatal consequences for the organism. By contrast, naturalistic environs are loaded with potentially lethal consequences that unmercifully befall those who happen to perform hazardous errors. For this reason, it would be exceedingly injudicious to rely primarily upon trial-and-error and successive approximations methods in teaching children to swim, adolescents to drive automobiles, or adults to master complex occupational and social tasks. If rodents and pigeons toiling in Skinner boxes and various mazes could likewise get electrocuted, dismembered, or extensively bruised for errors that inevitably occur during early phases of learning, it is a reasonably safe prediction that few of these venturesome subjects would ever survive the shaping process.

[2] For example, “teaching children to swim, adolescents to drive automobiles, or adults to master complex occupational and social tasks” (Bandura, 1965, p. 2)

[3] The phrase “reproduce the stimuli accurately” refers to a process of mediation through which “in the course  of observation, transitory sensory and perceptual phenomena are converted to retrievable images of the modeled sequences of behavior” (Bandura, 1965, p. 10), or through which “in addition to the acquisition of imaginal responses, once verbal lables have become attached to objective stimuli, the observer acquires, during the period of exposure, verbal equivalents of the model’s behavior” (Bandura, 1965, p. 11).

[4] Although some of the material in this publication was actually printed six years earlier in a monograph entitled Social Learning Theory (Bandura, 1971), the 1977 publication is cited in Bandura’s biographical sketch (Pajares) as the book that “dramatically altered the direction psychology was to take in the 1980s” (para. 18). Since Bandura links to this biographical sketch from his own professional website (Bandura), it seems reasonable to accept it as an accurate representation of events and attribution of importance to the 1977 publication.

[5] Bandura was not the first, or the only, theorist to use the term social learning theory. Phares (2001) credits Dollard and Miller as being among the first to use the term social learning and Rotter as presenting the first comprehensive social learning theory (p. 1564). Note also that distinguishing his theory from other theories with similar names was one of Bandura’s reasons for later changing the name of his own theory to Social Cognitive Theory (Bandura, 2006a, p. 65).

[6] Later referred to as triadic reciprocal causation (see, for example, Bandura, 1994b; 1996).

Leave a Comment