11-3: Psychological Explanations of Observational Learning
Psychology of Learning
Module 11: Observational Learning 1
Part 3: Psychological Explanations of Observational Learning
Looking Back
Part 1 introduced observational learning as learning from others’ experiences rather than direct personal experience. Part 2 explored observational learning across animal species, from Galef’s food preference research in rats to mate choice copying in quail to tool use learning in primates. We saw that social learning is widespread, adaptive, & can be explained by various mechanisms ranging from simple stimulus enhancement to complex cognitive processes. Now we turn to psychological theories explaining how observational learning works in humans.
Behavioral Explanation of Social Learning
Miller & Dollard (1941) attempted to explain social learning using purely behavioristic principles, without invoking cognitive processes. They proposed that imitation functions as a generalized imitation—a learned operant response. If we have been reinforced for copying others’ actions on previous occasions, copying becomes a learned behavior that we perform whenever we observe someone acting.
Consider a young child learning to imitate. Initially, the child might randomly perform behaviors when observing others. Sometimes these behaviors match the model’s actions, sometimes they don’t. When the child’s behavior matches the model’s behavior, reinforcement follows—parents praise the child, the behavior produces desired outcomes, or social interaction continues smoothly. Through this reinforcement history, imitation itself becomes a learned operant response. Baer, Peterson, & Sherman (1967) reinforced severely mentally-challenged children for imitating various behaviors exhibited by their teacher including standing up, nodding yes, & opening a door. Following these experiences with reinforcement, the children continued to imitate their teacher in the absence of reinforcement—apparently having learned a generalized imitation response.
However, this explanation faces significant limitations. It struggles to explain immediate imitation of novel behaviors—if imitation requires prior reinforcement history, how can organisms imitate completely new behaviors on first exposure? It doesn’t account for deferred imitation where delays separate observation from performance. It fails to explain selective imitation—people don’t copy everything they observe, but choose particular models & behaviors. These limitations motivated more sophisticated explanations acknowledging cognitive processes.
Bandura’s Social-Cognitive Theory
Bandura & Walters (1963) revolutionized understanding of observational learning by arguing that learning can occur through vicarious learning—observing others’ behaviors & their consequences, without direct personal experience. Their work challenged behaviorist orthodoxy by demonstrating learning without direct reinforcement of the learner. This theoretical advance has had lasting influence on clinical practice, with guided mastery therapy & self-efficacy interventions continuing to demonstrate effectiveness in treating anxiety disorders & promoting health behavior change (Scott & Cervone, 2024).
Modeling is a term introduced by Bandura used interchangeably with observational learning & imitation. It emphasizes the process by which observers pattern their behavior after a model, learning through watching others perform behaviors. Modeling can be narrowly displayed with imitation of specific behaviors, or much broader, such as when a boy identifies with his father & tries to “become him” over a period of years.
Three Types of Behavioral Change Through Modeling
Learning by modeling incorporates three distinctly different types of behavioral change (Bandura & Walters, 1963). First, in the imitative sense, a learner can directly copy the model’s behavior—learning a completely new response that was not previously in their repertoire. A child who has never seen anyone tie shoelaces observes a parent demonstrating the behavior & learns this novel skill.
Second, in many cases the model’s demonstrated behavior invokes inhibition or disinhibition of behaviors the learner already knows. The learner can learn through modeling which behaviors to perform (or not perform) in certain contexts. A person might know how to act aggressively but inhibit such behavior due to anticipated consequences; observing a model act aggressively without punishment may weaken these inhibitions. The teacher hopes that disciplining a student for screaming will not only reduce the probability that the naughty student will scream again (through operant conditioning), but will also model the inhibition of screaming to other students in the class. The other kids already know how to scream, but they need to learn when & where screaming is unacceptable.
Third, modeling can elicit an already learned response. When one child sees some crayons & paper & begins coloring, they become a model for other children, eliciting their coloring behavior as well. This resembles social facilitation—the model’s behavior triggers similar behavior in observers. In the process of direct imitation (the first type), the observer does not need to observe the consequences of the behavior. However, the second two types (inhibition/disinhibition & elicitation) require viewing consequences of the behavior to the model.
Recent psychoanalytic critiques have suggested that deeper psychological mechanisms—including identification with the aggressor—may also play a role in children’s aggressive imitation (Galanaki & Malafantis, 2022). However, social-cognitive theory remains the dominant explanation for observational learning.
Factors Affecting the Probability a Model Will Be Imitated
Though children are exposed to many different models whose behavior could be imitated, they do not imitate all of them. Mischel (1971) describes factors of rewardingness, dominance, similarity, & sincerity as determining which models will be imitated. The existence of these factors suggests that imitation is selective & strategic, not blind copying.
Rewardingness: Models Who Provide Reinforcement. A model’s rewardingness refers to how much general reinforcement the observer receives from that source. People who provide attention, approval, affection, or tangible rewards become powerful models. Bandura & Huston (1961) had nursery school students observe an adult play with toys. For some children, the adult was warm & nurturing (high rewardingness); for others, the adult was distant & cold (low rewardingness). When children later observed the adult make several distinctive movements & noises during a game, children who had received nurturing treatment were significantly more likely to imitate these incidental behaviors. The authors argued that this explains why children tend to imitate their parents—who are typically their primary source of rewards.
Dominance: Models Who Hold Power. Mischel & Grusec (1966) demonstrated that children are more likely to imitate an adult introduced as “the new teacher” (implying ongoing authority over the children) than an adult introduced as “a visiting teacher from out of town” (implying no continuing relationship). Presumably, the “new teacher” will hold some dominance & power over the children, hence the children’s increased probability of imitation. Research also shows that dominant children in classrooms are more frequently imitated by their peers.
Similarity: Models Who Resemble the Observer. Burnstein, Stotland, & Zander (1961) showed that children are more likely to imitate a model who is the same gender, same age, or has similar interests. This makes adaptive sense—consequences experienced by similar others are more likely to apply to oneself than consequences experienced by dissimilar others.
Sincerity: Models Who Appear Genuine. Klass (1979) showed that children are more likely to imitate a model whom they perceive to be sincere. If a model’s behavior seems fake or manipulative, children are less likely to copy it.
The Social-Cognitive Theory of Learning
In his social-cognitive theory of learning, Bandura (1971, 1977) argues that the process of learning by observation of others differs very little from traditional operant conditioning when properly analyzed. Traditional operant conditioning can be understood as observational learning of the consequences of our own actions followed by a cognitive appraisal of the probability of similar consequences if we repeat the behavior. That is, observational learning & operant conditioning both involve expectations about the consequences of our future actions. In one case these expectations arise based on our own past experiences, whereas in the other case, the expectation arises due to our witnessing the experiences of others.
Key principles of social-cognitive theory include: People can learn by observing the behavior of others, as well as by observing the outcomes of those behaviors. Most learning occurs through observation rather than direct experience. Learning can occur without a change in behavior—knowledge can be acquired & stored without being immediately performed. Learning plus motivation equals behavior. The consequences of behavior play a role in learning, including reinforcement & punishment of the model. Cognition plays a role in learning—people gain expectancies about behaviors learned through observation, not just blind imitation. People can have considerable control over their actions & their environments through self-regulation.
The Four Processes of Observational Learning
The social-cognitive theory proposes that learning by observation requires four processes in order to be effective: Attention, Retention, Reproduction, & Motivation (Bandura, 1977).
Attention: We do not learn from everything we observe. If you are sleeping in chemistry class while the professor is demonstrating an experiment you are supposed to perform, you are not paying attention & there is no probability that you will be able to produce the required steps when necessary. Attention is influenced by model characteristics (more attention to attractive, prestigious, or similar models), observer characteristics (arousal level, perceptual capabilities, past reinforcement for attending), & stimulus characteristics (distinctiveness, complexity, prevalence).
Retention: We must be able to retain the memory of the behaviors we have observed for some length of time. The memory must be accessible at the time the new behavior is to be performed. Retention involves forming mental representations of observed behavior through imaginal (visual) & verbal (linguistic) encoding systems. Rehearsal—whether cognitive or physical practice—strengthens retention.
Reproduction: The subject must have the physical motor skills to reproduce the observed behavior. You might be able to pay attention while watching a juggling demonstration & retain the memory of how the juggling was performed, but when you try to reproduce the behavior you might find that your dexterity is poor & you continually drop the balls, despite “knowing” how to juggle. Reproduction involves translating cognitive representations into motor performance, requiring physical capability & skill development through practice.
Motivation: Even if you have paid attention, retained the memory, & have the required motor skills, you may not perform a new behavior without the right motivation or incentives. Motivation determines whether learned behaviors are actually performed based on expected consequences. One family tried a novel approach to toilet training their young son. They bought a video demonstrating the process. The video was bright & colorful so the boy paid attention. The parents discussed the video over & over with the boy, to ensure that he retained the knowledge. Whenever the boy needed to poop, he demanded a diaper. Because he was able to wait for the diaper, he obviously had the motor skills required for the behavior. But unfortunately for the parents, the boy lacked the motivation to use the potty.
Vicarious Reinforcement & Punishment
Vicarious reinforcement occurs when observing a model receive reinforcement for behavior increases the observer’s motivation to perform similar behaviors based on expectancies of similar rewards. Vicarious punishment occurs when observing a model punished for behavior decreases the observer’s likelihood of performing that behavior, without the observer personally experiencing punishment. Importantly, vicarious consequences affect motivation to perform, not learning itself—observers can learn behaviors equally well regardless of whether models are rewarded or punished, but vicarious reinforcement increases & vicarious punishment decreases spontaneous performance of learned behaviors.
Self-Efficacy & the Agentic Perspective
Self-efficacy is an individual’s belief in their capability to execute behaviors necessary to produce specific outcomes. Self-efficacy influences activity choice, effort expenditure, & persistence in the face of obstacles (Bandura, 1997). People with high self-efficacy for a task approach it with confidence, persist through difficulties, & recover quickly from setbacks. People with low self-efficacy avoid challenging tasks, give up easily, & experience stress & depression when facing obstacles.
Observational learning powerfully influences self-efficacy. Watching similar others succeed at a task increases self-efficacy (“If they can do it, so can I”), while watching similar others fail decreases it. Vicarious experience is one of four sources of self-efficacy information, along with mastery experiences (personal success/failure), verbal persuasion (encouragement from others), & physiological states (anxiety or arousal interpreted as capability or incapability). A recent meta-analysis of 31 studies found that interventions targeting teacher self-efficacy through modeling & guided mastery significantly improved teaching effectiveness (Täschner et al., 2024).
Bandura’s agentic perspective emphasizes that people are not merely shaped by environmental forces but actively contribute to their own development & functioning through intentional action, forethought, self-regulation, & self-reflection (Bandura, 2023). Humans are agents of their own lives, capable of exercising control over their thoughts, feelings, & actions. This perspective contrasts with purely behavioristic views that see behavior as controlled entirely by environmental contingencies.
Reciprocal Determinism
Reciprocal determinism is Bandura’s concept that behavior, personal factors (cognitions, emotions, biological events), & environmental factors all influence one another in a continuous, dynamic, triadic interaction. This stands in contrast to unidirectional models that see behavior as controlled solely by the environment (behaviorism) or solely by internal forces (psychoanalysis). In reciprocal determinism, each element of the triad both influences & is influenced by the others—we observe our environment, we interact with & change the very environment we are observing, & our behavior feeds back to modify our cognitions & the environment’s effects on us.
Looking Forward
Module 12 examines the biological foundations of observational learning—evolutionary explanations, mirror neurons, & implications for understanding autism. We then turn to applications of observational learning theory, exploring how it explains the acquisition of aggressive behavior (the Bobo doll experiments), media violence effects, achievement motivation, & cultural transmission.