Why the study matters

The outline asks about three kinds of importance: the contribution to psychology, the influence on theory, and the influence on practice. This page takes them in that order, then looks at what later research did to the study's claims.

Contribution to psychology

The study made one point so clearly that it changed the field: people can learn new behaviour simply by watching someone else, without doing it themselves and without any reward. In 1961 this was a challenge to the strongest theory of learning, which held that new behaviour is learned through its consequences.

The design made the point hard to argue with. The children reproduced unusual acts, a particular way of sitting on the doll and punching its nose, and phrases such as "He keeps coming back for more", which they would not have invented themselves. And they did so in a different room, with the model gone. The study gave psychology a simple, repeatable method, the Bobo paradigm, that other researchers could use and extend.

Influence on theory

The study was the first step in a theory that grew for the next 25 years.

From the Bobo doll study to social cognitive theory

• Study
• Theory
• Practice and policy

1. 1941

   Theory: Miller and Dollard, Social Learning and Imitation

   Imitation is learned because copying has been rewarded. The view Bandura set out to go beyond.

2. 1961

   Study: The Bobo doll study

   Children copy a live model's new aggressive acts, without reward, in a new room.

3. 1963

   Study: Film and cartoon models

   Children copy a filmed adult as much as a live one. A cartoon cat is copied a little less but raises aggression just as much.

4. 1963

   Theory: Bandura and Walters, Social Learning and Personality Development

   The first book-length statement of social learning theory.

5. 1965

   Study: Rewarded, punished or no consequences

   Seeing the model punished reduces copying, but offering children rewards reveals they learned the acts anyway.

6. 1967

   Practice and policy: Modelling as therapy

   Children afraid of dogs become less fearful after watching another child approach and play with a dog.

7. 1972

   Practice and policy: US Surgeon General's report on television violence

   Research in this tradition enters national policy debate on television.

8. 1977

   Theory: Social Learning Theory and self-efficacy

   Observational learning is described as four processes. Self-efficacy, belief in one's own ability, becomes a central idea.

9. 1986

   Theory: Social cognitive theory

   Bandura renames his approach, placing thinking, self-regulation and agency at its centre.

Four ideas that came from the Bobo studies

1. Observational learning. People learn by watching others. Bandura (1977) later described four processes: attention to the model, retention of what was seen in memory, reproduction of the behaviour, and motivation to perform it. The 1961 study fits all four, although it measured only whether children reproduced the acts and inferred the rest.

2. Learning is not the same as doing. Bandura's 1965 study showed this directly. Children watched a film in which a man, later called Rocky, attacked a Bobo doll. In one version Rocky was praised and given sweets, in another he was told off and spanked with a rolled-up magazine, and in a third nothing happened to him. Children who saw him punished copied less. But when the experimenter then offered every child stickers and juice for showing what Rocky had done, the differences between the three groups disappeared. All the groups had learned the acts equally well. Seeing consequences changed what children chose to do, not what they learned.

The 1965 results in more detail

The study had 66 children, 33 boys and 33 girls. In free play, children who saw the model punished copied fewer acts than both other groups (both p < .025), but children who saw the model rewarded did not copy more than children who saw no consequences (t = 0.55, not significant). The effect came mainly from the girls. After the incentive was offered, the treatment differences vanished, but girls still reproduced somewhat fewer acts than boys (Bandura, 1965, pp. 592-594).

3. Vicarious reinforcement. Vicarious reinforcement means learning about the likely consequences of an act by watching what happens to someone else who does it. The 1965 study, and a 1963 study in which children watched Rocky either win or be beaten, showed that children are sensitive to what happens to models.

4. Self-efficacy and agency. In later work Bandura argued that people's beliefs about their own abilities, their self-efficacy, shape what they attempt and how long they persist. His approach moved from explaining behaviour by the environment to seeing people as active agents who think, plan and regulate themselves. This is reciprocal determinism: the person, their behaviour and their environment all influence each other.

Influence on practice

Television and media policy. The 1961 paper never mentions television. The link came with the 1963 follow-up, which showed that children copy filmed and cartoon models. In 1969 the US Surgeon General set up a scientific committee on television and social behaviour. The television networks were allowed to veto candidates, and ABC, NBC and the broadcasters' trade association all named Bandura as unacceptable (Paisley, 1972). The committee's 1972 report was cautious. It found "a preliminary and tentative indication of a causal relation between viewing violence on television and aggressive behavior", operating "only on some children (who are predisposed to be aggressive)". It also noted that some questioned whether laboratory behaviour could be called "aggressive" in the everyday sense. Decades of hearings, rating systems and parental warnings followed.

Therapy. Modelling became a tool for helping people. In a study by Bandura, Grusec and Menlove (1967), young children who were afraid of dogs watched another child approach and play with a dog, and afterwards they avoided dogs less. Bandura's later work on self-efficacy grew out of modelling treatment for adults with snake phobias. Modelling is now used in treatments for phobias, in social skills training and in video modelling for children with autism.

Parents. The idea that children learn from what adults do, not only from what adults say, is built into many parent-training programmes. Indian child-protection guidance uses the same reasoning against corporal punishment, as the India page shows.

Entertainment-education, and an Indian example. In the 1970s the Mexican producer Miguel Sabido designed television serials whose characters modelled behaviours such as joining adult literacy classes. He drew directly on Bandura's social learning theory. Hum Log, the first Indian-made soap opera on Doordarshan, broadcast from July 1984 to December 1985, was the first use of Sabido's method outside Mexico. It was hugely popular, with an average audience of about 50 million. But its own evaluators, Arvind Singhal and Everett Rogers, reported mixed results. The family-planning theme was almost dropped after the 13th episode because viewers disliked the sermons. Many viewers identified most with Bhagwanti, the self-sacrificing mother who was meant to be a negative role model for women's equality. Watching the serial was linked to behaviours showing more freedom of choice for women, but these effects, though statistically significant, were "all quite small" (Singhal & Rogers, 1991, p. 22). Modelling through media works, but viewers choose their own models.

What later research did to the claims

The main finding, that children copy models, has held up well. It was repeated in Bandura's own later studies and by others, including a study in which boys who had watched an aggressive film attacked a person dressed as a clown, not just a doll (Hanratty, Liebert, Morris & Fernandez, 1969). But the bigger claim that grew out of it, that violent media make people violent, has been fiercely disputed.

• Evidence of long-term links. Huesmann and colleagues (2003) began with 557 children from the Chicago area, tested in 1977-78, and re-interviewed 329 of them in their early twenties. Children who watched more television violence were more aggressive as young adults, for both men and women (correlations of about .2), even after accounting for social class, intelligence and parenting. This is a correlational study, one that measures how two things vary together without changing either, so it cannot prove cause, although the authors found the link ran more strongly from early viewing to later aggression than the reverse.
• Evidence that effects are small or overstated. A meta-analysis combines the results of many studies into one estimate. Christopher Ferguson's 2015 meta-analysis of 101 studies of children and teenagers found a very small link between violent video games and aggression (r = .06). Hilgard, Engelhardt and Rouder (2017) found signs of publication bias in the laboratory studies: studies that found an effect were more likely to be published than studies that did not, which makes the published effect look bigger than it is. Mathur and VanderWeele (2019), comparing rival meta-analyses, concluded that the effects are "nearly always detrimental in direction but are rarely stronger than" a small standardised effect.
• Courts and professional bodies. In 2011 the US Supreme Court struck down a California law that banned selling violent video games to minors. The Court's opinion, joined by five of the nine justices, called the state's evidence "not compelling": the studies "do not prove that violent video games cause minors to act aggressively". None of the opinions cited the Bobo studies. In 2020 the American Psychological Association revised its position. As reported by its media psychology division, it described "a small, reliable association between violent video game use and aggressive outcomes, such as yelling and pushing", but said these findings are "difficult to extend to more violent outcomes".

Why effect sizes matter here

A correlation (r) runs from 0, no relationship, to 1, a perfect relationship. An r of .06 means that violent game use accounts for well under 1% of the differences between young people in aggression. Huesmann's r of about .2 is larger, and the authors argued it matters at the level of a whole population. Small effects can still matter when millions of people are exposed, but they do not justify claims that media are a main cause of violence.

So the study's importance today is mainly as a demonstration of how children learn from models, which is widely accepted, rather than as proof about media violence, which remains contested.

Try this

A school principal reads about the Bobo study and bans all superhero toys, saying the research proves they cause violence.

What does the research actually support, and what would be a better use of it in a school?

Show the reasoning

The research supports the claim that young children readily copy specific acts they have just seen a model perform, especially when nothing bad happens to the model. It does not show that toys cause violence, and later debate shows that links between media and serious violence are small and contested.

A better use of the research is to focus on models. Children copy what adults and older children do and what happens to them. So the school could make sure staff model calm conflict resolution, respond consistently when children hurt each other, and praise children who help and share. The 1965 finding that consequences shape what children choose to do supports this more than it supports banning toys.