Last Updated on October 9, 2026 by Karl Thompson
Bandura, Ross and Ross’s Bobo doll experiment is a classic laboratory experiment which found that young children imitate aggression they have watched an adult perform. This post covers:
- The aim of the study
- The sample and the procedure, in three stages
- The findings, including the differences between boys and girls
- What Bandura concluded: social learning theory
- The strengths and limitations of the study, practical, ethical and theoretical
- How to use it in A-level sociology — for research methods, and for crime and the media
The study belongs to the research methods part of A-level sociology, where it is one of the standard examples of a laboratory experiment. It is also useful in crime and deviance and in the media module, in debates about whether watching violence makes people violent.
Aim
Albert Bandura, Dorothea Ross and Sheila Ross wanted to find out whether children would copy aggressive behaviour they had seen an adult perform, in a new setting where the adult was no longer present.
They also had two further predictions. Children who watched a quiet, non-aggressive adult would be less aggressive than children who had seen no adult at all. And because aggression was seen as a typically masculine behaviour, boys would imitate it more than girls, especially when the adult was a man.
Sample
The sample was 72 children, 36 boys and 36 girls, from the Stanford University Nursery School in California. They were aged between 37 and 69 months — roughly three to six years old — with an average age of 52 months.
Before the experiment, the children were rated for how aggressive they normally were, by the experimenter and a nursery teacher who both knew them well. The researchers then grouped the children into threes with similar aggression scores and assigned one from each three at random to each condition. This is called matching, and it matters: it means the groups started out equally aggressive, so any difference afterwards cannot be put down to one group simply being made up of more aggressive children.
Procedure
The children were split into three conditions of 24:
- 24 watched an aggressive adult
- 24 watched a non-aggressive adult
- 24 were a control group who saw no adult at all
Within the two adult conditions, half the children saw a model of their own sex and half saw a model of the opposite sex. There were two models, one man and one woman.
Stage one: watching the model
Each child was taken on their own into a playroom and settled in one corner with potato prints and stickers to play with. The adult model sat in the opposite corner with a set of tinker toys, a mallet and a five-foot inflatable Bobo doll — a clown-shaped toy weighted at the bottom so that it bounces back up when it is knocked over.
In the non-aggressive condition, the adult played quietly with the tinker toys for the whole ten minutes and ignored the doll.
In the aggressive condition, the adult played with the tinker toys for about a minute, then turned on the doll for the rest of the session. The model laid the doll on its side, sat on it and punched it on the nose, hit it on the head with the mallet, threw it in the air and kicked it around the room, repeating the sequence about three times. While doing this, the model said things like “Sock him in the nose…”, “Hit him down…”, “Kick him…” and “Pow…”, and two non-aggressive remarks: “He keeps coming back for more” and “He sure is a tough fella.”
The point of these distinctive acts and phrases was that a child would be unlikely to come up with them on their own. If a child later did exactly these things, it would be clear evidence of imitation rather than ordinary rough play.
Stage two: mild frustration
Each child was then taken to a different room full of attractive toys — a fire engine, a jet plane, a doll set and so on — and allowed to start playing. After about two minutes, the experimenter told the child that these were her very best toys and she had decided to save them for the other children.
All the children went through this, including the control group. The idea was to make sure every child was a little frustrated before the test, so that each one had some reason to feel aggressive.
Stage three: the test
Finally each child was taken into a third room containing a mix of aggressive toys (a three-foot Bobo doll, a mallet, two dart guns and a ball with a face painted on it) and non-aggressive ones (a tea set, crayons, dolls, toy animals, cars).
The child spent 20 minutes in the room. Observers watched through a one-way mirror and recorded what the child was doing every five seconds, giving 240 observations per child. They counted imitative aggression (doing or saying what the model had done or said), partial imitation, and non-imitative aggression.
Findings
Children who had watched the aggressive model were far more aggressive than the other two groups. They reproduced a good deal of the model’s physical and verbal aggression, often in almost exactly the same form. Children in the non-aggressive and control groups showed virtually no imitative aggression. All but two of the children in the aggressive condition copied the model’s distinctive aggressive acts.
They copied the non-aggressive remarks too. About a third of the children who saw the aggressive model repeated “He keeps coming back for more” or “He sure is a tough fella”. None of the other children did.
They also invented aggression of their own. Children who had seen the aggressive model showed more aggression that the model had not demonstrated, such as hitting other toys or making hostile remarks, than children who had seen the quiet model.
Boys imitated more physical aggression than girls, but there was no difference between boys and girls in verbal aggression.
The male model had more influence than the female model. Boys who saw the aggressive man were especially aggressive. Some children’s comments showed they thought aggression from a woman was odd (“That’s not the way for a lady to behave”), while aggression from a man was seen as normal.
Watching a quiet adult made children quieter. Children who saw the non-aggressive model, especially the man, were generally less aggressive than the control group, and spent more than twice as much time as the aggressive group simply sitting quietly.
What Bandura concluded
Bandura, Ross and Ross concluded that children can learn new behaviour simply by watching someone else do it, without being rewarded for it themselves. Watching an adult behave aggressively both taught the children specific aggressive acts and seemed to signal that aggression was acceptable.
This became the foundation of social learning theory: the idea that behaviour, including aggression, is learned through observing and imitating others, particularly people we see as role models.
In a follow-up study published in 1963, Bandura, Ross and Ross tested whether children would also imitate aggression shown on film. That is why the Bobo doll research is cited so often in debates about the effects of violence on television and in other media.
Evaluating the Bobo doll experiment
Strengths
Control over variables, so a clear cause and effect. Everything except the model’s behaviour was kept the same: the same rooms, the same toys arranged in the same order, the same frustration, the same 20 minutes. Because the groups were also matched for aggression beforehand, the only plausible explanation for the difference in behaviour is what the children watched. This is the defining strength of a laboratory experiment.
Standardised procedure, so high reliability. The procedure is described in enough detail to be repeated exactly, which is why it has been repeated many times since.
Careful measurement. Behaviour was recorded in five-second intervals against fixed categories, and the researchers checked their own consistency. When two people rated the same children’s aggression independently, their scores correlated at .89, and agreement between the observers who scored the test sessions was in the .90s. That is high inter-rater reliability.
Quantitative data. The results are numbers, which allow precise comparisons between groups — boys and girls, male and female models, aggressive and non-aggressive conditions.
Limitations
Low ecological validity. This is the criticism that matters most for sociologists. A Bobo doll is designed to be hit; it bounces straight back up. Hitting it is closer to play than to aggression towards a person, who would cry, complain or hit back. The setting was an unfamiliar playroom with an adult stranger. We cannot assume the children would behave the same way towards other children in a playground.
It only measured the short term. The children were tested minutes after watching the model. The study tells us nothing about whether one exposure to aggression has any lasting effect, which is the question that matters for debates about media violence.
Observer bias. The person who scored all 72 test sessions was the male model. He did not know which group most children were in — but, as the researchers admit in the paper, children who had seen the aggressive model were easy to identify because they copied its distinctive behaviour. So the observer could usually tell which condition a child had been in.
The experimenter stayed in the room. Many of the children would not stay alone, so the experimenter sat in the corner doing paperwork during the test. Her presence may have affected how the children behaved.
An unrepresentative sample. All 72 children came from one university nursery school in California in the early 1960s. That is a narrow base for claims about how children in general learn aggression.
Ethical problems
Children were deliberately exposed to aggression and frustrated. The researchers set out to make three- to six-year-olds feel frustrated, and half the experimental group watched an adult act violently. Some may have gone on to behave more aggressively afterwards.
Children cannot give informed consent. Very young children cannot understand what they are agreeing to, and they were not told the real purpose of any part of the study. It is hard to see the study passing a modern ethics committee in the same form; see the British Sociological Association’s statement of ethical practice for current expectations.
Using the Bobo doll experiment in A-level sociology
For research methods, it is one of the best examples of what a laboratory experiment can and cannot do. It establishes cause and effect very cleanly, through control and matching, and at the same time shows the method’s two big weaknesses: low ecological validity and ethical problems. Pair it with Milgram’s obedience experiment and you have the two classic lab examples sociologists use.
For crime and deviance, it supports the idea that criminal behaviour can be learned from others rather than being innate, and links to arguments about the influence of violent role models in the family and in the media.
For the media, it is the starting point for the effects debate. HOWEVER, be careful. The study shows that young children copy specific acts against a toy, straight after watching them, in a laboratory. It does not show that watching violence makes people violent in real life. Saying that clearly is exactly the kind of evaluation examiners reward.
This video gives a short summary of the study — although be warned, it is a bit cringeworthy:
Theory and Methods A Level Sociology Revision Bundle
Theory and Methods A-level sociology revision bundle cover
If you like this sort of thing, then you might like my Theory and Methods Revision Bundle – specifically designed to get students through the theory and methods sections of A-level sociology papers 1 and 3.
Contents include:
- 74 pages of revision notes
- 15 mind maps on various topics within theory and methods
- Five theory and methods essays
- ‘How to write methods in context essays’
Signposting and Related Posts
This post has been written primarily for students studying research methods as part of the AQA’s A-level sociology course, and is also relevant to crime and deviance and the media.
The Bobo doll study is an example of a laboratory experiment, so the place to start is Laboratory Experiments in Sociology, which covers the strengths and limitations of the method in full.
Milgram’s Obedience Experiment is the other classic laboratory experiment that usually gets wheeled out for use in sociology.
The Stanford Prison Experiment is a third, and a good example of how a laboratory study can go wrong.
For the basic logic of experiments — hypotheses, independent and dependent variables — see Experiments in Sociology: An Introduction.
For everything else on this topic, see the main research methods page.
Source: Bandura, A., Ross, D. and Ross, S. A. (1961) ‘Transmission of aggression through imitation of aggressive models’, Journal of Abnormal and Social Psychology, 63, pp. 575–582. Full text at Classics in the History of Psychology, York University.
Please click here to return to the homepage – ReviseSociology.com