The Bobo Doll Experiment: Can Children Learn Aggression by Watching?

Mymind Elevation

11 minutes ago

4 min read

What the famous study found-and what it did not prove

Imagine a young child watching an adult strike, kick and shout at an inflatable toy. When the adult leaves, will the child invent the same actions-even without being directly rewarded?

Historically reconstructed 1960s laboratory playroom showing a preschool child safely observing an adult model demonstrate non-graphic aggressive actions toward a weighted inflatable clown doll.

The 1961 experiment asked whether children would reproduce behaviour they had only observed.

The Question Behind the Bobo Doll Experiment

In 1961, Albert Bandura, Dorothea Ross and Sheila Ross investigated whether children could learn new aggressive responses by observing an adult model. At the time, influential behaviourist accounts placed strong emphasis on direct reinforcement: behaviour followed by reward becomes more likely; behaviour followed by punishment becomes less likely. Bandura’s question was more social. Could watching another person be enough for learning to occur?

The researchers studied 72 children-36 girls and 36 boys-aged 37 to 69 months from Stanford University Nursery School. Children were assigned to an aggressive-model condition, a non-aggressive-model condition or a control condition with no prior model exposure. The experiment also varied the sex of the adult model.

What Did the Researchers Do?

In the aggressive condition, a child watched an adult play briefly and then direct a distinctive sequence of physical and verbal actions toward a large inflatable Bobo doll. The non-aggressive model played quietly and ignored the doll. Control children did not observe a model.

The child was then taken to another room containing attractive toys. After a short period, the toys were withdrawn and described as reserved for other children-a deliberate attempt to create mild frustration. Finally, the child entered a playroom containing a Bobo doll, a mallet, other potentially aggressive toys and ordinary non-aggressive toys. Observers recorded behaviour through a one-way mirror.

Three text-free historical reconstruction panels showing observation of a model, mild toy frustration and later structured play with a Bobo doll and neutral toys.

The 1961 procedure moved from observation to mild frustration and then to a structured play session.

What Happened?

Children who observed the aggressive model produced substantially more actions closely resembling the model’s behaviour than children in the non-aggressive and control conditions. They copied distinctive physical actions and repeated modelled verbal expressions. This mattered because the responses were not merely common rough play; several had been specifically demonstrated by the adult.

The sex-related findings were more complicated than the familiar summary suggests. Boys displayed more physical aggression overall, especially after observing a male aggressive model. Some same-sex modelling effects appeared, but they were not uniform across every measure. These results emerged within the gender expectations and limited nursery-school sample of the early 1960s and should not be treated as timeless biological rules.

What Did the Experiment Appear to Demonstrate?

The strongest conclusion was that children can acquire new patterns of behaviour through observation. Direct reward to the child was not necessary for the behaviour to be learned and reproduced. This work became central to social learning theory and, later, Bandura’s broader social cognitive theory.

Observational learning is not automatic copying. Whether a person performs what was observed depends on attention, memory, ability, expected consequences, identification with the model and motivation. A child can therefore learn a behaviour without displaying it immediately.

Learning Is Not the Same as Performance

A later 1965 experiment made this distinction especially clear. Children watched a filmed model behave aggressively and then saw the model rewarded, punished or receive no consequence. Punishment reduced spontaneous imitation. However, when children were later offered incentives to reproduce what they had seen, differences between conditions became much smaller. The punished group had not necessarily failed to learn the behaviour; they were less willing to perform it without a reason.

Three text-free historical reconstruction panels showing an adult model receiving approval, disapproval or no visible consequence while a child observes.

Later research-not the original 1961 experiment-tested how reward, disapproval or no consequence influenced whether children expressed what they had learned.

What Later Evidence Clarified

Bandura and colleagues showed in 1963 that aggressive actions presented through live models, film and a cartoon model could all be imitated in the subsequent laboratory setting. The studies helped establish that models do not need to be physically present to influence learning.

Nevertheless, the original experiments cannot carry every claim later attached to them. They measured short-term behaviour in an artificial playroom after a scripted demonstration and deliberate frustration. Striking a weighted inflatable toy designed to spring upright is not equivalent to harming a person. The studies therefore support immediate observational learning more directly than they support claims about enduring aggression, criminal violence or the effect of one television programme on a particular child.

The responsible interpretation is therefore narrow but important: observation can support the acquisition and short-term expression of new behaviour under particular conditions. Behaviour outside the laboratory develops through interacting personal, relational, cultural and environmental influences.

Ethical Concerns

By contemporary standards, deliberately exposing preschool children to modelled aggression and then provoking frustration raises important questions. These include possible distress, the teaching or normalisation of harmful behaviour, the adequacy of parental consent and child assent, the child’s freedom to withdraw, debriefing and whether any learned behaviour was actively corrected afterwards.

Historical research should be interpreted within its period, but historical importance does not remove ethical scrutiny. A direct recreation with young children would require a substantially different risk–benefit justification and stronger safeguards today.

What It Means Today

The enduring lesson is broader than aggression: adults teach through what they repeatedly demonstrate, tolerate, reward and repair. Children observe how caregivers manage frustration, how teachers respond to mistakes, how peers gain status and how online personalities are rewarded. Modelling can transmit patience, cooperation and emotional regulation just as it can transmit hostility.

This does not mean that every unwanted behaviour is copied from a parent or screen, and it should never be used to assign simplistic blame. It means that instruction becomes more credible when behaviour matches it. “Calm down” teaches less than an adult visibly pausing, regulating their voice and repairing a conflict.

Final Takeaway

The Bobo Doll Experiment demonstrated that children can learn novel behaviour by watching another person. It did not prove that imitation is inevitable, permanent or produced by observation alone. Its real legacy is the recognition that human learning is social: we watch what people do, notice what happens to them and use that information when choosing what to do ourselves.