Transmission of Aggression



Albert Bandura is a Canadian psychologist who carried out a famous study on the transmission of aggression in children, commonly known as the “Bobo doll experiment”.

Albert Bandura

Bandura played a major role in developing the “social learning theory”. This theory revolves around the belief that children develop by learning from other people, particularly through observation and imitation. Social development is seen as a continuous learning process involving behaviour, cognitions and an individual’s environment.

The social learning theory is rooted in classic behaviourist theories, such as classical conditioning and operant conditioning. Behaviourists focus strictly on observable, measurable behaviour but social learning theorists also give importance to cognitive processes. These cognitive processes are obviously not explicitly observable; we make conclusions about cognitive processes by observing and measuring behaviour.

Bandura is often classed as a behaviourist but he described his own approach as “social cognitivism”.


The sample in Bandura’s experiment consisted of 72 children (36 boys and 36 girls) aged between 37 to 69 months (3 to 5 year olds). All children were from the Stanford University nursery school and their parents had given consent.


This study had four hypotheses:

  1. Children exposed to aggressive models will reproduce aggressive acts that resemble the models.
  2. Children exposed to non-aggressive models will reproduce less aggressive acts.
  3. Children will imitate the behaviour of a same-sex model more than a model of the opposite sex.
  4. Boys will be more predisposed towards imitating aggression than girls.


Bandura separated the children into three groups: two experimental groups and one control group. Each group was made up of 24 children. The two experimental groups were divided further by gender: six boys with a male model, six boys with a female model, six girls with a female model and six girls with a male model. The children in the control group did not see any model.

Since the number of children in each experimental group sub-group was small (only six children), the researchers needed to reduce the chances of individual differences. For example, if every child in one sub-group was naturally more aggressive than the other children, the results wouldn’t be accurate. The researchers addressed this by testing the children before using them in the experiment. They observed the children as they played in their nursery and rated their aggressiveness on a five-point scale. The observers were a female experimenter, a female nursery teacher and the model for male aggression. The children were rated on the following qualities:

  • Physical aggression
  • Verbal aggression
  • Aggression towards inanimate objects
  • Aggressive inhibition

In this way, the children all had similar levels of natural aggression.

The experiment consisted of three stages:

Stage I
Children were first tested individually. A child was brought into the experimental room by the experimenter. The room was arranged in a way that targeted children; one corner was a child’s play area with a table and chair, potato prints and stickers. The child was seated in this corner and the model, who was standing outside in the hallway, would be invited by the experimenter to come inside and play. The adult model was then escorted to a corner of the room with a small table and chair, a tinker-toy set, a mallet and a five-foot inflatable Bobo doll.

The original Bobo doll, oooh!

After the model was seated, the experimenter left the room.

In the non-aggressive condition, the model ignored the Bobo doll and played with the tinker-toys quietly. In the aggressive condition, the model started arranging the tinker-toys but, after a minute, turned to the Bobo doll and became aggressive.

  • An example of physical aggression was when the model “raised the Bobo doll and pommelled it on the head with a mallet.”
  • An example of verbal aggression was when the model said “Pow!” and “Sock him in the nose!”

After 10 minutes, the experimenter returned and took the child to a new room. They child was told that the new room was another games room.

Stage II
In the second stage, the child experienced “mild aggression arousal”. This began by taking the child to the new room with attractive toys. When the child started to play, the experimenter would tell the child that these were the experimenter’s best toys and that she had decided to keep them safe for some other children. This was a means of control to ensure that all children began the next part of the experiment at a relatively same level of aggression.

Stage III
The child was taken to a third room where he or she was told that they could play with any of the toys in the room. The experimenter stayed in the room because some children refused to stay alone or would end up leaving before the end of the experiment. The room had different toys; some toys were aggressive (mallet, peg board, dat guns and a three-foot Bobo doll) and some were non-aggressive (tea set, crayons, three bears and plastic farm animals).

The child stayed in this room for 20 minutes, during which their behaviour was observed through a one-way mirror. Observations were made after five-second intervals (so a total of 240 responses were recorded for each child).

The results were very detailed and encompassed a whole load of things. The results were found as follows.

Three measures of imitation (similar to the model’s behaviour) were noted:

  1. Imitation of physical aggression (e.g. punching the doll in the nose)
  2. Imitation of verbal aggression (e.g. repeating the phrase “Pow!”)
  3. Imitation of non-aggressive verbal responses (e.g. saying, “He keeps coming back for more.”)

Two types of non-imitated behaviour (not similar to what the model was doing):

  1. Mallet aggression (e.g. striking toys with a mallet instead of hitting the Bobo doll)
  2. Sitting on Bobo (the child sat on Bobo but didn’t behave aggressively)

Three non-imitated aggressive behaviours:

  1. Punching the Bobo doll
  2. Physical and verbal aggression that wasn’t done by the model
  3. Aggressive gun-play

Probably shouldn’t get on the wrong side of this kid, folks.

The other findings of this study are as follows:

  • Children in the aggressive condition made more aggressive responses than children in the non-aggressive condition (supporting the hypothesis).
  • Boys engaged in more aggressive behaviour than girls (supporting the hypothesis).
  • Boys in the aggressive condition showed more aggressive behaviour if the model was male instead of female (supporting the hypothesis).
  • Girls in the aggressive condition showed more physically aggressive behaviour when the model was MALE but engaged in more verbally aggressive behaviour if the model was FEMALE (not totally supporting the hypothesis).

Bandura also reported that aggressive female models contradicted with the children’s thinking. This is probably because these children grew up in a culture that did not include aggressive females. Some examples of this cultural background can be seen in these quotes from the children:

  • “Who is that lady? That’s not the way for a lady to behave. Ladies are supposed to act like ladies…”
  • “You should have seen what that girl did in there! She was punching and fighting but no swearing.”

Unsurprisingly, aggressive male models aligned with the children’s cultural stereotypes of appropriate gender-based behaviour. For example, one boy and one girl said, respectively:

  • “Al’s a good sucker, he beat up Bobo. I want to sock like Al.”

“That man is a strong fighter, he punched and punched and he could hit Bobo right down to the floor and if Bobo got up he said ‘Punch your nose!’ He’s a good fighter like Daddy.”

Aggressive adult, aggressive boy, aggressive girl… Poor Bobo.

Type of research method
This was a laboratory experiment.

Independent variable
There were three IVs: the condition the child experienced (aggressive/non-aggressive), the sex of the model (male/female) and the sex of the child (male/female).

Dependent variable
The degree of aggression shown by the children.


  1. High level of control: There was a strong level of control over extraneous variables in this study. The experimenters even matched the children according to their natural levels of aggression before the experiment had begun.
  2. Replicable: Since the procedure was standardised and tightly controlled, it can be replicated and the reliability of the results can be checked.


  1. Lack of ecological validity: The experiment was conducted in a laboratory, not a natural environment. This means the situation was not very realistic (one child and one adult, a random room, limited interaction).
  2. Snapshot: This study is an example of snapshot results. The children’s level of aggression was measured straight after their interaction with the model(or no interaction, depending on the condition) so this does not give us a long-term look at the children’s learned behaviour.
  3. Not totally generalisable: Although the study used young boys and girls, they were all from the same city and background so this study cannot be applied to children from other cultures.

Ethical issues

  1. Informed consent: Bandura did get informed consent from the children’s parents.
  2. Deception: There was not much deception, although the children were obviously unaware about the nature of this experiment.
  3. Confidentiality: The children’s identities were not publicised.
  4. Emotional or physical harm: There was no physical harm. However, there is some criticism that Bandura may have affected the children’s emotional state in the long-run. This is unlikely but also a point to consider.
  5. The right to withdraw: Nobody was forced to stay part of the study. However, it was not made clear about how the children could withdraw from the experimental setting.
  6. Debriefing: The children were not debriefed, although it is possible that their parents may have been informed.