Why did Bandura do the Bobo doll experiment?

These findings are still debated about over 40 years later.

In the modern world, there are many concerns about the effect of social influences on the development and growth of a child's personality and morality.

Television, computer games, food additives, music and the lack of role models are all cited as reasons for a supposed breakdown in society, and an increased tendency towards violence.

These concerns have existed for many years, even before the media turned these factors into sensationalist stories, to try and sell more newspapers. During the 1960's, there was a lot of concern and debate about whether a child's development was down to genetics, environmental factors or social learning from others around them.

For this purpose, Bandura designed the Bobo Doll Experiment to try and prove that children would copy an adult role model's behavior. He wanted to show, by using aggressive and non-aggressive actors, that a child would tend to imitate and learn from the behavior of a trusted adult.

The Bobo doll is an inflatable toy about five feet tall, designed to spring back upright when knocked over.

Children were chosen as subjects for the study, because they have less social conditioning; they have also had less instruction and teaching of the rules of society than adult subjects.

Hypotheses and Predictions

Bandura had a number of predictions about the outcomes of the Bobo Doll Experiment, fitting with his views on the theories of social learning.

1. Children witnessing an adult role model behaving in an overly aggressive manner would be likely to replicate similar behavior themselves, even if the adult was not present.
2. Subjects who had observed a non-aggressive adult would be the least likely to show violent tendencies, even if the adult was not present. They would be even less likely to exhibit this type of aggression than the control group of children, who had seen no role model at all.
3. Bandura believed that children would be much more likely to copy the behavior of a role model of the same sex. He wanted to show that it was much easier for a child to identify and interact with an adult of the same gender.
4. The final prediction was that male children would tend to be more aggressive than female children, because society has always tolerated and advocated violent behavior in men more than women.

Setting Up the Experiment

For the Bobo Doll Experiment, Bandura selected a number of children from the local Stanford Nursery School, varying in age from 3 to 6 years, with the average age being 4 years and 4 months.

To test the prediction that boys would be more prone to aggression than girls, he picked 36 subjects of each sex.

The control group, which would not see an adult role model at all, consisted of 24 children, 12 boys and 12 girls.

The second group, which would be exposed to an adult showing aggressive tendencies, was similarly made up of 24 children of either sex. Both of the resulting groups of 12 were further divided; half would be tested with a female role model, half with a male role model.

The third group was structured in exactly the same way as the second, the only difference being that they would be exposed to a passive adult.

For the Bobo Doll Experiment, it was necessary to pre-select and sort the children, to try and ensure that there was an even spread of personality types across the test groups; some subjects already known to be more aggressive in personality than others.

For this, one of the teachers from the nursery worked with the experimenter, to rate each child's personality and attempt to construct well balanced groups.

It must also be noted that each subject was tested alone and individually, to ensure that the effects and reactions of their classmates would bear no influence on the final results or findings of the experiment.

The Bobo Doll Experiment proper began by placing one of the children from the test groups in a room with an adult. The subject sat in one corner of the room, with a few appealing toys to play with, such as potato prints and sticker activities.

The adult sat in the other corner of the room, with a few toys, as well as a Bobo doll and mallet. The child was not permitted to play or interact with these toys.

For the children in group two, after one minute of playing with the toys, the adult would begin to verbally and physically attack the doll for a period of 10 minutes.

For the third group tested, the adult would sit quietly and play peacefully with the toys for ten minutes.

The control group, of course, sat in the room for ten minutes with no adult present.

The next stage of the Bobo Doll Experiment was to take the subject into another room, which was filled with interesting toys. The child was not permitted to play with these toys, being told that they were reserved for other children to play with. This was intended to build up the levels of frustration within the subject.

The child was then taken into yet another room filled with interesting toys, some of an aggressive type, some non-aggressive; the room also contained the Bobo doll and the mallet. The subject was watched through a one-way mirror, and a number of types of behavior were assessed.

The first factor measured was physical aggression, consisting of hitting the doll with the mallet or punching, kicking or sitting on the doll.

Verbal aggression was also assessed, whether it was general abuse or an imitation of phrases used by the adult role-model.

The third measurement was the amount of times the mallet was used to display other forms of aggression than hitting the doll. The final behaviors studied were modes of aggression, shown by the subject, which were not direct imitations of the role-model's behavior.

Results

The results for the Bobo Doll Experiment showed, as expected by prediction one, that children who were exposed to the aggressive model were more likely to show imitative aggressive behavior themselves.

Prediction four was proved correct in that boys were nearly three times more likely to replicate physically violent behavior than girls.

The measurements for verbally aggressive behavior again showed that children exposed to aggressive role models were more likely to imitate this behavior. The levels of verbal aggression expressed were about the same for boys and girls.

Subjects in the Bobo Doll Experiment exposed to the non-aggressive model, or no model at all, showed little imitative aggressive behavior. This finding partially proved prediction two, with children exposed to a passive role model showing less imitative aggression.

However, the results did not fully prove this prediction, as there was no discernible difference in the imitative aggression levels between groups one and three.

Male subjects exposed to non-aggressive role models were less likely to use the mallet to hit the Bobo doll. Strangely, male subjects placed with non-aggressive female models were more likely to use the mallet than the control group.

Conclusion

The findings of the Bobo Doll Experiment proved to be a little inconclusive with most of the predictions not being fully proved.

It is not certain that children learn socially, but it is likely that children observing an adult model utilizing violence are more likely to believe that this type of behavior is normal. They may, therefore, be more likely to use this type of action themselves when confronted by similar situations.

Bandura found that girls were much less likely to be physically violent, but were equally as prone to verbal aggression as boys. This is something often encountered in society, where bullying at school, by boys, is more often of a physical nature; intimidation amongst girls tends to be more verbal and social.

There were a few criticisms of the experiment; the Bobo doll springs back upright when it is hit and there is a strong possibility that the children saw it as a game rather than anything else.

There was a follow up experiment, in 1963, which used the same methodology but showed the subjects violence via video; this had a much less defined response than the initial experiment.

Another refinement of the Bobo Doll Experiment, in 1965, tried to establish the effects of rewarding or punishing bad and violent behavior. Children, who witnessed the model being punished for aggressive behavior, were much less likely to follow suit. Interestingly, there was no change in aggression when the model was rewarded for bad behavior.

In 1961, the Canadian-American psychologist, Albert Bandura (1925-) conducted a controversial experiment examining the process by which new forms of behavior - and in particular, aggression - are learnt. The initial study, along with Bandura’s follow-up research, would later be known as the Bobo doll experiment. The experiment revealed that children imitate the aggressive behavior of adults. The findings support Bandura’s social learning theory, which emphasises the influence of observational learning on behavior.

Bandura also conducted a number of follow-up studies during the 1960s which examined how witnessing a third party being rewarded or punished for behaving in a particular manner can influence a bystander’s own actions. He concluded that vicarious reinforcement, as well as direct rewards and punishments, can impact on an observer’s behavior.

Prior to Bandura’s experiments, conditioning dominated the behaviorist view of learning. During the 1890s, the influential Russian physiologist Ivan Pavlov had found that dogs would learn new behavior through classical conditioning. When a single stimulus was repeatedly paired with a particular event, such as the ringing of a bell with feeding time, salivation would begin to occur in response to the sound. Behaviorist B. F. Skinner further developed Pavlov’s theory, and proposed operant conditioning, whereby reinforcements lead to new forms of behavior being learnt.

Bandura viewed such conditioning as being reductionist in its understanding of human learning as a simple process of acquiring new ‘responses’ to stimuli. Instead, he turned his attention to the imitative behavior of children who watch, and then attempt to copy, the behavior of others.

Bandura et al (1961)

Bandura, Ross and Ross (1961) devised an experiment in which participants would observe an adult behaving in a violent manner towards a Bobo doll toy. The toys, which were popular during the 1960s, feature an image of a clown and were designed to self-right when pushed over.

The experiment took place at Stanford University, where Bandura was then working in a teaching position as a professor. The participants - children who attended the Stanford University nursery - were divided into groups. Children in one of these groups were placed in a room in which they witnessed an adult hitting a Bobo doll in an aggressive manner. They were later given the opportunity to play with the dolls for themselves.

The researchers found that the group of children who had observed an adult behaving violently towards the toy were more likely to act aggressively towards it themselves when given the opportunity. These findings indicate that learning takes place not only when individuals are rewarded or punished for their own behavior, but also when they observe another person exhibiting violent behavior - a process called observational learning.

In 1963, Bandura conducted a second experiment which replicated many aspects of the previous study. However, instead of observing an adult’s violent behavior firsthand, they watched a video of the Bobo doll being struck. As in the 1961 experiment, those participants who watched the film of a person being aggressive were more likely to behave violently towards the toy than participants in a control group. The study indicates that indirect exposure to violent behavior through film or television may lead to actions being imitated in a similar way to behavior observed in person (Bandura et al, 1963).

In a third study, Bandura tested whether the types of reinforcement that Skinner had used to encourage and discourage behavior (operant conditioning) would influence the behavior of an observer who witnessed a third party being rewarded or punished for his or her actions.

Bandura shows a film to participants in which a person again beat the toy. One group watched as the person’s behavior was reinforced by way of a food reward, whilst a second group saw a video in which the person was criticised for their violent behavior. The researchers found that the children who watched the video in which positive reinforcements were given were more likely to subsequently behave violently themselves (Bandura, 1965). This process of vicarious reinforcement suggests that learning takes place not just through direct observation, but also through the media that a person consumes.

Social Learning Theory

In 1977, drawing on his previous experimental research, Bandura outlined his social learning theory, which attempts to explain the effect of social interactions on learning. According to Bandura's theory, a person may observe the behavior of people around them.

At an early age, such people primarily consist of the parents or primary caregiver, siblings and later, classmates. A child may also observe the behavior of fictional characters on television and in films. Bandura argues that through observational learning, an individual may imitate the behavior of others. Furthermore, when a person sees another individual being punished or rewarded for their actions, their evaluation of the behavior will be further influenced, even if their own behavior has not been reinforced directly.

Aggression

One area of focus of the Bobo doll studies was the way in which children imitate the aggressive behavior of an adult. However, as the dolls were designed to be hit and pushed, and to rebound after being knocked over, some have suggested that the participants were not exhibiting aggressive behavior, but merely playing with the toy as it was intended to be used. This has led critics of the study to consider its experiment design to be flawed, as its participants were conforming to the demand characteristics of the situation.

Nonetheless, studies carried out in the decades since Bandura’s initial research have lent further support to his observation that violent behavior on-screen can influence the actions of those viewing it. Heuessman, Lagerspetz and Eron (1984) studied children’s behavior after they had watched television programs containing violence. The researchers found that the participants, and in particular males, were more likely to behave in an aggressive manner if they had seen such behavior on television.

Criticisms

The findings of Bandura’s Bobo doll experiments remain influential in the study of learning and aggression. His research furthered our understanding of how children learn from their parents, siblings and friends through imitation. It also provided early evidence that violence in films and other media can negatively influence viewers' behavior.

However, the Bobo doll studies have also drawn criticism for the methodology that Bandura and his colleagues used:

• Selection bias: The sample that Bandura used in his studies attended the nursery school at Stanford University, and so the study has been criticised for its selection bias. Participants may be expected to be from a more privileged background in terms of family education and income than the general population. Therefore, it is difficult to generalise Bandura’s findings to individuals from more diverse backgrounds.

• External validity: As the age of participants in Bandura’s experiments were in a narrow range (i.e. nursery school age), the findings lack high external validity. Whilst the observational learning that he identified may occur in children at an early age, it may be the case that the imitation of adults ceases as a person grows older. As a result, findings may not apply to the wider population.

• Critics have also raised question regarding the ethics of the methodology used in the Bobo doll experiments. A sample of children observed an adult behaving aggressively towards an anthropomorphic toy, whilst the researchers would have been aware that this behavior might be imitated by at least a proportion of the participants.