Strengths and limitations

The outline asks for at least three of each. Below are more than three, so you can choose the ones that fit your argument. For each one, the reason is given, because a strength or limitation without a reason earns little credit.

Strengths

1. Strong control, so the cause is clear

Everything except the model's behaviour was kept the same: the rooms, the toys and their positions, the timing, the frustration step and the experimenter's script. The aggressive model performed the same sequence of acts with every child. Because of this control, and because children were assigned at random, the difference between groups can be traced to what the children watched. This is high internal validity, meaning confidence that the study measured a real cause and effect within its own setting.

2. Groups matched on existing aggression

The researchers did not rely on chance alone with groups of six. They rated every child's aggressiveness beforehand and matched children in triplets. This removes the most obvious alternative explanation: that one group simply contained more aggressive children from the start.

3. Unusual acts made learning visible

This is the study's cleverest feature. Ordinary punching tells us little, because children punch dolls without being shown. So the model performed acts children were unlikely to invent: sitting on the doll to punch its nose, hitting its head with a mallet, saying "He keeps coming back for more". When children reproduced these exactly, the only reasonable source was the model. The results confirm the logic: ordinary punching did not differ between groups, but the unusual acts did.

4. Clear, reliable measurement

The behaviour categories were defined in advance and were concrete. Two observers agreed closely, with correlations in the .90s. Time sampling every five seconds gave detailed, numerical data that could be compared across groups.

5. Replicable, and replicated

Because the procedure was described so precisely, it could be repeated. Bandura and his colleagues used the same paradigm in later studies with filmed and cartoon models (1963) and with models who were rewarded or punished (1965), and the copying effect appeared again each time. Other researchers later used filmed models with real people as victims (for example Hanratty, O'Neal and Sulzer, 1972).

Limitations

1. A toy built to be hit

A Bobo doll is designed to be punched and to bounce back. Hitting it may be play, not aggression. Most definitions of aggression require an intention to harm a target that would want to avoid harm, and a Bobo doll can be neither. Tedeschi and Quigley (1996) argued that the Bobo paradigm may not measure aggression at all, but imitation of rough-and-tumble play (as quoted by Ferguson, 2020). So the study shows that children copy, which is important, but its link to aggression towards people is indirect.

2. Demand characteristics

Children may have worked out what the adults expected. They had just watched an adult hit a Bobo doll, then were taken to a room with a Bobo doll, with an adult present who did not object. Copying may partly reflect the children's reading of what was wanted or allowed in this place, rather than learning that would carry over to ordinary life.

3. Short term only

Behaviour was measured once, a few minutes after watching. The study cannot say whether the learning lasted, or whether it would ever appear outside the laboratory.

4. Scoring was not blind

The male model scored all 72 sessions, and he knew the condition of the children he had modelled for. The second observer scored only half the children. The authors admit that children who had seen the aggressive model "could be readily identified" by their behaviour. Observers who know or can guess the condition may, without meaning to, score borderline acts in the expected direction. This is called observer bias. The very high agreement between observers reduces this concern but does not remove it.

5. Small groups and many comparisons

Each experimental group had six children. Several conclusions, especially about the sex of the model and the calm male model, rest on comparisons between these small subgroups, and some of those differences only just reach the usual standard for being called reliable. The authors themselves warn that such comparisons may "capitalize on chance". With one man and one woman as models, differences between "male" and "female" models could also be due to those two individuals.

6. A narrow sample

All 72 children came from one university nursery school in California. The paper gives no information on social class or ethnicity. The results may not apply to children in other social classes, cultures or countries, or to older children. Section D looks at this in detail.

7. Ethical problems by today's standards

The paper reports no consent procedure, no debriefing and no follow-up. Children were deceived, deliberately frustrated by having toys taken away, shown adult aggression and possibly taught new aggressive acts. They were observed without knowing. Under current codes, such as the British Psychological Society's Code of Human Research Ethics and the American Psychological Association's ethics code, a study like this would need parental consent, the child's own assent (agreement in a form suited to their age), a real chance to stop, minimal deception, and a debriefing that undoes any harm. Formal ethical review of research in the United States only became a requirement from the mid-1960s, so judge the researchers against their time as well as ours.

Weighing strengths against limitations

Notice how the strengths and limitations connect. The tight control that gives the study its internal validity is the same thing that makes it artificial. A toy built to be hit made unusual acts easy to see and count, and is also the main reason critics doubt that the behaviour was aggression. Pointing out connections like these is what turns a list into an evaluation.

Try this

You have space for only one strength and one limitation in your journal, and you want them to be linked.

Which pair would you choose, and how would you connect them?

Show the reasoning

A strong choice is "unusual acts made learning visible" as the strength and "a toy built to be hit" as the limitation. The study's design used a toy that invited a specific, unusual set of acts, and the precise copying of those acts is the best evidence of observational learning in the paper. But the same toy is the reason critics argue that the behaviour was play rather than aggression.

So the connected argument is: the design gives convincing evidence that children learn new acts by watching, and weaker evidence that they learn aggression in the sense of intent to harm another person. Another strong pair is "strong control" with "short term only and artificial setting", the classic trade-off between internal validity and external validity, which means how far results apply outside the study.