4.1 Core study 1: Bandura et al.

(aggression)
Bandura, A., Ross, D., & Ross, S. A. (1961). Transmission of aggression
through imitation of aggressive models. Journal of abnormal and social
psychology, 63(3), 575–582.

The psychology being investigated

Children copy adults. One reason for this is because the immediate
social setting encourages the child to imitate what he or she is watching
(Figure 4.1). This helps influence their behavior, making it more likely
that the child will do what others are doing around them. Alternatively,
the observation of a behavior can lead a child to acquire a new response
that he or she could reproduce independently. If this is the case, the new
behavior should generalize to new settings and so would be produced in
the absence of an adult model.

Albert Bandura had previously conducted experiments on social

learning and was interested in studying social learning in the context of
aggression. To learn from others, the observer (e.g., a child) must be
paying attention to the behavior of a model. They must retain
(remember) the behavior they have observed, in order to reproduce it.
When social learning occurs, it could potentially lead to either aggressive
or non-aggressive behavior. Bandura expected then that watching an
aggressive model should lead to more aggressive behaviors being
demonstrated, and that observing a non-aggressive model should lead
to more non-aggressive behavior being produced, i.e., even less
aggressive behavior than normal.

Key Words

• Model: a person who inspires or encourages others to

imitate positive or negative behaviors
• Social learning: the learning of a new behavior that is
observed in a role model and imitated later in the absence of that model
• Aggression: behavior that is aimed at harming others either
physically or psychologically
4.1 Nghiên cứu cốt lõi 1: Bandura và cộng sự (hành vi hung hăng)
Bandura, A., Ross, D., & Ross, S. A. (1961). Sự truyền tải hành vi hung
hăng thông qua việc bắt chước các mô hình hành vi hung hăng. Tạp chí
tâm lý bất thường và xã hội, 63(3), 575–582.

Tâm lý học được điều tra

Trẻ em bắt chước người lớn. Một lý do cho điều này là do môi trường xã
hội xung quanh khuyến khích trẻ bắt chước những gì mà chúng đang
quan sát (Hình 4.1). Điều này ảnh hưởng đến hành vi của chúng, làm
tăng khả năng trẻ sẽ làm những gì người khác đang làm xung quanh.
Ngoài ra, việc quan sát một hành vi có thể dẫn đến việc trẻ tiếp thu một
phản ứng mới mà chúng có thể tự tái hiện. Nếu đúng như vậy, hành vi
mới sẽ được khái quát hóa sang các bối cảnh mới và sẽ được sản sinh
mà không cần có sự hiện diện của một mô hình người lớn.

Albert Bandura trước đây đã thực hiện các thí nghiệm về học tập xã hội
và quan tâm đến việc nghiên cứu học tập xã hội trong bối cảnh hành vi
hung hăng. Để học từ người khác, người quan sát (ví dụ: một đứa trẻ)
phải chú ý đến hành vi của một mô hình. Chúng phải lưu giữ (ghi nhớ)
hành vi mà chúng đã quan sát được, để có thể tái hiện lại. Khi học tập
xã hội diễn ra, nó có thể dẫn đến hành vi hung hăng hoặc không hung
hăng. Bandura cho rằng việc quan sát một mô hình hung hăng sẽ dẫn
đến việc biểu hiện nhiều hành vi hung hăng hơn, trong khi quan sát một
mô hình không hung hăng sẽ dẫn đến việc biểu hiện nhiều hành vi
không hung hăng hơn, tức là hành vi ít hung hăng hơn bình thường.

Từ khóa

• Mô hình (model): một người truyền cảm hứng hoặc khuyến

khích người khác bắt chước các hành vi tích cực hoặc tiêu cực
• Học tập xã hội (social learning): việc học một hành vi mới
được quan sát từ một mô hình và bắt chước lại sau khi mô hình đó
không còn xuất hiện
• Hành vi hung hăng (aggression): hành vi nhằm làm hại
người khác về thể chất hoặc tâm lý
Background

Previous research had shown that children imitated the behavior of a

model when in the presence of the model. Bandura et al. wanted to
investigate whether social learning theory could be used to explain
aggression, specifically when the child was no longer in the presence of
an aggressive model.

Children are rewarded in different ways for imitating adults. In general (at
least, in the USA in the 1960s when this study took place), boys were
rewarded for behaviors considered to be sex-appropriate and punished
for inappropriate ones, such as cooking or “playing mother.” Similarly, for
girls, rewards and punishments would be applied to discourage what
was considered sex-inappropriate behaviors. This, Bandura et al.
suggested, would lead to two kinds of differences.

First, boys and girls should be more likely to imitate same-sex models
and, second, they should differ in the readiness with which they imitate
aggression, with boys doing so more readily as this was seen as a more
masculine-type behavior.

Aim

The aim was to investigate whether a child would learn aggression by

observing a model and would reproduce this behavior in the absence of
the model, and whether the sex of the role model was important.
Specifically, there were four hypotheses:
1. Observed aggressive behavior will be imitated, so children
who see an aggressive model will be more aggressive than those seeing
a non-aggressive model or no model.
2. Observed non-aggressive behavior will be imitated, so
children seeing non-aggressive models will be less aggressive than
those seeing no model.
3. Children are more likely to copy a same-sex model.
4. Boys will be more likely to copy aggression than girls.

Issues and Debates

 • Nature versus nurture describes the importance of the
environment, or nurture, which can be seen in the role adults take as
models for children, as well as in the rewards and punishments adults
give to children.

Key Words

• Nature: innate, genetic factors which influence behavior

• Nurture: environmental influences on behavior

Method

Research method and design

This was a laboratory experiment; the environment was not the normal
place where the children played, and the situation was controlled. The
design of the experiment was that of independent measures as different
children were used in each of the levels of the independent variable.
Variables (IVs)

Although these children were matched for aggression in threes (see the
Research Methods box), there were three IVs:
• Model type: whether the child saw an aggressive model, a
non-aggressive model, or no model.
• Model sex: same sex as the child (boys watching a male
model and girls watching a female model) or different sex (boys
watching a female model and girls watching a male model).
• Learner sex: whether the child was a boy or a girl.

The dependent variable (DV) was the behavior the child displayed. This
was measured through a controlled observation of the children, and
measures of aggressive behavior were recorded.

Research Methods
In the matched participants design described here, the participants were
divided into threes, all with very similar initial aggression levels. One of
each of these individuals was placed into each of the three different
conditions of model type (aggressive model, non-aggressive model, and
control).

Sample

Seventy-two children aged 3–6 years (36 boys and 36 girls) were
recruited from Stanford University nursery school.

Procedure

Prior to the experimental part of the study, the children were observed in
their nursery school by the experimenter and a teacher who knew them
well. They were rated on four different measures of physical aggression,
verbal aggression, aggression to inanimate objects, and aggression
inhibition (anxiety), each on a five-point scale. They were then assigned
to three groups, ensuring that the aggression levels of the children in
each group were matched.

Of the 51 children rated by both observers (the rest were rated by only
one observer), similar ratings were generally produced. Their ratings
were compared as a measure of inter-rater reliability, which showed a
high correlation between the observers, of r = 0.89.
Research Methods

Inter-rater reliability is the extent to which two researchers rate the

same activity that they have observed, heard in an interview, etc., in the
same way. This is judged using a correlation (an r value) between the
two ratings, which will be high (close to 1) if they are reliable.

Twelve boys and 12 girls were allocated to control groups who saw no
model. The remaining children were divided equally by sex between
aggressive and non-aggressive model groups and, within those,
between same- and opposite-sex models.
The experimenter and child entered the observation room, where the
experimenter showed the child to a table and chair in their “play area,”
where they were shown how to make potato prints and sticker pictures:
activities previously identified as interesting for children.

The opposite corner of the room also contained a table and chair, a
Tinkertoy set (a wooden building kit), a mallet, and a five-foot (152 cm)
Bobo doll: an inflatable clown-like doll that bounced back when hit
(Figure 4.2). This is where the model sat, in those conditions where
there was one. The experimenter remained in the room so that the child
would not refuse to be alone or try to leave early, but they appeared to
be working quietly at their desk.

Figure 4.2: Observing and imitating aggressive behaviors with a

Bobo doll

The three groups were then treated differently. In the non-aggressive

condition, the model assembled the Tinkertoys for 10 minutes. In the
aggressive condition, this lasted only 1 minute after which the model
attacked the Bobo doll.

Here is the text extracted and translated into Vietnamese:

The Bobo Doll Experiment Details

The doll was laid on its side, sat on, and punched in the nose, picked up
and hit on the head with a mallet, tossed up in the air, and kicked. This
sequence was performed three times over 9 minutes, accompanied by
aggressive comments such as “Kick him” and two non-aggressive
comments such as “He sure is a tough fella.” Of children in the model
groups, half saw a same-sex model, and the others saw a model of the
opposite sex. A control group did not see any model and, therefore, saw
no aggression.

The experimental procedure continued with a stage in which all

participants were deliberately mildly annoyed. This was done for two
reasons:
 • Because watching aggression may reduce the production of
aggression by the observer (even if it has been learned) and it was
necessary to see evidence of learning.
• To ensure that even the non-aggressive condition and
control participants would be likely to express aggression, so that any
reduction in that tendency could be measured.

Research Methods

The non-aggressive model group might appear to be a control group

because the key factor of “aggression” is missing. However, the
important aspect is the presence of a model—and in the non-aggressive
group, a model was still present. This means the real control group is
where the IV is absent: the group where there is no model at all.

A test of the child’s aggression then followed in which the child was
observed for 20 minutes using a one-way mirror. The mirror appeared
transparent on the researcher’s side (so they could observe behavior)
but appeared as a normal mirror on the child’s side (so they could not
see that they were being observed from another room). For the
aggressive model group, this was a test of delayed imitation.

The experimental room contained a three-foot (92 cm) Bobo doll, a

mallet and peg board, two dart guns, and a tether ball with a face
painted on it which hung from the ceiling. It also contained some non-
aggressive toys, including a tea set, crayons and coloring paper, a ball,
two dolls, three bears, cars and trucks, and plastic farm animals. These
toys were always presented in the same order.

Children’s Behavior Observation and Measures

The children’s behaviors were observed in 5-second intervals (240

response units per child). There were three response measures of the
children’s imitation, with a range of possible activities in each:
• Imitative physical aggression: striking the Bobo doll with
the mallet, sitting on the doll and punching it in the nose, kicking the doll,
and tossing it in the air.
 • Imitative verbal aggression: repetition of the phrases,
“Sock him,” “Hit him down,” “Kick him,” “Throw him in the air,” or “Pow.”
• Imitative non-aggressive verbal responses: repetition of
“He keeps coming back for more” or “He sure is a tough fella.”

Partially imitative aggression was scored if the child imitated these

behaviors incompletely. The two behaviors here were:
• Mallet aggression: striking objects other than the Bobo doll
aggressively with the mallet.
• Sits on Bobo doll: laying the Bobo doll on its side and
sitting on it, without attacking it.

Two further categories were:

• Aggressive gun play: shooting darts or aiming a gun and
firing imaginary shots at objects in the room.
• Non-imitative physical and verbal aggression: physically
aggressive acts directed toward objects other than the Bobo doll and
any hostile remarks except for those in the verbal imitation category
(e.g., “Shoot the Bobo,” “Cut him,” “Stupid ball,” “Horses fighting, biting,”
“Knock over people”).

Finally, behavior units were also counted for non-aggressive play and
sitting quietly not playing at all, and records were kept of the children’s
remarks about the situation.

The male model scored all the children’s behaviors. Except for those
conditions in which the male was the model, he was unaware of which
condition the child had been in. However, the condition in which the child
had been was visually obvious, as in the case of the aggressive model
children as they performed the specific behaviors exhibited by the
model. To test the reliability of the scorer, a second scorer independently
rated the behavior of half of the children. The reliability score was high at
approximately r = 0.9 for different categories of behavior.
Results

Children exposed to aggressive models imitated their exact behaviors

and were significantly more aggressive, both physically and verbally,
than those children in the non-aggressive model or control groups. One-
third of the children in the aggressive condition also copied the model’s
non-aggressive verbal responses, but none of the children in either the
non-aggressive or control groups made such remarks. The mean
aggression scores can be seen in Table 4.1.

The mean for imitative physical aggression for boys with a male model
(25.8) was much higher than that for girls (7.2). This indicates that the
boys imitated the physical aggression of a male model more than the
girls did. However, with a female model, girls imitated less (5.5) than with
the male model. Girls imitated more verbal aggression of the same-sex
model than boys (although not significantly so). Children were also more
likely to imitate a same-sex model than an opposite-sex model; this
effect was stronger for boys than for girls.

Children seeing a non-aggressive model were much less likely than

either the aggressive model group or controls to exhibit mallet
aggression, and this pattern was especially apparent for girls. Although
the aggressive model did not appear to affect levels of gun play or
punching the Bobo doll, non-imitative physical and verbal aggression
other than these activities were higher following exposure to an
aggressive model compared to the other two conditions.

There were also differences in non-aggressive play. Girls played more

with dolls, tea sets, and coloring, and boys engaged in more exploratory
play and gun play. There were no sex differences in play with farm
animals, cars, or the tether ball. Both boys and girls seeing the non-
aggressive model engaged in more non-aggressive play with dolls than
either of the other groups and spent more than twice as much time
sitting quietly, not playing.

Activity 4.1

Consider the data for the mean aggressive behaviors and the non-
imitative verbal responses.
1. Which data is qualitative and which is quantitative?
 2. Draw a bar chart showing the results for imitative physical
aggression. Include the mean scores for the experimental groups and
the control group. Make sure you give the chart labels and an
appropriate title.

Issues and Debates

In the nature versus nurture debate, we can consider why the boys and
girls showed different responses. This could be because they are
genetically different; a nature explanation. Boys might be biologically
predisposed to be aggressive, so more likely to copy aggressive models.
Alternatively, boys might be more likely to copy aggressive models
because they have been rewarded for aggressive behaviors more than
girls have. This would be a nurture argument.

Additional Observations and Remarks

In addition to the observations, records of the remarks about the

aggressive models revealed differences, both between reactions to the
actions of the male and female models and between boys and girls.
Some comments appeared to be based on previous knowledge of sex-
typed behavior, such as:
• “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
just acting like a man. I never saw a girl act like that before. She was
punching and fighting but not swearing.”

Whereas comments about the female model’s behavior were

disapproving, those about the male model were not. This was more likely
to be seen as appropriate and approved by both boys and girls, for
example in comments such as:
• “Al’s a good socketer; 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.”
Key Word

• Sex-typed behavior: actions that are typically performed by

one particular sex and are seen in society as more appropriate for that
sex. For example, aggression is seen as masculine-type behavior and
was more commonly imitated by boys in the study.

Conclusion

The results strongly suggest that observation and imitation can account
for the learning of specific acts without reinforcement of either the model
or observer. All four hypotheses were supported:
• Observed aggressive behaviors are imitated: children who
see aggressive models are likely to be more aggressive than those
seeing a non-aggressive model or no model.
• Observed non-aggressive behaviors are imitated: children
seeing non-aggressive models will be less aggressive than those seeing
no model.
• Children are more likely to copy a same-sex model, although
this may depend on the extent to which this behavior is sex-typed.
• Boys are more likely to copy aggression than girls.
Strengths and Weaknesses

The main method was a laboratory experiment. This means that it was
possible to control extraneous variables such as ensuring there was a
possibility that the children in any condition would show aggressive
behavior. This was done by showing them nice toys but then taking them
to another room. Also, all children in both experimental groups saw a
model for the same length of time, and in each condition their behaviors
were standardized. This means the research was more valid—the
researchers could be sure that the differences in results between
conditions were due to the differences between the models—and more
reliable, because each child within a condition experienced exactly the
same exposure. Inter-observer reliability was also checked for both
the initial observations of aggressiveness and for the data recording, and
was very high.

The pre-testing of the children’s aggressiveness as part of the matched

pairs design was another factor that increased validity because it
ensured that differences between conditions were due to the models and
not to individual differences between the children who happened to be in
each group.
The main measure of the DV was through observation. As the observers
were behind a one-way mirror, the children were unaware that they were
being watched. This increases validity as they were likely to behave
naturally rather than responding to demand characteristics as they
might have done had they known they were being observed. The
observation period was divided into time intervals (of 5 seconds), and
the categories were clearly defined (e.g., imitative and non-imitative
behaviors), which also helped to improve both validity and reliability.

Among the weaknesses of the study is that only six children were used
in each experimental condition and, although they were matched to
reduce the risk of participant variables confounding the results, it is still
a small sample. Furthermore, it is possible that the children were quite
similar, as they all attended the same nursery.

Key Word

• Extraneous variable: this either acts randomly, affecting the

DV in all levels of the IV, or systematically (i.e., on one level of the IV,
called a confounding variable) so can obscure the effect of the IV,
making the results difficult to interpret.

Ethical Issues

One ethical issue with the study was some of the children might have
been harmed by becoming more aggressive during the research. For
example, they could have physically injured themselves with the toys
they were given to play with after watching an aggressive model. Even if
this were not the case, the children were still deliberately annoyed in the
procedure of the study. This could have been psychologically distressing
for the children. These aspects of the study go against the ethical
guideline of protecting participants from physical and psychological
harm.

Summary
Bandura et al.’s study used adults being aggressive to a Bobo doll to
show that children’s behavior can be affected by the behavior of a
model. Exact aggressive behaviors were imitated although the study
showed that non-aggressive modeling was also effective. Children were
more likely to copy a same-sex model. Additionally, boys engaged in
more aggressive imitation than girls. This was a well-controlled
laboratory experiment measuring the dependent variable through
objective observations that were reliable. Qualitative data suggested that
the children recognized sex-typing and were surprised by behavior that
did not fit the pattern. The findings suggest practical applications for
protecting children, e.g., through film certification and the use of parental
controls on media devices.

Questions

1. Why might Bandura et al. have chosen to record the

children’s behavior in 5-second intervals, rather than 1-second or 1-
minute intervals?
2. Consider the data for the mean aggressive behaviors and
the non-imitative verbal responses:
a. Which are qualitative data?
b. Which are quantitative data?
3. The procedure was standardized in many ways:
a. Identify the ways in which the procedure was standardized.
b. Explain why standardization was necessary—what might have
happened if each participant had not been treated in the same way?