Join 3,564 readers in helping fund MetaFilter (Hide)


He keeps Bopping his Boppo!
June 1, 2008 10:08 PM   Subscribe

In 1961 Albert Bandura published a study titled "Transmission of Aggression through Imitation of Aggresive Models," better known as the Bobo Doll Experiment, in which young children were shown video of a woman beating up on an inflatable Bobo doll in various ways, the video of the woman and the results is quite interesting/shocking and sums up the general experiment quite nicely if you don't want to do too much reading.
posted by Del Far (29 comments total) 3 users marked this as a favorite

 
Regarding the video - not sure I'm buying it. The Bobo doll would have to reside somewhere in the uncanny valley for there to a correlation to a real human being. It's a doll designed specifically to be punched, and then punched again, and then punched again. If you put anyone of any age in that room they'd be more interested in playing with the doll than pretending to bake a cake.
posted by quadog at 10:22 PM on June 1, 2008


I have to think that there might have been a better way to test if children imitate adult behavior than to use a toy specifically designed to be punched repeatedly (I believe that Bobo's nose even goes "squeak") and then trying to drawn conclusions about aggression.
posted by Robin Kestrel at 10:29 PM on June 1, 2008


Regarding the video - not sure I'm buying it. The Bobo doll would have to reside somewhere in the uncanny valley for there to a correlation to a real human being.

Which is why they had a control group who didn't see the film and never exerted any violent behavior towards it.
posted by delmoi at 10:30 PM on June 1, 2008


Robin Kestrel: did you watch the flim? The second subject used another doll to assault the bobo doll. And both children used a toy gun to shoot the doll. If they were simply attracted to the dolls shape, what would have made them want to shoot it?

By the way, seeing that cute little girl in a skirt literally jump on the doll and smash it's face with a mallet is pretty damn funny.
posted by delmoi at 10:32 PM on June 1, 2008


(er, not no aggressive behavior, but much less aggressive behavior then the children who saw the models)
posted by delmoi at 10:38 PM on June 1, 2008


I recall beating up Bobo the punching doll when I was young and I never saw any adults assault him. He pretty much begs to be hit.

...Damn creepy clowns...
posted by chairface at 10:43 PM on June 1, 2008


I don't know if this teaches us much about aggression or violence. I tend to think the main cause of violence isn't the initial desire to do it, but the lack of judgment or ethics that allows one to go through with it. Haven't many previous psychological and neurological studies (particularly those involving the frontal lobe) correlated violence and aggression to poor impulse control in general?

Also, I'm not really sure if you can conflate violence against an object with violence against a person, and I think you need to use slightly older subjects. Little kids tend a bit much toward mindlessly mimicking adults at any task, whether they understand it or not. Whether the little kids are beating up the dolls due to aggression or they're just copying is hard to say.

This would perhaps be a stronger study if you showed them videos of realistic violence and then had them beat up on a guy in one of those armored suits they use with dangerous dogs... although that starts to strain ethics a little.
posted by Mitrovarr at 11:05 PM on June 1, 2008


in which young children were shown video of a woman beating up on an inflatable Bobo doll

Just for the record, according to the link above, the models were live and not on video: "After having settled the subject in his corner, the experimenter escorted the model to the opposite corner of the room. . ."
posted by Staggering Jack at 11:11 PM on June 1, 2008


DIE CLOWN DIE!!!!1!
posted by maryh at 11:11 PM on June 1, 2008


much less aggressive behavior then the children who saw the models

Another way to view this is the children who imitated the models were inspired to add other elements to the play-time, such as shooting the target with guns. That's not unreasonable, and nothing to get alarmed over. Isn't it what we'd expect a clever child to do, immitate but then innovate? Isn't that a good thing, portrayed as bad through the poor choice of stimulus?

What this experiment needs is both versions of the model interacting with the SAME TOY, one in a calm way, and one in an agitated way. And the toy should be something uncompelling, without an "expected mode of play" that the children would know or intuit. Off the top of my head, here's an alternate version of this experiment that may be less biased:

1. Group 1 (calm model) views an adult calmly playing with various toys, including a nesting doll. The adult takes the dolls apart and puts them back together several times. He/she also hides other small toys inside the dolls.

2. Group 2 (agitated model) views an adult calmly playing with various toys, including a nesting doll. The adult takes the dolls apart, then becomes agitated, stacks up the dolls like a tower, and then knocks them down with a ball repeatedly while making "bang bang" noises". The adult throws the individual dolls at the walls and floor so that they burst open like a hand grenade, and kicks the halves of the dolls around the room. The adult makes "explosion" noises while doing so.

3. Group 3 (neutral model; the control group) views an adult calmly playing with various toys, including a nesting doll. The adult does not pay particular attention to the nesting doll in favor of the other toys.
posted by Robin Kestrel at 11:15 PM on June 1, 2008 [1 favorite]


So the researchers determined that kids will imitate adults? Having a look at the data from the experiment, it seems equally plausible to conclude that kids are often aggressive, but will act less aggressive if non-aggressive behaviour is modelled to them.

Which is why they had a control group who didn't see the film and never exerted any violent behavior towards it.

The control group and the group shown aggressive behaviour both punched the doll approximately equally (and the same holds for aggressive gun play).
posted by ssg at 11:26 PM on June 1, 2008


The control group and the group shown aggressive behaviour both punched the doll approximately equally (and the same holds for aggressive gun play).

What? That's the opposite of what happened.
posted by delmoi at 12:09 AM on June 2, 2008 [1 favorite]


Delmoi -- watch the video again (or just go to 4:22). "Children in the control group never exhibited the novel forms of agression." In other words, they beat the crap out of Bobo too, they just didn't improvise beyond the standard punching and hitting.

Not much of a surprise, this. Children who were show examples of how to interact with Bobo already had a basic understanding of the rule set and started to test out interaction variations much faster than the control group. Odds are that if the control group had been exposed to Bobo a second time, they would have started exhibiting novel forms of aggression as well.

To compare it to Grand Theft Auto 4, people who have never been exposed to previous series are likely to follow the missions for a while as it teaches them the basics of the game in addition to providing an interesting narrative. Returning players, however, tend to ignore the missions at the expense of narrative, so that they can explore the world and determine the new ways that they can interact with the game engine.

Essentially, greater domain knowledge (whether it's with Bobo, GTA or chemistry) results in a greater degree of experimentation.

Color me unimpressed.
posted by bpm140 at 1:39 AM on June 2, 2008


bpm140 writes 'Delmoi -- watch the video again (or just go to 4:22). "Children in the control group never exhibited the novel forms of agression." In other words, they beat the crap out of Bobo too, they just didn't improvise beyond the standard punching and hitting.'

As I understand it, the 'novel forms' are the specific forms of aggression displayed by the woman in the film the children were shown, it doesn't refer to the kids making up new moves. In other words, all kids hit the doll, but the kids exposed to the film copied the 'novel forms' of violence; kicking the doll across the room, wrestling it to the ground and punching it in the head with both fists, etc. The guy narrating certainly makes things confusing, though, referring to the 'novel forms' of action and speech in the intro, then again later, then going on to talk about the kids improvising new types of violence... so you might be right, I dunno.
posted by jack_mo at 2:22 AM on June 2, 2008


I've always had a problem with this experiment because the Bobo doll is a DOLL. All this experiment proves is that children imitate adults and experiment with their imitations, it says nothing about violence. A simple, effective way of fixing this experiment so that it actually measures imitations of violence would be to replace the doll with an actual person. There's eight worlds of difference between an inflatable doll that squeaks playfully when you smack it and a living, breathing, person.

Of course, such an experiment would be highly unethical and could never be done, so we're back to where we started.
posted by Ndwright at 4:08 AM on June 2, 2008


Another novel form of aggression inspired by imitation.
posted by Kinbote at 5:20 AM on June 2, 2008


*plunks out "Stuck in the Middle with You" on his Playskool Piano, smashes off one half of his safety scissors, then approaches Bobo with a maniacal gleam in his eyes*
posted by adipocere at 5:29 AM on June 2, 2008


I've mentioned before that, for fun, I sometimes beat people with sticks. As such, I know an amazing number of nasty things I could do to my fellow man, some of which you can wear equipment to protect against and some you can't. Previously on Metafilter.

The implication that comes packaged with this study is that I should be expected to be a violent individual. Yet, somehow, in twenty years of doing this I have never caught myself beating someone who wasn't in armour and kind of expecting it of me or using any of the nasty brutal techniques that would cripple your playmates on any of my playmates. Or coworkers.

Of course, in terms of cognitive development, I was pretty much an adult when I started doing this. But if we could find some children being taught that sometimes it's OK to beat your little friends with a stick..... I haven't heard any horror stories in this regard, but I will be the first to state that anecdote is not the singular of data.

NOTE: This comment may contain trace amounts of self-linking.
posted by Kid Charlemagne at 6:05 AM on June 2, 2008


What? That's the opposite of what happened.

No, it's exactly what happened. The data are in the second link.

For punching, boys in the control group who never saw a model punched the doll at a rate of 15.7, though a brief scan of the article doesn't say what the units are. Boys who saw an aggressive model punched at a rate of 18.9 or 11.9 depending on the sex of the model. Statistically, these differences aren't discernible.

For "aggressive gun play," the control group rate for boys was 14.3. The aggressive-model rate was 7.3 for a female model and 15.9 for a male model. In other words, if they saw a woman bashing the doll around, they were less likely to point a gun at it, though the article doesn't seem to test for significance in the wrong direction. Interestingly, the same is true for nonaggressive models.

Frankly, bringing up the issue of gun play in the filmed discussion when there was no statistically discernible effect there is not even borderline dishonesty; it's a plain and simple lie.
posted by ROU_Xenophobe at 7:35 AM on June 2, 2008


Which is why they had a control group who didn't see the film and never exerted any violent behavior towards it.

This study showed that children who were shown the intended use of a Bobo doll tended to use it more in the intended way. It shows very little about violence. Let's say there was a little dumbbell for curls in the room, and some children were shown an adult doing curls while some children were controls. The test children would probably do more curls than the control children. but would anyone try to draw anything profound from that?

This is why no one likes the soft sciences.
posted by TheOnlyCoolTim at 8:49 AM on June 2, 2008 [1 favorite]


All I want to know is this: can I cite this study the next time I beat up a clown?

"But officer, I'm doing it for Science!"
posted by quin at 9:55 AM on June 2, 2008


Of course, such an experiment would be highly unethical and could never be done, so we're back to where we started.

I think our scientific research ethics have become too namby-pamby, e.g. I quite frankly think that the Milgram experiment was fine and that if you feel emotional distress at finding out that you're a Good German, then good, have a learning experience. (And, notably, 92% of Milgram's subjects responded to a survey some years later with 84% of them viewing it as a positive and 15% viewing it as neutral.)

So I think we could have an interesting scientific experiment seeing if a five year old will beat up a burly volunteer. It would probably be unethical, however, to attempt to expand on this experiment to win Ask Metafilter by seeing how many five year olds a burly individual can beat up.
posted by TheOnlyCoolTim at 11:02 AM on June 2, 2008


This study showed that children who were shown the intended use of a Bobo doll tended to use it more in the intended way.

If by the intended way, you mean punching the doll, then no, they didn't. The data simply doesn't back up that assertion.
posted by ssg at 1:45 PM on June 2, 2008


I meant more its general use as a target or object to be agressive towards, the throwing around the room and all that which I guess they picked up more statistically significantly than just the punching.
posted by TheOnlyCoolTim at 1:53 PM on June 2, 2008


Yes, they did find statistically significant imitative aggression, but no statistically significant non-imitative aggression versus the control group. So yes, I think your initial point is correct: kids imitate adults. Violence has absolutely nothing to do with it.
posted by ssg at 2:30 PM on June 2, 2008


For punching, boys in the control group who never saw a model punched the doll at a rate of 15.7, though a brief scan of the article doesn't say what the units are. Boys who saw an aggressive model punched at a rate of 18.9 or 11.9 depending on the sex of the model. Statistically, these differences aren't discernible.

Those aren't even statistics which can be spoken of in "significance terms" - those are F scores, which are then translated into significance in Table 2.

Aggressive vs. Nonaggressive: p <>Aggressive vs. Control: p <>Nonaggressive vs. Control: p = .09

This means that Control and Nonaggressive groups are statistically significantly different than the Aggressive condition group in exhibiting Physical aggression imitative behaviour. In laymen's terms: the aggressive group produced more physical aggression than both the nonaggressive and the control groups; when the children see the adult models displaying aggressive behaviour, they produce more themselves than if they had otherwise seen a nonaggressive adult play with the doll or had no model at all.
posted by tybeet at 10:49 AM on June 3, 2008


(erm.. this was garbled in previous comment)

Aggressive vs. Nonaggressive: p < .001
Aggressive vs. Control: p < .001
Nonaggressive vs. Control: p = .09
posted by tybeet at 10:50 AM on June 3, 2008


If by the intended way, you mean punching the doll, then no, they didn't. The data simply doesn't back up that assertion.

The data certainly backs the assertion. You're missing the statistical interpretation entirely.
posted by tybeet at 10:53 AM on June 3, 2008


The data certainly backs the assertion. You're missing the statistical interpretation entirely.

Table 2 indicates that there is no significant difference between any of the groups for punching the doll. The data is right there, under "Punches Bobo doll" and clearly indicates that there was no significant difference (that's what ns means).

This means that Control and Nonaggressive groups are statistically significantly different than the Aggressive condition group in exhibiting Physical aggression imitative behaviour.

This is true.

In laymen's terms: the aggressive group produced more physical aggression than both the nonaggressive and the control groups; when the children see the adult models displaying aggressive behaviour, they produce more themselves than if they had otherwise seen a nonaggressive adult play with the doll or had no model at all.

But this is misleading. The important point is not that they were more aggressive (which is true), but that they were not more aggressive in a non-imitative way than the control group. In other words, the kids imitated the adult models, but the modelling of aggressive behaviour did not make them behave more aggressively in general.
posted by ssg at 11:19 AM on June 3, 2008 [1 favorite]


« Older The continuity I have in mind has to do with the n...  |  Two Spanish women meet in thei... Newer »


This thread has been archived and is closed to new comments