Haney, Banks and Zimbardo (prison simulation)
Psychology Essay - Sampling
Haney, Banks and Zimbardo (prison simulation)
i)
The method of selecting the sample for the prison simulation study was very extensive - it began very simply, with a newspaper advert offering $15 a day to take part in a psychological study on 'prison life', and progressed from there. 75 potential subjects responded to the ad, and of these, 22 were finally selected to take part in the study.
They were made to complete an extensive questionnaire, designed to find out about their family background, physical and mental health history, and their experiences with and attitude towards psychopathological tendencies (including any criminal history). Every respondent was also interviewed by one of the experimenters. The final selection were chosen because they appeared to the experimenters to be the most mature and mentally and physically stable, and the least likely to become involved in antisocial behaviour. Therefore, they were chosen on the basis of their 'normality', to demonstrate the effects that prison life could have on apparently ordinary, non-criminal people.
The subjects were described as 22 'normal, healthy male college students', all of who happened to be in the Stanford area during the summer the study was carried out. They were mostly of middle-class socio-economic status, and Caucasian (with the exception of one Oriental man), and prior to the study, they were all strangers. This was to avoid any pre-existing friendship tendencies, which might have arisen during the study.
This is clearly not a representative sample - the subjects were all male, almost entirely Caucasian, and largely middle-class. There were also not enough subjects for the sample to be representative - with only 22 subjects taking part in the study, it was unlikely that they would represent a large percentage of the world.
ii)
The sampling method was, at least in my opinion, a very successful one - it was incredibly thorough (with an extensive questionnaire, and an interview by one of the experimenters), and therefore provided exactly the kind of sample that Haney, Banks and Zimbardo were looking for. The fact that the sampling method was so thorough was more than likely an influence on the results - it certainly made the results more reliable, as no one could say that the subjects already had violent and antisocial tendencies before they became part of the study's mock prison. In ...
This is a preview of the whole essay
ii)
The sampling method was, at least in my opinion, a very successful one - it was incredibly thorough (with an extensive questionnaire, and an interview by one of the experimenters), and therefore provided exactly the kind of sample that Haney, Banks and Zimbardo were looking for. The fact that the sampling method was so thorough was more than likely an influence on the results - it certainly made the results more reliable, as no one could say that the subjects already had violent and antisocial tendencies before they became part of the study's mock prison. In this particular study, the sample was much more important than most, because with a less carefully chosen sample, the results could have been entirely different, and far less remarkable.
Another strength of the sampling method is that it was very ethical - all the subjects knew what the study was about, and they were never lied to. With the slight exception of the prisoners, they all gave informed consent (the prisoners were not told they would be arrested at their homes at the start of the study). Although the rest of the study was certainly less than ethical - for example, the prisoners had the threat of potential physical harm from other subjects, and certainly suffered from mental harm - the sampling method was very honest and decent.
One of the only real weaknesses of the sampling method was the fact that it did not produce a representative sample, and there was no way it could have done. From the beginning, the newspaper ad asked for only male participants, proving it impossible for the sample to become truly representative. It was also only applicable to students in the Stanford area, which erases quite a large portion of the world. It then went on to narrow down the already-limited sample to people who were mature, non-violent, and mentally and physically stable - and in the end, produced an incredibly limited selection of people.
iii)
There are many problems with sampling in psychological studies, the most evident of which is the almost impossible task of getting a representative sample. No matter how close the sample is to being representative - it could include thousands of people, of all ages, genders, and races - there would undoubtedly be something about it that made it unacceptable as a representative sample. The only way to do it, really, would be to subject every person in the world to the same psychology study - which is essentially impossible, and definitely a waste of time.
This problem is made evident in almost every study that we've investigated so far - for example, Loftus and Palmer's study on automobile destruction. In their second experiment, they used a wide range of subjects - a hundred and fifty - with a fairly equal variation of genders. But the sample was not representative, because every one of the subjects was a student, and therefore they did not cover a wide enough range of people.
Another problem with sampling is the possibility that the results are very little to do with the actual study, and much more to do with the sample that is selected. Because most samples are fairly small (the previously mentioned sample in Loftus and Palmer's experiment is probably one of the biggest by quite a long way), it's possible that the results gained during the study are based on the fact that a small amount of people think or behave the same way entirely by chance. If the study were done on a larger amount of people, and a wider range, the results could be completely different.
A good example for this is Henri Tajfel's study on intergroup discrimination. When Tajfel performed the study on a group of just over 60 teenaged boys, he found that almost every subject discriminated against the people in the opposing group. However, when we performed a small-scale version of this study in class, we found absolutely no evidence of what Tajfel found - most people assigned their points randomly, and the results were due to chance. Therefore, the actual sample greatly affects the final results.
iv)
I believe the Haney, Banks and Zimbardo study is one that would be most greatly affected by a change in the sample, as the results depend entirely on the individual personalities of the subjects, and the way they take on their role. I think it would be most interesting to perform this study on a selection of females - perhaps between the ages of 18 and 25. There are several ways in which this could affect the results.
It could result in the outcome becoming even more shocking, due to the fact that women are more likely to gang up on each other and pick fights over nothing, especially if they are spending an abnormal amount of time together. It's possible that it would end up in a small-scale war between the guards and the prisoners, with each group firmly uniting against the other, and the abuse going beyond the level it did with males. Whereas men are more likely to deal with issues like that alone, women are far more likely to group together.
It could also result in the guards ignoring the prisoners, and the prisoners socialising with each other. As women are generally more open and talkative than men, I find it much easier to believe that they'd spend their time in the cells talking amongst themselves and discussing their lives outside the study, rather than talking of little else other than the guards' treatment of them, as the men did. This would prove the original hypothesis completely wrong, confirming that it was the people inside the prison who make it a bad place to be, and not the general atmosphere.
Finally, the change in sample could result in complete interaction between the prisoners and the guards, splitting off into separate social groups which did not concern the study. As women who are bored tend to start talking as quickly as possible (women at bus stops, or in queues, for example), it would be unsurprising to see that the roles made no difference whatsoever to who interacted with who.