Not up to snuff:
The Ethical Issues in the SPE
Any discussion on the failings of an experiment should begin with a discussion of the ethical issues. Ethics guide researchers toward valid results via set standards of experimental design. These standards dictate everything from how questions are worded to how procedures are designed. There are several aspects of the Stanford Prison Experiment which violate the ethical standards of research today.
Sample
The first area which could be improved upon is the makeup of the subject pool. Zimbardo was interested in showing that the environment people are in is heavily influential on their personality and behavior. Despite his overarching desire to show this was true in all types of people, he included only college-aged males. It would have been fine to do so if he was only interested in saying that men are influenced heavily by their environment, but that was not his goal. He may have had plans to do this with women, and different age groups, but from reading his book there is no indication that this was his plan. Another issue with his sample, is that they were all college students. During college, people are still figuring their life out and are more likely to change their personality or behavior based on the situation they are in than older people with stable careers. Another problem with their sample is that it was not reflective of an actual prison, where age and other demographics vary greatly between prisoners.
Demand Characteristics
The design is also flawed in that the confederates (people hired by the research staff, but not a part of the actual study) and researchers played the parts of the warden and upper-level staff. This would have created many issues. Firstly, this would have induced demand characteristics of the guards and maybe even the prisoners. They would have seen if Zimbardo wanted more cruelty from the guards or expected more fear from the prisoners and might have unconsciously changed their behavior to meet his expectations. There were also accusations that Zimbardo and his research team coached the guards. If they had a participant playing the role of warden, they would not have been tempted to do so, and their results would be much more meaningful and genuine. This would have also allowed them to see the effect of power in a situation like the prison.
What Should have happened...
When designing this experiment, Zimbardo could have done a few things differently which would have made his results more valid.
Firstly, he should have canvased for a more diverse sample, rather than sticking to the convenience of the college campus. He could have done this with flyers posted around town and in businesses off campus to draw in a different crowd. This would have made his condition both more externally valid and more realistic of an actual prison.
To solve the issue of demand characteristics, it would be even simpler. He would only have had to recruit a few more participants to assume the roles of warden and other higher-up staff. This would have reduced the risk of demand characteristics as well as shown more clearly what people will do when put in a situation such as prison at varying levels of power.
