Aims: To establish the extent to which behaviour is shaped by stereotyped expectations of role.
Procedure: A set of participants, screened to ensure a "normal" psychological profile, are randomly separated into one of two groups: "prisoners" and "guards". The "prisoners" are subject to a humiliating "arrest" procedure, and placed in a mock-prison block, with the "guards" having control over them. While no physical violence is permitted, the guards are otherwise free to exercise authority over the prisoners in order to run the prison in an effective manner.
Findings: Within a very short space of time, the participants began to act out their roles, with the prisoners initially attempting a rebellion and then falling into passivity, and the guards becoming increasingly authoritarian. Before the experiment was halfway through its planned fourteen day duration, it had to be abandoned, as the prisoners were showing signs of severe stress.
Conclusion: While it may not be possible to determine to what extent the behaviour shown was due to participants simply "acting out" a role as opposed to actually experiencing it, nevertheless the experiment shows that behaviour can be very strongly influenced by the expectations that people have of specific social situations.
Strength: The fact that the "prisoners" and "guards" were randomly chosen from a pool of similar candidates, would seem to indicate that it was the situation, rather than participants' personality, which influenced the subsequent behaviour.
Weakness 1: The experiment was only carried out once by Zimbardo, which means that the results were not replicated. It may be that in this study, the random selection of guards just happened to pick participants who were predisposed to abuse their position. In fact a subsequent recent attempt to re-play the a similar study ("The Experiment" by BBC tv) produced quite different results, suggesting that personality has a significant influence on behaviour in this type of study.
Weakness 2: Zimbardo did not remain a detached observer in the case of this study; rather he himself assumed the role of "prison governor". He himself admitted that as a result of this, the experiment was not abandoned early enough, and their was a risk of psychological harm to the participants.