Re-evaluating the SPE. And its Critics.

It’s probably fair to say that Philip Zimbardo’s Stanford Prison Experiment (SPE) has, particularly over the last few years, attracted a great deal of critical attention – something that should, on the face of things, make it relatively easy for students to evaluate.

The problem, however, is that while the SPE has thrown-up a number of well-known criticisms these don’t always seem to be presented in a way that’s been clearly thought-through in many contemporary psychology textbooks and websites. While students are presented with criticisms of the study, there’s no real attempt to evaluate them in the light of our changing understanding of the study – something this post is designed to correct.

The Critique…

In general students are encouraged to criticise the SPE in terms of three ideas:

• the study is invalidated by demand characteristics.

• its reliability has been called into question by its lack of replication, particularly in relation to Haslem and Reicher’s (2002) BBC Prison Study.

• the study involved a number of highly-questionable ethical choices.

We can examine each of the above to, firstly, understand the basis of the criticism and, secondly, suggest a “critical response” students can make as a means of actually evaluating the SPE and the claims of its detractors.

Demand Characteristics

The most potentially damning criticism raised against the SPE is that Zimbardo and his collaborating colleagues weren’t simply dispassionately observing a situation play-out in real time. Rather, the argument here is that Zimbardo in particular was heavily-invested in ensuring that the outcome of the experiment matched his predictions about what would happen.

In basic terms, Zimbardo’s situational psychology argued that people are shaped by the environments in which they find themselves. Thus, if a group of perfectly ordinary, non-violent, college students are given positions of complete power and authority, where they are expected to keep order by whatever means necessary, they will adapt to the situation by behaving in ways that demonstrate that power.

In this respect, critics of the experiment argue that Zimbardo acted in ways – some subtle, some not – that biased the experiment and effectively created the outcomes he predicted. As has emerged, for example, from the Stanford Prison archive in films like Quiet Rage, Zimbardo was heavily-involved in directing the experiment – from designing the prison (including its punishment cell…) and intimidating uniforms to briefing the Guards about their role and the importance of maintaining order. The fact Zimbardo took-on a dual role as both the lead researcher and the Superintendent in charge of overseeing everything that happened in the prison has also led to the accusation that this gave him the power, which he enthusiastically exercised, to determine the outcome of the experiment.

However…

While it’s clear demand characteristics played a part in the Stanford Prison Experiment, the question is whether or not this, in itself, is a sufficient criticism? The real evaluation here, perhaps, is the question of whether demand characteristics were an integral part of the experiment itself?

The argument here is that Zimbardo needed to embed clear demand cues into the experiment in order to enhance its ecological validity (an idea we’ll discuss further in a moment in relation to realism and replication). The important point here is that in order to test the brutalising effects of prison conditions on both guards and prisoners, Zimbardo had to create an initial environment, both physical in terms of the building and mental in terms of the guards’ orientation to their role, that at least partially reflected the real structure of American prisons.

In other words, for the experiment to work as designed it was absolutely necessary to create an environment that encouraged, in some small ways, the guards to develop authoritarian personalities. That, after all, was the point. To argue Zimbardo’s demand cues biased the experiment by compromising his objectivity isn’t sustained by the evidence: once Zimbardo set the parameters for the experiment – how the prison was set-up, the punishments available to the guards, how the guards should orientate themselves towards their role – it was allowed to play-out until things got out of hand…

The key point here is that in a real prison system the individual guard or prisoner is dropped into an already existing power structure they cannot control: they can either conform to its demands or oppose it and face the consequences. And since this is what the Stanford Prison Experiment tried to simulate, the demand characteristics introduced by Zimbardo, it could be argued, were an important and necessary part of the experiment.

Ethics

A second contemporary source of critical discussion around the SPE involves the ethical standards applied in the experiment. Judged by contemporary standards there’s little argument the SPE would have spectacularly failed to uphold the current ethical standards relating to such things as:

• Freedom from psychological and physical harm
• The right to withdraw at any time
• Debriefing following the conclusion of the study.

However…

There are a number of critical considerations we can throw into the mix:

Firstly, at the time of the experiment the American Psychological Association’s ethical guidelines were much less prescriptive than contemporary ethical codes. They were both more aspirational – standards a researcher should strive to achieve – and largely left to the professional judgement of the researcher. A much stricter set of behavioural rules wasn’t introduced until 1973 – two years after the conclusion of the SPE. While there’s little doubt that the kinds of behaviour displayed by the Stanford researchers – from prolonged psychological distress experienced by both some guards and prisoners to pressure being placed on participants not to withdraw – would not be tolerated today, this wasn’t necessarily the case in 1971.

The critical question here, therefore, is whether or not it’s valid to hold the Stanford researchers to the ethical standards pertaining at the time of the experiment or to judge their behaviour in the light of contemporary standards?

Secondly, this evaluation its given greater clarity if we understand how the SPE was actually judged by the ethical standards in place at the time. Zimbardo, for example, was given permission to carry-out the study by the relevant Stanford University authorities and, more-significantly he requested a post-experiment review from the American Psychological Association that concluded that, under the 1963 guidelines operating at the time of the experiment:

• informed consent had been gained from the participants
• no explicit use of physical violence, banned under the code, was in evidence.
• there was no material breach of the existing guidelines.

The question here, therefore, is whether we judge a study by the ethical standards of the time or by contemporary standards – a question that, in itself, is clearly up-for-both discussion and further evaluation…

A third critical consideration raised by the SPE relates more-generally to the purpose and nature of psychological research and the question of whether some kinds of social research are effectively off-limits to researchers because they run the risk of breaking current ethical standards?

While this isn’t a question we can resolve here, it raises further evaluative questions that students can reasonably ask concerning both the subject of psychological studies and the ethical codes that surround such research. Would contemporary researchers, for example, be allowed to replicate the conditions under which something like the SPE took place or would they be effectively banned from producing such research because of current ethical guidelines?

Reliability and Replication

A common criticism of the SPE, particularly at A-level, is its lack of reliability. While the fact no-one has been able to successfully replicate the study in the 50 years since its inception might be significant, we need to keep in mind no-one has actually tried to replicate the full study over this time. The closest we get is probably Haslem and Reicher’s BBC Prison Study (2002) and the argument it not only failed to replicate the SPE, but that it also exposed some of its fundamental flaws – particularly the idea it was fatally compromised by Zimbardo’s demand cues.

Le Texier, for example, argues the fact Haslem and Reicher chose not to put themselves into the narrative by taking on leadership / superintendent roles or provide guidance to the guards resulted in their very different findings – the guards in the BBC Study, for example, did not develop brutal attitudes towards the prisoners in their charge. This, he claims, supports his argument “that Zimbardo’s guidance and demand characteristics likely played a major role in the outcome of the SPE”.

However…

While the failure of the BBC Study to replicate the SPE is frequently cited as evidence that the conclusions drawn by Zimbardo from his research are invalid, a closer inspection of the BBC Study leads to both a different conclusion and a more-nuanced interpretation of the relationship between the two studies.

One criticism of the “replication failure” argument is that the two studies were only superficially-similar. When they commissioned the study the BBC may have thought they were getting a replication of the SPE, Hasler and Reicher seem to have had a slightly different agenda, one that involved applying their ideas about social identity theory to an understanding of the social dynamics between two different, but carefully matched, groups in a conveniently confined space (a pretend prison).

In this respect, although the setting was similar, the objectives of each study were different:

• For Zimbardo, the SPE reflected his life-long attachment to situational psychology: the idea, in a nutshell, that social environments condition and determine individual behaviours.

• Haslem and Reicher, on the other hand, were much more interested in using the opportunity to test the idea that people do not blindly conform to social roles, as Zimbardo effectively claimed. Rather, the extent to which individuals identify with social groups, such as “Guards” or “Prisoners” determines whether they conform to. or resist, authority. In other words, social identity theory argues people don’t just obey because an authority tells them to: they obey when they identify with what they believe to be the authority’s legitimate aims and goals.

The sampling process used in each study to allocate guard and prisoner roles reflected the different objectives of the two studies. The SPE used a simple random selection of guards and prisoners, while the BBC study used a more-sophisticated stratified random sample to assign guard and prisoner roles to the 15 “normal, decent, well-adjusted individuals” selected for the experiment. This simple difference in how participants were allocated to different roles had one very important effect: it conditioned how the participants in each study were encouraged to perceive their roles.

In the SPE the guards were immediately treated very differently to the prisoners and, most-crucially, were aware, because Zimbardo and his colleagues made it very clear, of their position in a prison power structure: Zimbardo, in short, had their back. In the BBC study the guards, in the minority, were largely left to their own devices by researchers who declined to intervene and were powerless to resist when the prisoners took over…

The argument here, therefore, is that aside from the setting, the SPE and BBC Prison Study have little in common. We’re not comparing like-with-like and so any attempt to see the one either replicating or failing to replicate the other is like criticising a banana for not being an apple. The mistake, in this respect, is assuming that because they were experiments set in the same situation they were designed to measure the same thing.

One of the key things about the SPE is that it reflected Zimbardo’s life-long attachment to situational psychology. Unlike the SPE, in the BBC Study there was no attempt by the authors to “brutalise” the incarceration process. In some instances, such as when they introduced a trained Trade Union negotiator as a new prisoner, the reverse was true: there seemed to be deliberate attempts to defuse any sources of tension.

In the BBC Study the Guards were disunited and unsure about their ability to exercise power – they were, of course, aware that, if push-came-to-shove they had very little real power or authority in this situation – unlike in a real prison system where the power of individual guards is backed by a wide range of structural sanctions that can be brought to bear on prisoners.

Realism (Ecological Validity)

As we’ve seen, one of the key criticisms of the SPE, articulated by critics such as Le Texier (2019), is its lack of realism (or ecological validity). This, in particular, was directed towards the “prison” itself which, as Zimbardo demonstrates in Quiet Rage, was a basement corridor hastily shaped into makeshift prison cells. The infamous “hole” into which miscreant prisoners were thrown for bad behaviour was actually an old broom cupboard. Although some aspects of the experience were real – the “arrests” of the prisoners, their transportation to the prison and various mortification processes such as blindfolding and fingerprinting – others were plainly not: each prisoner was given a nylon head cap to wear in lieu of having their heads shaved…

This lack of “realism” does, of course, extend to the prison experience itself insofar as all of the participants were being asked to play roles. Playing the part of a prisoner for money (each was paid around $115 per day, adjusted for inflation) knowing it’s only for a couple of weeks (the experiment actually only lasted 6 days) is clearly very different to being banged-up in a maximum-security prison for months or even years.

However…

Two main evaluative questions arise from this, the first being that, of course, it wasn’t real. While the SPE crossed many ethical lines, we have to assume that Zimbardo didn’t seriously consider kidnapping innocent students and subjecting them to a rigorous prison regime against their will. As daft as this may sound, it does raise an important point about social psychological research in that if we want to study areas that potentially involve tricky ethical issues there has to exist an element of non-realism. Either we accept the limitations of this situation or we don’t engage in this kind of research.

The critical thing to note here is that ethical considerations are, by their very nature, bound-up in cultural conventions – something we’ve seen earlier with the observation that behaviour considered generally acceptable in the 1970s is no-longer considered acceptable in the early 21st century. The point here is that while arguments about ecological validity are pertinent in many contexts, this is arguably not one of them because of a second, highly-significant, evaluative question, namely, what was the SPE actually about? What, in other words, was it testing?

Critical evaluations of the SPE generally accept Zimbardo’s interpretation of the experiment and the conclusions he draws from it: working within a fairly tight situational framework, the objective was to show how, in basic terms, brutal environments produce brutal and brutalised individuals (a theme to which he returned in 2007’s “The Lucifer Effect”). This, of course, is where problems of realism and ecological validity raise their head precisely because everyone involved in the experiment know it “wasn’t real” – and if everyone involved in the prison interaction knew this then how can any conclusions drawn from the respective behaviours of the guards and prisoners be valid?

As Haslem and Reicher succinctly put it “people do not automatically assume roles that are given to them in the manner suggested by the role account that is typically used to explain events in the SPE”.

Social Identity Theory

And this brings us back to the BBC Prison Study and Haslem and Reicher’s application of social identity theory. If, rather than accepting Zimbardo’s interpretation of the experiment and criticising his argument accordingly, we apply Haslem and Reicher’s interpretation of both studies, the problem of ecological validity largely disappears. This follows because social identity explanations describe relationships and social dynamics that are tied to social groups, such as “guards” and “prisoners” (or, as in these contexts, people playing these roles) not the physical location (a pretend prison) of those groups.

The argument here is that social identity explanations of social interaction apply as equally to so-called real-world situations as they would to contrived situations such as field or lab experiments. What the researchers are studying in these contexts is the process by which social identities are formed, develop, change and generally play-out – and these processes are not necessarily bound by artificial constructions  of “realism”.

In this respect, criticism of Zimbardo and the SPE is not that the experiment lacked ecological validity, it’s that he employed the wrong theory to interpret the data collected during the experiment. A simple example here is the oft-stated observation that although all of the guards were placed in the same stressful position, not all of them responded in the way Zimbardo predicted. The fact some guards were brutal but others weren’t, even in a situation where “brutality” was both encouraged and tacitly endorsed, suggests we need a much more subtle interpretation of the situation, one that arguably Hasler and Reicher provide.

Rather than failing to replicate the SPE, therefore, the BBC Prison Study interpreted in the light of social identity theory might actually be used to confirm its findings…

References

Thibault Le Texier (2019) “Debunking the Stanford Prison Experiment”

Alex Haslem and Steven Reicher (2002) The BBC Prison Study