Experiment in using TV as a science lab for research

Harriet Swain considers the ethics of the increasing involvement of scholars in fly-on-the-wall TV programmes, while Alex Haslam and Steve Reicher argue that any risks are far outweighed by the potential benefits.

Last December, we ran a major study into the psychology of social inequality. Over ten days, we examined the behaviour of 15 men who were randomly assigned to the role of prisoner and guard in a purpose-built prison. In the next fortnight, aspects of the study will be broadcast on BBC2 as The Experiment .

This has been a groundbreaking collaboration between academia and the media, with science as the guiding principle for us and the BBC. As academics, we designed, ran and analysed the study. The BBC provided intellectual and emotional support, technology and funding at a level that we could never have hoped to obtain from research councils.

We believe that the result will be compelling television and an important contribution to academic and public debate. But it was apparent all along that this outcome could be achieved (and justified) only if we were able to steer our way through some important and, at times, controversial issues.

The first revolved around the ethics of the science. It arose primarily from the link between our experiment and Philip Zimbardo's Stanford Prison Study. In this, the random allocation of college students to the role of prisoner or guard in a simulated prison seemingly led to brutality from the former and psychopathology in the latter. For the researchers, this was evidence of the capacity of powerful roles to render irrelevant the concepts of human judgement, will and decency with which we seek to understand and explain behaviour.

So shocking were its consequences that the study has not been replicated. Zimbardo has argued that it never should be. The result has been that, although they feature prominently in textbooks and the wider culture, his conclusions have remained inviolate. We believe that those conclusions are wrong: intellectually and morally. They are wrong in the sense that the Stanford guards did not immediately become tyrants, they did not all become tyrants and to the extent that any did, they were encouraged to do so. This is apparent in the briefing of the guards when they were instructed by Zimbardo (who acted as governor) to "create in the prisoners feelings of boredom, a sense of fear to some degree... a notion of arbitrariness that their life is totally controlled by us, by the system, you, me". For us, such instructions are incompatible with the argument that "guard aggression was emitted simply as a 'natural' consequence of being in the uniform of a 'guard' and asserting the power inherent in that role".

But as well as being intellectually contestable, such conclusions are morally problematic in that they allow tyrants to avoid responsibility for their acts. They legitimate the defence that "I couldn't help it, it was the uniform that made me do it", a message that suggests the struggle against tyranny is futile. If a model serves to perpetuate injustice, we have not only an intellectual but also a moral and political responsibility to challenge it. But this seems to leave us with a classic means-end dilemma. How can we do a study to challenge our understanding of tyranny without tyrannising research participants?

We spent almost nine months planning our experiment with the BBC, gaining ethical approval and seeking professional input. Our solution was to create a situation that was harsh but not harmful, in which the prisoners'

deprivations were more symbolic than material. For example, their meals were of inferior quality to those of the guards but were nutritionally adequate.

Next we had to devise a series of procedures to monitor the ethical boundaries and ensure that no breaches were possible. This involved: three-phase screening of participants by us, the BBC and independent clinical psychologists to ensure that they were not a risk to themselves or to others; constant monitoring of the participants by security and medical personnel; constant monitoring of the participants by clinical psychologists during the study, and provision of support after the study and its broadcast; and constant monitoring of the study and proactive screening of planned interventions by an independent five-person ethics panel. The clinical psychologists and the panel had the power to end the study at any stage. The participants were free to leave at any time. But neither the overseers nor the participants had serious concerns about the way the study was run or about the participants' welfare. The panel concluded that the study provided an exemplary instance of how to address powerful phenomena in an ethical manner. Although we finished the study slightly early (we had gone as far as we could with the science), they said they would have been happy for us to continue.

The televising of this study brought with it new layers of ethics to consider. Two thorny issues were those of informed consent and anonymity. We could warn the participants of the consequences of being talked about in the press, but how could we prepare them for a potential media scrum? We could warn them about being watched and recognised by millions, but how could we prepare them for the fact that they would have to account for actions that were performed in one context but judged in another?

We sought to address these problems by involving the participants in the process of analysis. They twice collectively viewed rough cuts of the programmes, and although they were not given editorial control, their observations informed our reading and interpretation of incidents. In this way, their perspective served to enrich ours and made the science and the programmes better. What is more, the whole point of the research was to understand action in context. If people behave in particular ways, the aim is to understand the antecedents and the consequences of their acts rather than to individualise and judge the participants.

Televising the experiment raised other issues that we needed to address. How would the presence of the cameras affect the behaviour of the participants? There is no doubt that our participants were aware that they were being observed and that this affected their behaviour. But to dismiss our study because of this would be to dismiss all scientific study in which participants know they are being observed - in other words, almost all psychological science including the Stanford study. It is worth noting that Zimbardo has called his study "one of the first examples of reality TV" in that the whole procedure was videotaped. The fact that people are under surveillance in complex ways involving multiple audiences, some at a long remove from the immediate context, is far from artificial. It is an increasing feature of contemporary life. The problem lies not in the fact of surveillance, but in elaborating models of action that ignore its impact.

Our participants were sometimes aware of the cameras and the wider television audience, but more often they were concerned with the experimenters watching them. And they were also aware of their fellow group members and of the members of the rival group. Television surveillance was therefore one factor among many. But being a constant, it cannot account for variations across time and in responses to our planned experimental interventions. It also cannot account for the strong effects we obtained on covert psychological and physiological measures. And to the extent that it does explain something, it enriches rather than undermines the scientific and social value of the study. Indeed, if it was the case that our guards failed to display brutality because they were being filmed, this is an important lesson in itself - for it goes against the view that brutality is an inevitable consequence of having a powerful role and points to the role accountability can play in curbing tyranny.

Critical voices could still argue that this is all very well, but that all the media wants is drama that will produce high viewing figures. We found no evidence of this. If we found anything distasteful, it was a tendency towards moral smugness among some academics who represented themselves as fearless warriors of truth and the BBC as the "dumbing-down" enemy. By contrast, the producers and executives we worked with were committed to the science and concerned with the issues. They invested massive amounts of time and energy in addressing ethical issues and communicating our scientific message. What we have ended up with is not an academic paper, but we do hope it might lead people to understand why we write such papers and the nature of the issues we try to address when we do.

We would freely admit that even if The Experiment is successful, it runs the risk of spawning imitations that try to cut corners and that compromise scientific standards.

The dangers are there, but surely the greatest danger would be to never take risks or rise to challenges that one sees as difficult but important. In academia, as everywhere, the only way to be certain of being completely irreproachable is to be completely irrelevant.

Alex Haslam is professor of psychology, University of Exeter, and editor of the European Journal of Social Psychology . Steve Reicher is reader in psychology, University of St Andrews, and co-editor of the British Journal of Social Psychology . The Experiment begins on BBC2 on May 14.