We have run a series of studies that include two (ethical) paradigms of Milgram's obedience studies. The first, which takes place online, asks participants to assign increasingly negative adjectives to increasingly positive groups of people. Using this paradigm, we have used the following manipulations: 'Peers Rebel' feedback, identification with the experimenter by gender, identification with science by participants' background, identification with science by using the 'three things' manipulation, and identification with the 'learner' by reminding participants of shared priviledge. In the second paradigm, we use virtual reality to investigate helping behaviour and stress levels of the participants in a Milgram-like environment.Since the end of the second world war, western social thought has been haunted by the shadow of the holocaust. In 1961 two events came together which have shaped our understanding of the human capacity for inhumanity. The first was the trial of Adolf Eichmann in Jerusalem and Hannah Arendt's subsequent book about Eichmann. She argued that, far from being a monster, what was truly terrifying about him was his ordinariness. She also argued that what made him capable of murdering millions was 'sheer inattention': Eichmann concentrated so hard on the details of his job that he lost sight of the consequences. This was summarised in her phrase 'the banality of evil'. While the trial was in progress, Stanley Milgram was conducting his famous 'obedience' studies in Yale. Participants were asked to administer an escalating series of electric shocks to a learner each time he made an error in a learning experiment (actually the shocks were not real and the learner was a confederate of the experimenter). To his surprise, many participants went all the way to 450v. This seemed to provide experimental rigour to the notion that ordinary people can act in extraordinarily harmful ways. What is more, Milgram took on board Arendt's explanation of toxic obedience. In the face of authority, he argued, people enter an agentic state whereby they focus exclusively on how well they follow instructions. This notion of the perpetrator as an 'inattentive bureaucrat' has come to dominate understanding of atrocity in psychology, across different academic disciplines and even in the public domain. It is regularly invoked by the media and commentators every time we see instances of toxic behaviour. However, despite its dominance, the explanation is deeply flawed and cannot explain variations in levels of obedience, including within Milgram's own studies. Based on a reanalysis of these studiesand also on historical evidence, we have argued that perpetrators are 'engaged followers' rather than 'inattentive bureaucrats'. This has three elements: first, people identify with a cause. Second, the see the cause as noble. Third, they see the authority who tells them to inflict harm in the name of the cause as a legitimate representative. That is, people don't do wrong because they are unaware of what they do, but because they believe it to be right. In this project, we will provide support for the 'engaged follower' perspective. The work has three elements. The first of these involves collecting new empirical data by employing innovative methods which overcome the ethical objections to Milgram's original research. We will start with conventional experimentation which examines how far people will go in derogating ever more benign groups. This allows us to investigate the impact of our core variables (identification with the cause, glorification of the cause, legitimacy of the representative) in a relatively benign setting. Next, we will draw on findings which show that people behave similarly in virtual reality simulations to conventional psychology experiments but without the same level of emotional intensity or distress. We will develop a virtual reality simulation of the Milgram paradigm and use it to, again, address the impact of our core variables on obedience. Finally, we will plan a full field study working with the police/armed services and drawing on exercises that would be conducted as a matter of course. The second element will be to organise a series of seminars involving psychologists, historians and practitioners with a view to sharing latest understandings of obedience and atrocity and to create a critical mass of people who can challenge the authority of the 'banality of evil' perspective. The third element involves dissemination to the public, both through involving the media in the seminar series, through producing a film to illustrate the 'engaged follower' perspective and also through producing a popular book aimed at a general audience.

Data was primarily collected using undergraduate students who had agreed to be emailed psychological studies. In some cases, post-grad students and staff members working in universities across the U.K. were also emailed invitations to take part. All of the online studies used the same experimental paradigm (see Haslam, Reicher, & Birney, 2014 for a more detailed description). The task involves selecting negative traits to describe groups of increasingly pleasant people, making the task increasingly noxious for participants. In all of the studies, the purpose is to explore participants' willingness to continue with this unpleasant task as a function of their identificaiton with experimenter, the study's purpose, or the 'victim.' Hence, we did not collect data on the actual words chosen to describe each respective group. Instead, we recorded how far participants went in the trial (and variations of this measure) as well as their answers to the questionnaires that both preceeded and proceeded the task.