PLoS ONE (Jan 2018)

Participant concerns for the Learner in a Virtual Reality replication of the Milgram obedience study.

  • Mar Gonzalez-Franco,
  • Mel Slater,
  • Megan E Birney,
  • David Swapp,
  • S Alexander Haslam,
  • Stephen D Reicher

DOI
https://doi.org/10.1371/journal.pone.0209704
Journal volume & issue
Vol. 13, no. 12
p. e0209704

Abstract

Read online

In Milgram's seminal obedience studies, participants' behaviour has traditionally been explained as a demonstration of people's tendency to enter into an 'agentic state' when in the presence of an authority figure: they attend only to the demands of that authority and are insensitive to the plight of their victims. There have been many criticisms of this view, but most rely on either indirect or anecdotal evidence. In this study, participants (n = 40) are taken through a Virtual Reality simulation of the Milgram paradigm. Compared to control participants (n = 20) who are not taken through the simulation, those in the experimental conditions are found to attempt to help the Learner more by putting greater emphasis on the correct word over the incorrect words. We also manipulate the extent to which participants identify with the science of the study and show that high identifiers both give more help, are less stressed, and are more hesitant to press the shock button than low identifiers. We conclude that these findings constitute a refutation of the 'agentic state' approach to obedience. Instead, we discuss implications for the alternative approaches such as 'engaged followership' which suggests that obedience is a function of relative identification with the science and with the victim in the study. Finally, we discuss the value of Virtual Reality as a technique for investigating hard-to-study psychological phenomena.