David Navarrete and his colleagues from Michigan State wondered how people would react to moral dilemmas within a virtual world. They used the classic runaway train scenario for their test, but rather than presenting it as a hypothetical situation, they immersed their subjects in a three-dimensional virtual world. The change in venue made little difference.
The scenario runs as follows. An unstoppable train is heading down a track toward five people who have no chance to escape. There is a switch in the track that can divert the train onto a sidetrack just in time, but there’s one person on that sidetrack. You control the switch. Do you do nothing and let the train kill the five people, or do you throw the switch and let the train kill the single person?
If you’re like most people, you’ll throw the switch. In Navarrete’s virtual study, over 90% of people chose to sacrifice one person to save five, a percentage which is consistent with nonvirtual tests.
I have a problem with this study. Being in a virtual world, however realistic, does not change the fact that the subjects know that no one is really being killed. Thus, I don’t see how this is an improvement on simply asking people what they’d do. If anything, you might be testing how much participants want to see virtual people get hit by trains.
You can see a 2D adaptation of the test below.