What Is Learned from 70,000 Responses to Trolley Scenarios?


A team of researchers has reported on its collection and analysis of 70,000 responses to three scenarios that frequently comprise versions of the trolley problem.

Appearing in Proceedings of the National Academy of Sciences of the United States of America, “Universals and variations in moral decisions made in 42 countries by 70,000 participants” lays out the results of the massive study that collected the responses, in ten languages, from 42 countries. The study was conducted by Edmond Awad (Exeter, MIT),  Sohan Dsouza (MIT)Azim Shariff (UBC)Iyad Rahwan (MIT, Max Planck Institute), and Jean-François Bonnefon (MIT, Toulouse).

The participants were confronted with three hypothetical scenarios in which a trolley will soon roll over and kill five workers unless the participant chooses to intervene and sacrifice the life of one other worker.

In “Switch” (aka “Bystander”) the participant’s options are to do nothing, in which case the five workers are killed by the trolley, or to pull a lever that diverts the trolley onto a spur track on which one worker is stuck, killing that one worker. In “Footbridge,” the participant’s options are to do nothing, in which case the five workers get hit by the trolley and die, or push a large person off of a bridge above the track so that his body stops the trolley, killing him but saving the others. In Loop, the participant’s options are to do nothing, in which case the five workers are killed by the trolley, or to pull a lever that diverts the trolley onto a side track that connects back up to the main line just before the location of the workers; ordinarily, this would do nothing to save them, but in this scenario there is a large person on that side track, and if the trolley hits him, it will stop before it gets back to the main line, killing the one on that track but saving the five on the main line.

What did the researchers find? First, “Participants endorsed sacrifice more for Switch (country-level average: 81%) than for Loop (country-level average: 72%), and for Loop more than for Footbridge (country-level average: 51%).”  The authors say that “data suggested that people from different cultures displayed a remarkable qualitative regularity, the Switch–Loop–Footbridge ordered pattern of preferences.”

Second, the data “also suggested that people from different cultures displayed quantitative variations in the exact degree to which they endorsed sacrifice in each of these scenarios.” This finding concerned relational mobility. Relational mobility refers to “how much freedom and opportunity a society or social context affords individuals to choose and dispose of interpersonal relationships based on personal preference. The researchers found a positive correlation “between relational mobility and the propensity to endorse sacrifice in each scenario variant.” Commenting on this finding, the authors write:

holding attitudes that put one at social risk is especially costly in low relational mobility societies, where alienating one’s current social partners is harder to recover from. This cost is likely lower (although not absent) in high relational mobility societies, as they offer abundant options to find new, like-minded partners. Accordingly, people in low relational mobility societies may be less likely to express and even hold attitudes that send a negative social signal. Endorsing sacrifice in the trolley problem is just such an attitude. Recent research has shown that people who endorse sacrifice in the trolley problem are perceived as less trustworthy, and less likely to be chosen as social partners. As a consequence, low relational mobility societies may feature more acute pressure against holding this unpopular opinion. Although it is possible that this pressure would discourage people who hold socially risky positions from expressing them, it could also change people’s attitudes, making certain ideas morally “unthinkable.”

The researchers have made their data publicly available on the Open Science Framework for others to examine and use.

Here’s one question for us: What, if anything, should philosophers take away from this study?


Related: “Trolley Problems: You’re Doing It All Wrong“, “On Trolley Problems

There are 23 comments

Your email address will not be published. Required fields are marked *

  
Please enter an e-mail address