The evidence
What the studies found
Many scenes in the game adapt questions that researchers have put to real people. Each page here says what one of those studies asked, what it found and where it was published. Every number is the study’s own.
A share of people is not a verdict. These pages report what people answered, not what is right.
The trolley problem: what people actually answerTurn the wagon and one dies instead of five. Most people say yes. Push a man to stop it, and most say no.The framing effect: lives saved or lives lostThe same two plans, described once by the lives they save and once by the deaths they cause.The identifiable victim effect: one voice against manyOne named person moves us more than a number, however large the number.In-group favouritism: your people firstPeople favour their own group even when the group was made up a minute ago.The ultimatum game: would you refuse an unfair offer?Take a small share, or refuse and both get nothing. Many people refuse.The dictator game: what people give when nobody can refuseYou split the money and the other side must take it. Most people still give something.Third-party punishment: paying to punish a strangerWould you pay to punish someone who cheated a person you will never meet?Omission bias: the harm you cause and the harm you allowA medicine that kills fewer children than the disease it prevents, and why many still refuse it.Obedience to authority: the Milgram experimentsHow far people went on hurting someone because a person in charge told them to.Conformity: when everyone else gives the wrong answerThe Asch line experiments: how often people went along with a group that was plainly wrong.Anchoring: the number you saw firstA number from a spun wheel moves the estimates of people who know it means nothing.Base-rate neglect: what a positive test really meansA test with a 5% error rate comes back positive. The chance it is right can be about 2%.The sunk cost effect: finishing what should be stoppedSpending more because of what has already been spent.The default effect: when doing nothing is the choiceAsked to opt in, 42% agreed to be organ donors. Made donors unless they opted out, 82% stayed donors.