Researchers agree this effect is real. They do not agree on how strong it is, and that argument is still going on. So you can trust the direction, but not any exact number.
Peter Thiel opens Zero to One with a question he says he puts to almost anyone he is evaluating. What important truth do very few people agree with you on? It sounds easy until you try to answer it out loud. Thiel reports that good answers are rare, and that the failures follow a pattern. Most people reach for something safe wearing a bold costume. Our education system is broken. Technology is changing everything. Statements like these feel daring, and yet everyone nods along, and that universal agreement is precisely what disqualifies them. A real answer demands something professional life spends years training out of people: a belief you reasoned your way into, that the room will dislike, spoken aloud to someone holding power over your future. Rehearsal barely helps, because a prepared contrarian answer collapses under a single follow-up. Two things get tested at once, whether genuinely independent thinking has ever happened in this person's head, and whether they carry the nerve to voice it when the voicing costs something.
Notice what most interviews actually reward. The candidate who reflects the company's own beliefs back with confidence tends to do well, which means many hiring processes are calibrated to filter out exactly the people this question was built to find. That observation is the useful part, and it stands regardless of what anyone thinks of Thiel himself. Before adopting the question, be honest about what it is: a good prompt for thinking, not a validated predictor. Nobody has shown that answering it well predicts performance in a role, and the claim that good answers are rare is one man's impression across his own interviews rather than a measurement. So use it to test your process rather than your candidate. If everyone who does well in your interviews agrees with you, that is worth knowing, and it is a real and common failure. Scoring people on their answer to a question nobody has validated is a different thing, and a riskier one.
Source: Peter Thiel with Blake Masters, Zero to One, 2014. The claim that good answers are rare is Thiel's own observation across his own interviews, not a measured finding, and the question has never been validated as a selection tool.
Zero to One
Peter Thiel with Blake Masters, 2014
Where the question appears, alongside the argument for why consensus and value usually sit in different places.
Draw your own card. It does not take long, and it rewards taking your time.