Source: "Free exchange: How to be wrong"
Human thought is not always (or even usually) rational. Humans don’t use new information to update their beliefs - instead, they cling to their beliefs in the face of contradictory evidence. Jean Tirole and Roland Benabou wrote a landmark paper in 2016 that looks at beliefs as consumption goods. This new perspective yields interesting insights and predictions.
Beliefs are goods, in the sense that people spend resources acquiring them and developing them and they derive value and benefits from their beliefs. A typical value involves signaling membership in a group. A less common belief value would be the way believing in a religion can shape your behavior (if I believe I am a good salesman, I can use the confidence generated by my belief to close more sales; if I accept an ascetic religion, my beliefs can help me avoid unhealthy behaviors).
Benabou and Tirole believe people engage in motivated reasoning to protect their hard-won beliefs from contradictory evidence. The first stage of motivated reasoning is strategic ignorance: here, the believer simply ignores the evidence. In stage two, reality denial, the believer acknowledges the new evidence but refuses to accept it as coming from a credible source. Finally, in self-signalling, the evidence is accepted as credible but the believer interprets it as actually strengthening their belief, not contradicting it (example: an unhealthy person might decide that their ability to still run a few miles is proof that they are indeed healthy).
Other work by Benabou suggests that “groupthink” is highest when people within a group share the same fate if their briefs are discredited. If a politicians fortunes are linked to his party’s performance, they will have little incentive to speak out against the group. Such groups become partisan and polarized. Such groups try to delegitimize independent voices (like research groups or watchdog groups).