Enlightenment Now: The Case for Reason, Science, Humanism, and Progress

Psychologists have long known that the human brain is infected with motivated reasoning (directing an argument toward a favored conclusion, rather than following it where it leads), biased evaluation (finding fault with evidence that disconfirms a favored position and giving a pass to evidence that supports it), and a My-Side bias (self-explanatory).²¹ In a classic experiment from 1954, the psychologists Al Hastorf and Hadley Cantril quizzed Dartmouth and Princeton students about a film of a recent bone-crushing, penalty-filled football game between the two schools, and found that each set of students saw more infractions by the other team.²²

We know today that political partisanship is like sports fandom: testosterone levels rise or fall on election night just as they do on Super Bowl Sunday.²³ And so it should not be surprising that political partisans—which include most of us—always see more infractions by the other team. In another classic study, the psychologists Charles Lord, Lee Ross, and Mark Lepper presented proponents and opponents of the death penalty with a pair of studies, one suggesting that capital punishment deterred homicide (murder rates went down the year after states adopted it), the other that it failed to do so (murder rates were higher in states that had capital punishment than in neighboring states that didn’t). The studies were fake but realistic, and the experimenters flipped the outcomes for half the participants just in case any of them found comparisons across time more convincing than comparisons across space or vice versa. The experimenters found that each group was momentarily swayed by the result they had just learned, but as soon as they had had a chance to read the details, they picked nits in whichever study was uncongenial to their starting position, saying things like “The evidence is meaningless without data about how the overall crime rate went up in those years,” or “There might be different circumstances between the two states even though they shared a border.” Thanks to this selective prosecution, the participants were more polarized after they had all been exposed to the same evidence than before: the antis were more anti, the pros more pro.²⁴

Engagement with politics is like sports fandom in another way: people seek and consume news to enhance the fan experience, not to make their opinions more accurate.²⁵ That explains another of Kahan’s findings: the better informed a person is about climate change, the more polarized his or her opinion.²⁶ Indeed, people needn’t even have a prior opinion to be polarized by the facts. When Kahan exposed people to a neutral, balanced presentation of the risks of nanotechnology (hardly a hot button on the cable news networks), they promptly split into factions that aligned with their views on nuclear power and genetically modified foods.²⁷

If these studies aren’t sobering enough, consider this one, described by one magazine as “The Most Depressing Discovery About the Brain, Ever.”²⁸ Kahan recruited a thousand Americans from all walks of life, assessed their politics and numeracy with standard questionnaires, and asked them to look at some data to evaluate the effectiveness of a new treatment for an ailment. The respondents were told that they had to pay close attention to the numbers, because the treatment was not expected to work a hundred percent of the time and might even make things worse, while sometimes the ailment got better on its own, without any treatment. The numbers had been jiggered so that one answer popped out (the treatment worked, because a larger number of treated people showed an improvement) but the other answer was correct (the treatment didn’t work, because a smaller proportion of the treated people showed an improvement). The knee-jerk answer could be overridden by a smidgen of mental math, namely eyeballing the ratios. In one version, the respondents were told that the ailment was a rash and the treatment was a skin cream. Here are the numbers they were shown:

Improved

Got Worse

Treatment

223

75

No Treatment

107

21