After conducting many studies like this, Kahan and his team concluded that if individuals are members of groups who have become polarized about a particular issue, and that polarization puts the group’s opinions at odds with scientific consensus, people will almost always go with what their group believes over what the preponderance of the evidence suggests. In another of Kahan’s studies, people were told they would be making sense of the raw numerical results of a separate bit of research that tested the effectiveness of skin cream. Now, the results were fake, and for half the subjects the cream was shown to be effective, and for the other half it was shown that it wasn’t. Kahan found that the better subjects were math, no matter their politics, the better they performed when it came to determining the effectiveness of the cream. But when those exact same numerical results were relabeled, and subjects were told the research tested the effectiveness of gun control, the better subjects were at math the worse they performed — but only if the political party they belonged to was openly opposed to what the numbers suggested. If the results suggested that gun control was effective, Republicans who were good at math became bad at math. If the results showed gun control was ineffective, Democrats who were good at math became bad at math. If their party favored the results, then once again math skills alone determined the subjects’ performance, the same as it had when the exact same results supposedly measured the effectiveness of skin cream. Kahan says that the better you are with numbers, the better you are manipulating them to protect your identity-connected, and in this case politically motivated, beliefs. Of course in the study none of the subjects had any idea they were doing this. They didn’t think their tribal loyalty was affecting their math ability. They all felt they were doing their best.