Motivated reasoning is the process by which we reach the conclusion we wanted and then construct the justification, while sincerely experiencing it as objective analysis. The uncomfortable finding from decades of research is that being smarter doesn't protect against it — it makes it worse, because greater analytical ability means a greater capacity to generate plausible-sounding reasons for what you already believed.
The tell is asymmetric scrutiny. When evidence supports what we want, we ask 'can I believe this?' and accept it readily. When it contradicts us, we ask 'must I believe this?' and hunt for flaws — and there are always flaws, in every study. Same person, same standards nominally, completely different thresholds. The only reliable defense isn't more intelligence; it's structural: seeking out disconfirming evidence deliberately, committing to what would change your mind before seeing the data, and exposing your reasoning to people who want to find its errors.
Dan Kahan and colleagues at Yale designed an experiment to test whether reasoning ability protects against bias. They gave participants a genuinely difficult data-interpretation problem — a two-by-two contingency table requiring careful analysis to get right. When the data was framed as being about the effectiveness of a skin cream, people with higher numeracy did much better, as you'd expect. Then they presented mathematically identical data framed as being about whether a gun control law reduced crime. Now the highest-numeracy participants were the most polarized: they got the answer right when the correct answer supported their political group's position, and got the same math wrong when it didn't. Their numerical skill wasn't being used to find the truth — it was being deployed to reach the conclusion their identity required. Kahan's conclusion was blunt: the problem is not a deficit of reasoning ability, and giving people more of it does not fix it.