[simonh]

>If you always respond to this by doing whatever the stranger advocates, you're likely to end up getting scammed or getting eaten up by a political movement or otherwise doing something you later regret.

That's amazing, it's an excellent summary of the theory the article proposes for why we instinctively discount evidence against our preconceived opinions. Basically it's to stop us getting scammed, but in the hunter-gatherer context in which we evolved, not a modern society with robust systems for validating evidence.

The pre-conceived opinions the studies test aren't always even things the test subjects actually care about though. They can be opinions about things they were only just exposed to and wouldn't be expected to have any personal investment in, such as opinions about fictional characters that only exist in the test. We're not talking about proving to conservatives that liberalism is right, or vice versa, some of the tests are literally concerning beliefs about issues that only exist in the test. It doesn't matter, as soon as an opinion is formed it's incredibly hard to change it no matter how strong the counter evidence, and even if it's shown conclusively the initial opinion was based on false data.