One of the most widely-held ideas in the popular consciousness is that a person’s access to correct facts determines their espousal of correct beliefs. In other words, we think we base what we think on the facts that are available to us. This is not true. We determine what we accept as fact based on what we already believe to be true.

This is as true for the scientist as it is for the priest. We all have ideological pre-commitments that determine what facts we are willing to accept. And this truth is never more firmly established than when we are confronted with “opposing” facts.

Recently, Snopes, Politifact, and other fact-checking sites have popped up to tell us whether some or other political or social story is true or not. You would think such reliable access to the “facts” would result in a reduction of the belief in myths. It hasn’t. In an interesting study, it turns out that fact-checking services can produce a “backfire” of even stronger belief:

In the experiment, participants were shown a fake newspaper article containing an actual George W. Bush quotation: “The tax relief stimulated economic vitality and growth and it has helped increase revenues to the Treasury.” In one version of the experiment, the article then contained a correction, refuting this claim; in another version it did not. It turned out that conservatives who read the correction believed Bush’s falsehood more strongly than did conservatives who never read the correction.

Similar studies have been done with Democrats. It works both ways. In our day, we implicitly believe that fact-checking is not trustworthy. If someone works very hard to dispel something, it’s probably because it’s true, right? It’s the epistemological equivalent of “the lady doth protest too much.” Our natural inclination is to decide what we believe is true, receive data that reinforces that belief, and reject oppposing data as “irrationally motivated” or “manipulated.”

So fact-checking often has the opposite effect than what its proponents intend. Perhaps fact-checking services just want to give people better, clearer information. But they have misjudged the process by which people come to conclusions. In most cases, people believe before they understand. And no amount of premise-tweaking will ever affect their conclusions.