In the last three posts I've considered the psychology behind whether people accept evolution or instead opt for an alternative like Intelligent Design (see Darwin Day Posts 1, 2, and 3). In this fourth and final Darwin Day Post, I turn to a complex question that runs throughout this series:
Does research concerning the psychological origins of a particular belief have implications for whether or not the belief accurately describes the world? In other words, if your mind made you believe it, is it true?
Several examples emerge in the context of evolution and Intelligent Design. For example, does learning that people are more likely to endorse Intelligent Design when they've been primed to feel powerless (the subject of Darwin Day Post #1) or if they misunderstand the nature of scientific theories (Post #3) have implications for the truth of Intelligent Design itself?
Judging a claim on the basis of its origins rather than its merits is considered a fallacy under some conditions, and is known as the "genetic" fallacy. Nonetheless, psychological explanations for belief are often presented as a way to undermine a belief's legitimacy. Consider the following cases:
In these cases, the psychological findings seem to provide good reason to doubt Andrea's and Bob's beliefs. But to investigate whether people actually find this line of argument compelling, I ran an informal pilot study. I found that 16 out of 21 people (76%) said that the psychological results made them less inclined to think Andrea's belief was true, with a comparable 19 out of 21 (90%) for Bob. In both cases, people tended towards "yes." 
There are two intuitive ways to understand why psychological explanations like this seem to challenge the truth of beliefs. Here's the first, which I'll call the "statistical argument." Assume that prior to learning about the psychological research, you believed that there was some low probability that Andrea and/or Bob was mistaken - say 5%. After learning about the psychological research, however, that number should go up. If 75% of people believe they are better-than-average drivers, for example, then at least 25% of people must be wrong. And if a majority of people overestimate their contributions to housework, than over 50% of people must be wrong.
Here's a second way to understand what's going on, which I'll call the "explanation argument." Prior to learning about the psychological research, you had a compelling explanation for why Andrea and Bob had their respective beliefs: Andrea believed that she was a better-than-average driver because she was a better-than-average driver, and Bob believed that he did 60% or more of the housework because he did 60% or more of the housework. But once the psychological findings are revealed, we have an alternative explanation for each belief. Andrea believes that she is a better-than-average driver because most people do as a result of particular psychological mechanisms—say a desire to think well of themselves. And Bob believes that he does 60% or more of the housework due to particular psychological mechanisms—say a tendency to effectively notice and recall one's own contributions to housework, but to fail to notice or recall those of one's partner. If these beliefs aren't explained by their truth, then the beliefs themselves no longer provide good evidence for their truth.
Nonetheless, neither of these arguments decisively shows that Andrea or Bob is wrong, which is why the genetic fallacy is a fallacy. If we wanted to know the truth with greater certainty, we'd need to turn to independent evidence: a driving test for Andrea, for example, and a daily observation log for Bob. (This evidence is "independent" in the sense that the outcome shouldn't depend on the accuracy of the beliefs in question. )
Now consider some trickier cases:
These cases aren't quite like (A) and (B). In those cases, the psychological research itself implied that some proportion of people's beliefs were false, supporting the statistical argument. In (C) and (D), the psychological research doesn't have this implication, so the statistical argument doesn't hold.
But the explanation argument still holds. If psychological mechanisms explain someone's acceptance or rejection of evolutionary theory, then we have an alternative explanation for their belief—one that doesn't hinge on the truth or falsity of the belief itself. This doesn't logically imply that the belief is false, but it arguably provides a reason for doubt, shifting the burden of proof.
For these cases, my informal study revealed a mixed pattern of responses: 14 out of 22 people (64%) indicated that the psychological results challenged the legitimacy of Catherine's belief, but only 9 of 21 (43%) said that the psychological results made Doug's belief less likely to be true.
Note that what I've said so far cuts both ways. A psychological explanation for the rejection of evolution in favor of, say, Intelligent Design might challenge the legitimacy of Intelligent Design, but a psychological explanation for why someone accepts evolution should be similarly damning. The asymmetry between scientific and supernatural beliefs emerges not from the applicability of the explanation argument itself, but from the sources and availability of independent evidence concerning the accuracy of beliefs.
Evolutionary theory is bolstered by an enormous body of empirical research. And from a scientific perspective, Intelligent Design and creationism are unfalsifiable at best and refuted by current evidence at worst. But for someone who endorses a form of creationism, holy texts, the testimony of others, and personal experience might be regarded as legitimate sources of evidence that provide a compelling explanation for belief, whatever the psychological mechanisms that ultimately mediate belief.
And that's why arguments concerning the psychology underlying belief are unlikely to persuade people on either side of the evolution/creation debate. For each side, independent evidence seems to support their own views, while the explanation argument potentially undermines the alternative.
 For the experimentally and statistically inclined, here is more information about my informal pilot study. There were 85 participants (71% male) recruited through Amazon Mechanical Turk, ranging in age from 18-67 with a mean age of 29. Each participant responded to a single item: A (Andrea), B (Bob), C (Catherine), or D (Doug). Responses for A and B were significantly more likely to be "yes" than "no," with responses for C and D no different from chance.
 For these cases, it is possible that the evidence isn't perfectly independent. Perhaps believing that you're a better-than-average driver boosts your confidence and performance on a driving test. And if Bob believes that he does 60% of the housework and knows he's being watched, he might in fact do more. But these complications aren't important for the points being made here.