I read an interesting article today, here are some excerpts:
To get at this issue, Munro polled a set of college students about their feelings about homosexuality, and then exposed them to a series of generic scientific abstracts that presented evidence that it was or wasn't a mental illness (a control group read the same abstracts with nonsense terms in place of sexual identities). By chance, these either challenged or confirmed the students' preconceptions. The subjects were then given the chance to state whether they accepted the information in the abstracts and, if not, why not.
Regardless of whether the information presented confirmed or contradicted the students' existing beliefs, all of them came away from the reading with their beliefs strengthened. As expected, a number of the subjects that had their beliefs challenged chose to indicate that the subject was beyond the ability of science to properly examine. This group then showed a weak tendency to extend that same logic to other areas, like scientific data on astrology and herbal remedies.
A second group went through the same initial abstract-reading process, but were then given an issue to research (the effectiveness of the death penalty as a deterrent to violent crime), and offered various sources of information on the issue. The group that chose to discount scientific information on the human behavior issue were more likely than their peers to evaluate nonscientific material when it came to making a decision about the death penalty.