The more I think about it, the more convinced I am that this sort of thing is ultimately caused by the non-overlapping magisteria model. If scientific methods are thought of as something special and other, separated by context and ritual from everyday life, then when someone like your average journalist looks at scientific results, they're parsed as essentially an argument from authority and get evaluated against the opinions of other, nonscientific authorities: for example, religious or ideological. The usual responses to scientifically illiterate arguments don't help: instead of trying to get the point across that this is an invalid way of evaluating results, they usually focus on getting the lay public to accept science as a domain authority above whatever its opponent institution of the moment is -- with predictable results when you're talking to someone that identifies with that institution but not with science.

If we're interested in convincing people to accept scientific results, this suggests to me that we'd be better off downplaying the authority of the journal or institute in question, and instead trying to get people to accept scientific methods as a fully general means of problem-solving.