Well, it turns out Dave Roberts has been going about this talking to climate skeptics thing all wrong. If you want to get people to consider data that doesn't fit with their pet worldview, you should make them think really hard about how great they are. Then they'll be putty in your hands! And if you don't believe me, have I mentioned how fetching you look today?
Here's how the study that showed this effect worked:
Nyhan and Reifler once again confronted partisans with information on these subjects that (presumably) contradicted their beliefs—but there was a twist. This time, the contradictory information was sometimes presented in the form of a convincing graph, showing a clear trend (in attacks, jobs, or temperatures). And second, sometimes the individuals went into the manipulation after having undergone a “self-affirmation” exercise, in which they were asked to describe a positive character attribute or value that they possessed, and a situation in which showing that attribute or trait made them feel good about themselves.
And in both cases, the manipulation worked—although by different means.
Super-convincing graphs worked because they were super-convincing. Having people contemplate being good enough, smart enough, and goshdarnit people liking them worked because … well, in theory it worked because it bolstered their sense of identity, making them more willing to reexamine a belief they'd staked their self-concept on. Possibly it just worked because they were like "fine, global warming, whatever, now go away so I can get back to making sex eyes at this handsome devil in the mirror."
Not all super-convincing graphs can do the job (cough hockey stick cough), so maybe rampant ego-stroking should be our new line of attack.