Felix Salmon has a gracious reply to my earlier post:
Apologies to John Sides and Jack Citrin for dismissing their science out of hand on my Tumblr: they really are careful and sophisticated researchers, and Sides is well within his rights to give me a good slap.
He then goes on to clarify his thoughts. He points out some additional aspects of "numeracy" that often people cannot achieve. For example:
And more generally, numeracy is about much more than estimating proportions and percentages. It's about comfort with numbers and number lines, and having an intuitive feel for how they work.
I agree.
He also continues to argue that, despite my "carefully chosen counter-examples," there is still reason to think that opinions resist facts:
The general public doesn't want its mind changed, and any changes which do happen are always going to happen slowly. Which is why presidential debates are almost never about who won the argument on any particular point, much as people like myself would love it if they were. The goal in such debates isn't to win some kind of intellectual argument: it's to make as many people as possible think that you think the same thing that they think. People will vote for the person they agree with - not the person with the sharpest wit or the cleverest arguments or the most apposite facts.
Again, I would be more circumspect. There are cases where the public changes its mind, and quickly. See, for example, the 2012 Republican presidential primary. Bachmann, Trump, Perry, Cain, Gingrich, Santorum: people's opinions about these candidates changed rapidly and sometimes changed again.
It's a separate question, however, whether specific factual information will itself change opinions-which was the subject of the paper by Jack Citrin and myself that Salmon originally criticized. Salmon is certainly correct that many studies have found that such information doesn't really change opinions. Our paper cites those studies, and they are probably more numerous than those which have found the opposite. So the challenge is really figuring out why "the facts" matter in some cases but not in others. In any case, I would agree with Salmon's point regarding presidential elections: they typically aren't won by having "the facts" on your side.