If Bryan Caplan ever does another Ask Me Anything, I think I'll ask him this:
In Myth of the Rational Voter, one of the important ways you prove voter irrationality is by comparing voters to experts. But there isn't much reflection in the book over who the relevant experts are. After all, you conclude here that sociology and economics study the same thing.
So if sociologist were polled and compared to the public, do you really think we'd find biases along the same lines you describe? Why shouldn't this be how we test voter irrationality?