Bryan Caplan once again tries to make the argument that, contrary to what many on the right might claim, the relationship between the amount of education people have and the rationality of their economic views is a positive one, and that the only reason it might seem otherwise is due to selection bias on the part of those who argue against the notion. Now, in a certain sense, what Caplan says is true enough: if we simply take, say, having a bachelor's degree as the minimum standard for calling someone "more educated", it will be true that those who have less education will also tend to have less sensible economic views than those who have more, and the same would likely still hold if we made the dividing line the possession of a master's degree.
The thing is, though, that while the above statements might seem to essentially concede Caplan's point, the real factor at work here is not the sheer amount of education one has, but the presence of very large numbers of people with substantial educational exposure to economics in our respective samples, and once we exclude such people, the difference between the economic attitudes of more vs. less educated people will likely vanish altogether - going by my own (admittedly subjective) experience, the average PhD holder outside economics and related fields is not a whit more reasonable in his views on free trade, deregulation, vouchers, unionism, minimum wages and the like than the typical BA holder who enters the job market after 4 years of higher education: if anything, the opposite tends to be the case, with diehard leftist economic views being an aberration most easily indulged in an academic cocoon, and those with PhDs being among the least likely to concede that their expertise in one subject does not in itself confer upon them greater insight into the unrelated field of economics, while that very sense of being an expert in something tends to foster the delusion that expertise is generally superior to just letting things play themselves out as is the case in a free-market scenario. Were we to divide the adult population into educational classes with a single class per number of years of education accumulated, I'm convinced that what we'd find if we were to administer an economic literacy test would be that there would indeed be a skew towards the more educated, but what we'd have would merely be a positively skewed gamma distribution, with the least economically literate people being found on both ends of the educational spectrum, and the skew being almost entirely attributable to the very large numbers of individuals who concentrate on economics, finance and related fields during their years in higher education.
To put it succintly, Arnold Kling has it right, and Bryan Caplan's faith in the superior economic judgment of the academic elite is for the most part misplaced: the all-pervasive liberal values which make the most educated take a less favorable view towards a manifestation of xenophobia such as protectionism are the very same set of values which make their attitudes towards many other economic matters so incredibly - to put it bluntly - stupid.