The Pew Research center takes a look at Americans' political leanings and, perhaps predictably, I'm horrified. 65 percent of the public "favor government guaranteeing health insurance for all"? 86 percent of those surveyed favor raising the minimum wage, including 80 percent of libertarians? Well, that last datum suggests that there's something fundamentally flawed with this study. It assigns categories to respondents not based on self-identification but on the basis of Pew's own notions about what constitutes "conservatism," "liberalism," "libertarianism," and "populism."
People were sorted into the four categories based on the combination of socially liberal (or conservative) and economically liberal (or conservative) answers they gave. To be included in one of the four groups, a person needed to provide at least two answers consistent with either the social or economic dimension and at least one consistent answer in the other dimension – while also giving no more than one inconsistent answer in each dimension.
In other words, liberals tended to give consistently liberal responses to the six questions we chose, while conservatives gave consistently conservative responses. Populists, by contrast, gave conservative responses to the social issue questions but liberal responses to the economics questions. Libertarians took the opposite approach, giving conservative responses to the economic questions and liberal responses in the social issue sphere.
The minimum-wage question was not one of the original six used to determine respondents' ideology. As a look at the Pew page will show, there are other results that are not just "surprising" but indicative of flawed methodology — only 37 percent of libertarians and 48 percent of conservatives think that Bush's tax cuts should be made permanent, for example. I'm prepared to accept that self-identified conservatives and libertarians might diverge from the picture of conservatism and libertarianism that emerges from an acquaintance with either ideology's literature — it's not inconceivable that self-identified conservatives may be less "conservative" than we think, for example. But the results are so far off from common-sense expectations on tax cuts and the minimum wage that a reasonable conclusion here is just that Pew has manufactured inaccurate categories.
It would be interesting to see how self-identified conservatives and libertarians match up against the categories assigned by Pew. My guess would be that most of Pew's "populists" would call themselves conservatives; and they do, according to Pew, self-identify as Republican or Republican-leaning by a significant plurality — as do Pew's "libertarians." I suspect the majority of the latter are actually just "moderate" Republicans, who could be expected to have "liberal" views on social issues while retaining mixed free-market opinions relative to "populists" and "liberals." Again, while allowing that self-identified libertarians might not fit the conventional stereotype exactly, it just defies reason to believe that 80 percent of libertarians support higher minimum wages and 63 percent don't think Bush's tax cuts should be made permanent.