This is easily explained, I think, but still striking to anyone who knows (or is) a doctrinaire libertarian. Love ’em or hate ’em, you know where they stand on most issues without needing to ask. They want a vastly more limited federal government both domestically and abroad and they want Americans left alone to pursue their happiness in whatever consensual ways they see fit. They want open borders, less aggressive police, and robust rights to privacy and to bear arms. In virtually all things, they’re anti-authoritarian. They’re not unanimous, as no movement is — there’s no dogmatic libertarian position on abortion, as far as I know, and the Pauls want stronger borders, not weaker ones — but I would have guessed that, immigration aside, there’d be 85 percent consensus on all of the foregoing issues if you polled them.

Nope.

When it comes to attitudes about the size and scope of government, people who say the term libertarian describes them well (and who are able to correctly define the term) are somewhat more likely than the public overall to say government regulation of business does more harm than good (56% vs. 47%). However, about four-in-ten libertarians say that government regulation of business is necessary to protect the public interest (41%).

The attitudes of libertarians similarly differ from the public on government aid to the poor; they are more likely than the public to say “government aid to the poor does more harm than good by making people too dependent on government assistance” (57% vs. 48%), yet about four-in-ten (38%) say it “does more good than harm because people can’t get out of poverty until their basic needs are met.”…

But there are only slight differences between libertarians and the public in views of the acceptability of homosexuality. And they are about as likely as others to favor allowing the police “to stop and search anyone who fits the general description of a crime suspect” (42% of libertarians, 41% of the public).

Yes, you’re reading that correctly: Self-described libertarians are slightly more likely than the general public to let police stop people who fit the “general description of a crime suspect.” And that’s not the most eye-popping result from the poll. Check out the second one here:

Not even Rand Paul, who’s spent the year trying to convince righties that he’s a Reaganesque “peace through strength” cautious hawk, would have joined the 43 percent there, I think. The whole thing reminds me of that ultra-weird Pew poll from a few years ago showing that 21 percent of atheists, er, believe in God. In both cases, my instinct is: You’re doing it wrong.

But like I say, this is easily explained. Pew used a two-step process to identify “libertarians,” in order to make sure people who claimed the label really knew what it meant. First they asked people whether they defined themselves that way; if they said yes, Pew then offered them a multiple-choice question asking which ideology best describes ““someone whose political views emphasize individual freedom by limiting the role of government.” The five choices presented were libertarian, progressive, authoritarian, Unitarian, and communist. That’s an awfully easy choice, and conspicuous in its omission of some close cousins of libertarianism, namely, conservativism and “Republicanism.” What you’re seeing in the poll results, I think, is a bunch of doctrinaire libertarians having their brand diluted by a bunch of conservatives/Republicans who are disgusted with those labels right now, for whatever reason, and are thus hoping to claim “libertarianism” for themselves. Do you support aggressive policing, a muscular foreign policy, and a social safety net but are disgusted with how big and intrusive the federal government’s gotten and how complacent the GOP has gotten about it? Congrats, you might be a “libertarian.” In fact, this reminds me of what David Frum said recently about the “libertarian moment”: It’s not so much that conservatives are turning into doctrinaire libertarians, he argued, as that they’re attracted in the age of Hopenchange to the broad libertarian critique that government is malignant, not merely inefficient and stupid. That’s how you get the sort of “libertarians” captured in the poll. They’re deeply distrustful of government writ large, but ask them about particular manifestations of government power — the welfare state, the police, etc — and they’re more simpatico.

All good news for Rand Paul, though. In theory, he’s got a big problem in the primaries: He’s branded himself as a conservative with libertarian tendencies but, on most actual policy matters, I suspect he leans solidly libertarian. The fake libertarians in the Pew poll are the opposite, people who lean conservative on actual policy but crave the anti-government libertarian brand. There’s a mismatch there, but exploiting that mismatch depends on educating conservative voters that Paul’s policies don’t match their own when, I think, most voters are attracted to candidates chiefly because of their “brands.” If you’re a hawkish low-information Republican who’s looking for the biggest middle finger to Washington in the field next year, are you going to let Paul’s foreign policy inclinations deter you from backing a libertarian? Maybe you will if Cruz is in the field too, but what if he isn’t?