Brexit polling: completely wrong and very dumb ✔️

The Brexit failure has many parents, but clearly one culprit is bad research. The polling didn’t work.

Part of the problem was, apparently, “shy” respondents. Speaking on a Charlie Rose panel, British historian Andrew Roberts speculated that “shy Brexiteers” had lied about their intentions, claiming “remain” when they were in fact “leave.”

Hence the total astonishment at the outcome of voting. Hence the failure to see this problem and head it off at the pass.

It’s a little clannish of me to say so, but surely this is precisely why you call on an anthropologist (or some other qualitative specialist).

An anthropologist is good at detecting shyness.

This may be a result of some personality strength or (“takes one to know one”) flaw. But it is also in the very nature of the method. It is tough for a respondent to dissemble successfully over 60 minutes and a series of sinewy questions. The truth will out.

Sometimes the truth is right there before you. As it was, on the Uber trip I took recently. I asked the driver what he thought the issues were for the American presidential campaign. He proceeded to roll out every blank in the Trump platform. Every single blank, phrased just as Trump likes to phrase it.

“So you’re a Trump supporter,” I ventured.

“Oh, no, I wouldn’t say that!” he replied a little huffily.

Of course, he wouldn’t. Clearly, this guy was living under a kind of cover. He didn’t know who I was. I might very well be a Mr. Smarty Pants, a member of the very elite Trump opposes. He wasn’t going to expose himself to my contempt. And who can blame him.

In other cases, detecting shyness takes a little more effort. But it is not so very hard to do. The respondent might lie or withhold. But the “drift” of their politics becomes clear enough.

But let’s be clear about one thing. (That’s the kind of thing they say on the Sunday Morning talk shows I’ve just been watching, and I have fallen, I hope briefly, under the spell of this rhetoric. Normally I don’t like ordering the reader around, but here’s goes.)

Let’s be clear about one thing. The pollsters were not just methodologically unsound, but magnificently naive. If there is ever any chance that the poll participant feels anything but a glassy neutrality for (and from) the pollster, that pollster is obliged to make allowances.

Less abstractly, if there is any chance the poll participant harbors even faint suspicion or hostility for the person holding the clipboard, it is ABSOLUTELY incumbent on the pollster to manage the problem. Don’t just stand there collecting data and checking off boxes. Find some way to allay the suspicions of the respondent. Or call an anthropologist.

In the Brexit case, respondent hostility was very much a possibility. Actually, it was a good deal of the issue at hand. Many people, and millions of Britons, were sick and tired of experts presuming they know better. And who is that person with the clip board if not an expert or someone who works for them?

We should have anticipated “shyness” from the outset. And that means proceeding without an anthropologist was very wrong and completely dumb.