I admit that my distrust of surveys, or polls, has little statistical evidence to back it up — not that I’ve researched the question. It’s just that they seem to be everywhere — especially in an election year — and they are reported in the media as hard evidence of what the larger, unpolled population actually thinks. Well-paid pundits and journalists then use the polls to announce their latest predictions about the “mood” of the country as a whole.
I’ve often wondered why we believe these polls in the first place. Thankfully, the media never ties to describe the complex laws of probability that, supposedly, justify the pollsters’ conclusions; we would all fast-exit if they did. I know I’d never consent to such torture, and only a truly committed sorehead would take on the challenge in the hope of exposing the “science” as a fraud.
And yet, secretly, don’t many of us think it is? But, if we’re not willing to put in the “hard labor” of learning this pseudoscience, does that mean we are to be forever besieged with polls that claim a national significance based on the droppings of 1,154 unidentified bozos?
I propose a half-measure as a kind of defense, at least until a hardier sorehead than me comes along to do the dirty work. Why not have each polled respondent answer a question of fact about the subject they are asked to have an opinion about? It should be a plain, simple fact, but one that everyone should be expected to know if we are to take their “informed opinions” seriously. Let’s say, Mitt Romney was the governor of what state? Or maybe, what is a governor?
I think this is important because the pollsters and pundits never explain why their volunteer army of pollees are representative of us. We are expected to take that on faith. A “plain fact” test — like the one I propose — might make them hesitate a little. In fact, I don’t think the poll interpreters should even know those “plain fact” answers until after they make their “scientific” conclusions about the other questions. It should be easy to hide the computerized responses from the pollsters, and then publish them widely. Maybe on the Comedy network.