Poll-y Want a Cracker

I’ve seen and read and even participated in a lot of polls in my life. Telephone polls (that’s a pun), lighting polls, barber polls (remember barber poles? Remember when jokes didn’t rely on puns?), probably hundreds of them (polls, not jokes … well, maybe jokes too). Many of them claimed to be scientific polls. Me? I’m a big believer in science. “Scientific” polls? Not so much. My take on polls after 60-plus years monitoring them? Sometimes they get it right, and sometimes they don’t. Sometimes you feel like a nut, sometimes you don’t. But I’ll save specific politicians for another time.

I find polls that claim to be scientific, suspect. I mean, what makes a poll “scientific”? Working out averages? Checking the skill levels of a specific group of pollutants (sorry, pollsters). Comparing the poll replies of individual ethnic groups against one another? Let’s take any individual from any of those groups. Or any individual from any group. Let’s say you’re a 30ish highly-paid coder working for a booming software startup in Palo Alto (California, not Bolivia). I reckon the current scientific polling is going to assume you’ll vote for a Democrat and count you in that column.

But what if you have stock options? Might that incline you, other considerations notwithstanding, to vote the other way? Would scientific polling somehow take that into account?

Or let’s take the polar opposite (okay, I’ll try to stop now). Suppose you’re a rancher in western Washington state who is strongly pro-second amendment, feels free to deal with local predators on his (or her) own terms, harbors an intense dislike of the mainstream media (why doesn’t anyone ever point out that Fox is part of the mainstream media?), and supports the breaking away of their part of the state to join with Idaho? Might polling reasonably assume this to be someone residing in the Republican camp?

‍

‍

So here’s the problem with both those examples and with polling people in general:

‍

1. You can’t generalize, especially about Americans. People lie.

2. When confronted with an actual ballot, people have this awkward tendency to — change their minds.

3. You can run the most expensive, the most comprehensive, the most scientific poll imaginable — but any of it is only marginally trustable because of those three mindsets. Our first example might declare to all and sundry, friends, relatives, coworkers, even one’s betrothed, that they’re for sure definitely positively voting for one when they will actually vote for the other. Or a third-party candidate.

‍

As for our second example, the same might be exactly save that they’re voting for one when in public they loudly declare their fealty to the other. Or a third-party candidate. Or write in Taylor Swift.

This is why I pay attention to but never fully believe in polls, scientific or otherwise. The problem with polls is that those answering the pollsters’ questions are human beings. The secrecy of the voting booth or in voting at home is tighter than that to be found in church confessionals. If you want to vote for someone who would make your friends quiver in disgust, you are free to do so. And in so doing, you are making a joke of “scientific” polling. Never mind that in primaries people sometimes vote for a weak candidate in the hope that they will garner enough votes to win the primary, thereby entering the actual election unable to defeat the opposition candidate that the voter really wanted to win in the first place.

Suspect polls needn’t be just about politics. Polling someone on who they think has the best fast-food french fries? Maybe they work for McDonald’s, or Cane’s. Which is the best airline? Perhaps your subject (or your subject’s spouse) works for United or Southwest?

I firmly believe that many of those folks who reply to polls provide answers that support their personal views rather than the actions they really intend to take. Consider human-caused climate change. More than 99.9% of peer-reviewed scientific papers agree that climate change is mainly caused by humans, according to a recent survey of 88,125 climate-related studies (source: Cornell Chronicle and elsewhere). A poll conducted among members of the 117th Congress shows that 139 of its members do not.

If we accept the first poll (survey) as well as the second, then either 139 members of Congress are complete idiots when it comes to climate science (not an impossibility, I suppose), or they’re lying due to the nature of their constituencies and for general political reasons. I would tend to go with the lying.

That’s a funny thing about science: you don’t get to pick and choose the parts you like and discard the rest.

But I reckon that’s a subject for a poll.