Why Polls are Junk

One thing Bitter and I agree on in regards to Trump is that he doesn’t give a rats ass about polling, and that it’s both refreshing and terrifying. My Congressman, Brian Fitzpatrick, a rare specimen of the endangered Republicanus Bloombergensus, I don’t think takes a shit without checking first to see how well it will poll. The problem with leadership by poll is that most polls are junk, and lobbyists pushing one issue or another are happy to put their junk polling in front o lawmakers and their staff. The one Salon is hawking to promote gun control is no exception. Looks good for the gun control folks doesn’t it? Eight percent say it’s their biggest issue! Of those four to one broke for Dems! Oh dear! I don’t want to be on the wrong side of history!

Except if you look at the AP survey, at its root it’s a self-selected poll, which has about zero validity:

We start by mailing a postcard to a random sample of registered voters in 25 states, inviting them to take our survey either online or by phone. We also try to reach those registered voters directly by phone. At the same time, we’re conducting a random-sample survey of registered voters nationwide. Finally, we survey self-identified registered voters in all 50 states using online panels, which allows us to interview a very large number of people in just a few days.

Emphasis mine. Even if your initial sampling is random, if they have to choose whether to participate, the numbers are going to be skewed over what they would be if they were truly random. In truth all polling has this flaw, which is why polling, for the most part, is bullshit. Telephone polling was a better method in the era when people answered their phones and would have felt social pressure to participate. But what’s replacing telephone polling is utter garbage. This survey has all those flaws in spades. No one should pay a lick of attention to this shit. I think an attribute of people who are politically successful after the realignment is they won’t pay any attention to polling. Polls, like candy, are the junk food of the political world.

How do you know what voters in your community think, if you’re the politician representing them? Get out there and talk to them. Get out there in the community. There’s no substitute for that. Don’t just talk to your donors. Talk to real people. Years ago, everyone knew this. Somehow, we’ve forgotten.

5 thoughts on “Why Polls are Junk”

  1. This is a major reason why I like Utah’s caucus system, and am glad to see that Idaho restored the system for their primaries. It encourages candidates to go and talk to delegates selected by caucuses.

    If a candidate can’t convince a sufficient number of delegates to vote for them, it *then* goes to a primary.

    In Utah, this has been short-circuited somewhat, by the legislature requiring parties to accept candidates based on petition (so that a person can go straight to the Primaries without going through the caucus system), but I hope that we can get to the point to remove that blemish.

    I noticed that a lot of people don’t like the caucus system, but that’s because their favorite candidates (Senator Bob Bennett didn’t make it through the caucus system) sometimes gets primaried — and they want to blame the “fringes” for this.

    The fact is, though, any rich person can get petition drives going. If you want to get money out of politics (or at least reduce it), and get candidates to mingle with people, a caucus system is a *very* good way to do that!

  2. I have doubts about your opening line. In contrast, i think Trump is a world class political strategist with a world class polling operation. He A/B tests his messaging and his ideas, then his data operation crunches the numbers. It’s how he stays close to his constituents and knows where we want action. it’s just that he never justifies any behavior based on polling, because polls can lie.

    1. Private polling is different from public polling. If nothing else, they sell accuracy, not opinion.

  3. The reason I have little faith in polling numbers is the way that polls are designed can change how the poll reads. I can design a poll about guns and whether or not people support the private ownership of guns, simply by how I state the question. And that is exactly how these groups that get 80% of people supporting gun registration or background checks go about getting such ridiculous numbers.

    1. Great point. Its quite easy to word a question to lead a person to answer the way you want. This is especially easy in gun law when most of the population doesn’t understand what the law currently is.

      I agree with Sebastian – all polls are bullshit.

Comments are closed.