The Polls - A Tale of Two Stories (the second one might surprise you)

[Jimmy Higgins]

So... the polls. Companies call up a bunch of people, ask them questions (or are hung up on), and put it in to a computer which spits out results. 2020 was a bit like 2016, and the polls didn't necessarily match up with the final result. But lets make one thing clear, the polls were not all wrong. In fact, some of the polls were actually right, despite being a bit far fetched.

The polls can be shuffled into four categories.

1. Accurate as all heck!
2. Barely right
3. Close to within the margin of error, but trending the wrong candidate
4. States that don't matter, so the numbers are probably wrong because models can't handle it.

Now, here is the thing. I just listed out four Categories. And other than 4, I bet most of the readers here don't know which states fall into which categories. I mean, if I told you that Georgia falls into Category 3, you'd probably nod and say yeah, that sounds right. But it wasn't. Georgia in aggregate, was almost dead-on.

Take a state like Iowa... the polls generally showed a 50/50 finish. Except one poll... the poll you should have paid attention to. Ann Selzer's poll. She had it at 7 pts. She did have it near tied, but then it broke heavy for Trump. What did Selzer know that all the other pollsters didn't? Selzer again, was crowned Emperor of Polling for Iowa. But the other polls in Iowa, much like Ohio were just wrong, by 6 to 10 points.

This was a theme of the Midwest (in general). Wisconsin and Michigan... were supposed to be safe for Biden, they were close and not within the margin of error. Indiana polling around 10 points, that was outside the margin of error (though not a lot of polling for Indiana).

So the Midwest was screwed up by the pol... wait... Minnesota. Minnesota was on by the polls. About 7 to 8 pt win for Biden. Polls flirted between 5 and 10. And this begins what is the frustration about the polls. How did the polls get Michigan, Wisconsin, Ohio, and Indiana wrong... but Minnesota right? And it wasn't just Minnesota. Virginia, the polls were showing a 2x margin of victory for Biden compared to Clinton. That happened! The polls indicated a 2x to 3x margin of victory for Biden in Colorado, relative to Clinton. That also happened!

This provides us with a list of Georgia, Minnesota, Colorado, and Virginia, states with varying margins of predicted victories... all pretty darn close to the actual result.

Meanwhile states like Nevada and Pennsylvania, while accurate winner wise, the margin of victory was half to a quarter of that predicted, though within varying percentages of the margin of error. Arizona was hugging hard on the margin of error.

Then we have states like North Carolina and Florida which were close to the margin of error, but floated to the wrong candidate. It is hard to tell if the results of the polls indicate they all consist of noise or whether some states were easier to poll because turnout demographics were easier to predict, though that argument for Georgia is harder to justify... especially with all of the mail-in voting. Or maybe the access to mail-in voting was the reason why they got Georgia right. But why Georgia but not Pennsylvania?

My conclusion is I have no conclusion and a deep statistical analysis of demographics is needed to help determine if the polls are worthless or if the polls in places like the Midwest and PA are missing a demographic shift.

So, we have states that went from anywhere between right-on with the polls and off by a good margin.

 

[T.G.G. Moogly]

Whenever I'm polled I always answer truthfully, I tell the person who I am going to vote for and that's what I do. Does everyone do that? How would one know? To know if responders are being truthful in their answers one would have to know where they lived and then do a mini analysis of persons polled in that area. If it's a red area and the polls were off there and not in a blue area, or at least not as much, then maybe right wingers like to lie when polled.

I did not get polled in this general election but got polled a dozen times in the last one because I was a super voter, a voter that hadn't missed a voting opportunity in a long time.

Anyway, knowing what we do about right wingers and their paranoia about "the media" I think they lie.