POLITICAL POLLING – WHO’S GOING TO WIN?

We now have about a month to go before election day on Tuesday November 3rd. And this is when the polling truly gets interesting because most people have made up their minds and there is just a small percentage left in each of the swing states that are still in play. And the 2020 presidential election will boil down to a small handful of states:

  • Pennsylvania
  • Wisconsin
  • Michigan
  • Minesota
  • Nevada
  • New Hampshire
This image has an empty alt attribute; its file name is IMG_3230-1024x768.jpg

If Biden wins these states, then the election is over, because when added to the states the Democrats never lose — like my home state of Maryland — he will have way more than the 270 electoral votes needed to win. So, these are the states we need to focus on. It really is that simple.

This image has an empty alt attribute; its file name is DSCN5030-1024x768.jpg

Trump needs to win the states he snagged the last time around, or peel off a Democrat state he didn’t win in 2016 — let’s say, Nevada or New Hampshire, though neither one have a lot of electoral votes. But Trump’s bigger problem is that he is losing some big states he won in 2016, and he absolutely, positively, can’t lose them in 2020. These include:

  • Florida
  • North Carolina
  • Arizona
This image has an empty alt attribute; its file name is DSCN5032-1024x768.jpg

It’s hard to imagine him winning without these three normally Republican states — maybe Arizona, if he picks up one of the Rust Belters — but definitely not Florida or North Carolina.

This image has an empty alt attribute; its file name is DSCN5037-1024x768.jpg

Even more troubling for the Trumpster is that it’s really close in Republican stalwart states like Texas, Ohio, and Georgia. If any one of those breaks for Biden, Trump is toast.

This image has an empty alt attribute; its file name is IMG_9275-1024x768.jpg

Okay, so let’s talk about how to figure out how polls work.

This image has an empty alt attribute; its file name is DSCN5025-1024x768.jpg

There are three things you need to focus on:

  • The percentages — how much support each candidate has.
  • The amount of undecideds left in the pool.
  • The margin of error of the poll.
This image has an empty alt attribute; its file name is Pennsylvania-Trip-9-8-2010-7-19-49-1024x768.jpg

And then there’s the magic number which is 50%. In theory, if more than 50% of the voters in a state are voting for one of the candidates, that candidate wins — no ifs, ands or buts. It’s simply mathematics.

This image has an empty alt attribute; its file name is Keith-Secola-DC-8-8-2009-9-44-04-PM-1024x768.jpg

Let’s look at an example from Saturday October 3rd, using the swing state of New Hampshire. The percentages below represent a weighted average of every credible poll that has been taken in the state since January, according to the website 538.

  • Biden 53.0% Trump – 43.5%
  • Biden has a lead of 9.4%
  • Add Biden’s and Trump’s numbers and that shows you there are still 3.5% of the voters who have not made up their minds yet.
  • There is a 3% – 4% margin of error in the poll.
  • Biden has 53.0% (3.0 % above the magic number).
  • Add the Undecideds (3.5) + Margin of Error (3.5%) = 7%.
  • Add the lead (9.4%) + the Magic Number margin (3.0) = 12.4%.
  • Subtract 7% from 12.4% = 5.4%
  • So, it’s over — Biden wins the state by about 5% — unless the numbers drastically change in the next few weeks, which isn’t going to happen absent a Trumpian Miracle.
This image has an empty alt attribute; its file name is DSCN0338.jpg

As we get closer to the election, the number of undecideds will shrink and the polling accuracy will increase.

This image has an empty alt attribute; its file name is DSCN5028-1024x768.jpg

And there are two more factors to consider:

  • How often has a candidate led in a state poll? For instance, Biden has never been behind in any legitimate poll taken in Pennsylvania, Wisconsin, Michigan, Minnesota, Nevada, New Hampshire, Arizona, Florida, or North Carolina. Not once. That means something. And absent some “October Surprise”, you would suspect that steady outcome would prevail.
  • All of the undecideds are not going to break for one candidate. Historically, they tend to divide about 50% – 50%. But either way, they are rarely statistically significant.
This image has an empty alt attribute; its file name is DSCN3602-1024x768.jpg

Which leaves us with the false notion that hordes of Democrats and Independents don’t tell pollsters — for whatever reason — they are voting for Trump. Yes, that can happen like it did in 2016, but it was studied extensively for the past four years by numerous think tanks and universities, and the broad consensus is that it rarely happens in great enough numbers to swing an election. And I would argue that in this election, due to pure embarrassment, more Republicans will tell pollsters they are voting for Trump, and then vote for Biden or not vote, than the other way around.

This image has an empty alt attribute; its file name is DSCN5034-1024x768.jpg

So, with no further adieu, I will now predict the winners of the 2020 election.

  • Electoral College Biden 280 Trump 258 (If the current polling numbers hold true it’s Biden 353 and Trump 185)
  • Senate Democrats 50 Republicans 50
  • House Democrats 235 Republicans 200
This image has an empty alt attribute; its file name is DSCN5977.jpg

Leave a Reply

Your email address will not be published. Required fields are marked *