13 October 2016

Battleground/National Report, 12 October, 2016 - T-minus 27 days (Clinton 344 / Trump 194)

Wednesday, October 12th, 2016, was exactly 27 days before the General Election on November 8, 2016. Less than one month to go...

Complete polling results for all states for 2016 to-date are here as an EXCEL table for 10/12/2016 in GOOGLE DOCS. All polls are hyperlinked.


So, my methodology for doing averages in this wild-and-wooly-sometimes-3-way election was explained in yesterday's starting Battleground Report, from 11 October, 2016.

The Battleground table is pretty darned detailed. The LEGEND for said table is HERE. In fact, you need to read it in order to develop an "eye" for the Battleground tab.

The polls from yesterday caused no real change in the battlegrounds. The nine states that were battlegrounds yesterday are still battlegrounds today:






First off, the percentage at the bottom is how much of the electoral college, as we like to call it, is represented by the Battleground states. So, right now, just about 21.4%, or a little more than 1/5th, of the EC can be found in the battlegrounds. Since a candidate needs to get to 270, then that would also be 42.8% of all the electors one would need to get to 270. This is why when the battlegrounds fall, they tend to fall like dominos for the winner on election night and that candidate usually comes easily over 270.

To know if there have been any Battleground polls from 10/11 to 10/12, all you have to do is to look at column G. If there is a value there, then you know that a poll for that state came in on 10/12. Notice also that I included the state of Maine in the Battleground table (outside of the 5% margin) since it is currently under 11%. I am in no way implying or inferring that ME is going to go for Trump, no more than I am inferring or implying that IN is going to go for Clinton, but fair is fair and I stick with the same method for all sides.

So, going down the list:

A Monmouth poll came in for Missouri showing Trump +5 points in front of Clinton in a three-way matchup and this number is so much smaller than the other numbers, it shifted his aggregate down two points. So a new poll in a relatively small polling DNA gene pool can make a very big splash. It's simple math.

Conversely, in Ohio, a Baldwin-Wallace poll came in showing Clinton at +10 in a 2-way and +9 in a 3-way. Those are landslide margins, btw. However, there are more polls in the Ohio gene pool than that of Missouri and so the shift is smaller, but yet significant. Clinton's aggregate in Ohio climbed from +1.88 to +3.18, a jump of 1.30 points and right around President Obama's end-polling statistic from 2008 (column K). I see no BW polls from the 2012 and based on this article, it looks like BW just shortly jumped into this kind of thing at the presidential level and will also be specifically polling Cuyahoga County, which is the motherlode County for a Democratic candidate in Ohio, so I really can't analyze whether there may be a L or R bias in the poll at all. That being said, the survey size and time frame and projected MoE all look kosher to me.

In Florida, an Opinion Savvy poll came out showing Clinton up on Trump by +2.8 in a three-way match, which shifted the aggregate 0.8 points in her direction, from +2.11 to +2.91, far above the end-polling statistic for 2012.

In Nevada, a PPP (D) poll came out showing Clinton with the same +4 lead over Trump as the Ipsos/Reuters poll from the day before, thus bumping her aggregate up by 0.57 points, from +2.48 to +3.05. This is UNDER President Obama's end-polling-statistic from 2012.

In Wisconsin, a Marquette University poll came out showing Clinton +7 in a 3-way and +4 in a 2-way against Trump, bumping up Clinton's aggregate 0.9 points, from +5.69 to +6.59. Marquette is pretty much one of two gold-standard polls for Wisconsin; however, it has a slight mathematical bias to the Left, based on it's past performance. It's final poll predicted Obama +8 in 2012, he won by +7, so the poll was off 1 point to the Left. A one point variance, however, is considered slight.


In non-Battleground polling, the shock poll of the day came out of Utah, where Y2 Analytics shows a tie between Clinton and Trump (26/26) and Evan McMullin, an independent candidate who is from Utah and who is also a Mormon, is at 22, while Johnson is at 14. So, it is essentially a three-way tie in the most Conservative state in the nation. Utah is traditionally an R+40 state, so seeing these numbers really is a shock. Then again, polling out of Utah all year long has been showing Trump struggling mightily. Could we have a situation with a one-hit wonder, ala Robert LaFollette in Wisconsin in 1924, where a third or fourth party candidate wins one single state as "favorite son"? Well, it certainly is possible, but considering that the Democrats never even consider campaigning here and the state is not in their electoral math, should the state go for McMullin or even Johnson, then that makes for 6 EV less for Donald Trump, plain and simple. So yeah, I am watching Utah very closely.

And as mentioned above, in Maine, an MPRC poll came in showing Clinton up by 8.7 in a 2-way and by 7.9 in a 3-way, but these numbers are considerably less than the +13 or so from UPI and Reuters, so the margin aggregate shrank considerably. Still, +10.78 is a hefty landslide margin. What is important to note is that the poll also polled by CD and in ME-02, Trump is up by +1, so the real possibility exists that, regardless of national outcome, that Maine may split it's electors this time around. I am keeping a watchful eye on Maine...

The electoral map from yesterday to today has not changed, it is still at:





On the national level, here the numbers:






If you check out yesterday's report, then you will see that the poll released on 05.10.2016 (european dating method) fell out of the statistic (one week time frame) and a new poll, from Insights West (a Canadian firm) came into the statistic. Some of the dates jump around because I just past in new polls from repeaters like Dorsife, etc. But those are the 13 polls from yesterday. Overall, Clinton's aggregate shrank slightly, from +5.74 to +5.52, a -0.21 shift. Were this a +0.5 race ala Bush vs Gore 2000, then a 0.2 variance would be big. Here it is slight and may also be considered statistical noise.

So, let's start keeping track as of today:

12 October 2016, Clinton +5.52, -0.21

This list will grow tomorrow, of course...


And some end-statistics: I first started collecting 2016 presidential polls at the beginning of October 2013 and since then I have logged 3,958 matchups from 1,869 individual polls. You can find the breakdown to this stuff at the EXCEL table, in the tab that says "poll totals".

For past reference, here is the BATTLEGROUND REPORT from 28 days before the GE 2012, from 10/10/2012And here is the BATTLEGROUND REPORT from 28 days before the GE 2008, from 10/08/2008.

-Statistikhengst

No comments:

Post a Comment

Constructive comments and critique are always welcome. Please keep it polite and respectful.