Tues. Models: Trump gains a bit in GA, ME, and NV, but no major changes

Quick State of the Race

Donald Trump’s continuing to gain a bit in the Georgia, Maine and Nevada Models as data from August continue to be replaced by September numbers that still show a tighter race. That said, there are indications over the last 24 hours or so that Trump’s gains last week have peaked and may be ebbing 1-3 points, at least nationally. That means the race is fundamentally closer than it was a few weeks ago, but some of Trump’s peaks last week may be just that — peaks. So, it’d likely take another fundamental shift to the race (debates, etc.) for him to either pull into a real lead or for Clinton to pull back into a healthy mid-digit and large electoral lead.

As it stands, Clinton is still fundamentally favored by a few points, and even as Trump continues to make some gains in many of these states, the Electoral Map is holding at Clinton 272 – Trump 190 without the Toss-Ups. This is happening because even though Trump is improving in states like OH, FL or NV, he’s moved the status of these states from Toss-Ups that favored Clinton to more of pure Toss-Ups. Meanwhile, Clinton hasn’t lost any “Leans Clinton” states to the Toss-Up category as states like Virginia, New Hampshire and Pennsylvania definitely remain in her favored column. Thus, until there are fundamental changes to those states – or Trump takes a real lead in a Toss-Up state – Clinton’s structural advantage will keep holding.

With that, today’s Models, including a breakdown of what’s going on in Maine:


Trump 46.5% (+0.9%)

Clinton 42.1% (–)

  • With every update, Georgia keeps trickling toward Trump. The race remains much closer than normal for a Presidential contest. Part of that is Trump under performing with a few key groups here. The other part is the changing nature of the state as Georgia is part of the new Southeast. It remains “redder” than North Carolina or Virginia, but the terrain of these states continues to change quickly as younger transplants are often much more liberal than the “Old South.” Additionally, the Latino population continues to surge across the Southeast. That said, it still looks like a tough climb for Clinton to actually win here, even if she cuts in half the Romney/Obama margin.
  • The state remains “Leans Trump.”

Maine Statewide:

Clinton 44.3%

Trump 38.0%

  • The Maine polling data certainly are getting more interesting. MPRC is a blue-leaning pollster, so to see their numbers close to a tie in the 4-way statewide and a 5-pt lead in the head-to-head is a bit shocking. But these numbers match a very consistent trend in the data. Simply put, if Clinton really only wins this state by mid-single digits, the 2nd CD Electoral Vote is all but assured to vote Trump as it trends much redder than the 1st CD. The Northern part of the state (the 2nd CD) is a remote, expansive area that heavily voted for Paul LePage (R) twice. So, the CD 2 numbers don’t surprise me, and those numbers are certainly making the state much closer overall. As it currently stands, Maine straddles the “Likely/Lean” Clinton line statewide, and it’s running as close as a Pennsylvania/Virginia-type state.
  • Now, with that said, Modeling out CD 2 is filled with data noise. So, I’ll walk you through three possible options:

Maine CD2:

Statewide extrapolation Model of CD2 w/Fundamental Band:

Trump +0.6%

  • In this formula, I simply take the statewide Model and extrapolate out to the 2nd CD, based on historic voting patterns. It’s a bit of a blunt object approach as it essentially takes away a lot of context from current CD2-specific data. But, it does show a result that would better match a fundamental break in the state.

CD 2 Only Data Model w/Fundamental Band:

Trump 42.5%

Clinton 39.6%

  • The second formula is a hybrid of the first and third. It focuses more on the CD-2 only data, then applies a fundamental/historic line to it. This allows for a snapshot view of what CD-2 would look like if it was in-line with present data, but still corrects back to some historic fundamentals when undecideds, etc. are applied. Trump’s lead grows to 2.9% here, still within the “Toss-Up” category, but breaking toward “Lean Trump.”

CD 2 Only Data Model w/o Fundamental Band:

Trump 42.0%

Clinton 36.3%

  • Finally, if Maine’s 2nd CD really is quite different from historical precedent, it’d break more Trump’s way as the CD-2 only data (without any bend toward historic fundamentals) show a 5.7% Trump lead, and much of this data are still a bit old as the newer and newer the data get, the more they break toward Trump. So, if more double-digit leads come in for him from the 2nd CD, this Model will of course climb as well.
  • In sum, the Model bands are now all in Trump’s favor but range from Trump +0.6% if historic fundamentals dominate (a Toss-Up scenario) to +2.9% if historic fundamentals come into play (getting closer to “Leans Trump”) to a year where the 2nd CD really is following its own path that departs from historic fundamentals (a “Likely Trump”) scenario.
  • As it stands, I’m going to keep this as a “Toss-Up” because of the volatility, but there’s a lot of data to surmise that Trump really may take this EV if current data trends continue into the election.


Model w/Fundamental Band:

Clinton 43.6% (-1.3%)

Trump 42.4% (–)

Model w/o Fundamental Band:

Trump 41.6%

Clinton 41.4%

  • Once again, I’m showing you two Models for Nevada. Why? Because each candidate is ahead, depending on which Model I use. Like the Maine data, the Fundamental band – which is historically more accurate – shows Clinton is ever so narrowly ahead in a blue-leaning state. But, her margin continues to close. Though, that’s because she’s dropped a bit; Trump remains stuck in the low-40s. But, IF the state doesn’t break a bit toward its historical footprint, Trump is actually up two-tenths.
  • But, again, either way you slice it, this one is a pure Toss-Up.

Polls Included in Current Models:

GA: Monmouth (9.18); Ipsos (9.15); Opinion Savvy (9.14); Emerson (9.13); NBC/WSJ/Marist (9/08); WaPo/SurveyMonkey (9.01).

ME: MPRC (9.17); Ipsos (9.15); Colby College/SurveyUSA (9/10); Emerson (9/05); WaPo/SurveyMonkey (9/01); MPRC (8/20); UNH (6.21).

NV: Rasmussen (9.18); Ipsos (9.15); Monmouth (9/13); NBC/WSJ/Marist (9/08); PPP (9/07); WaPo/SurveyMonkey (9/01; Pres Only).

Be the first to comment

Leave a Reply

Your email address will not be published.