Trump approval ratings thread 1.2

[Sorenroy]

Decided to try my hand at the Morning Consult poll. I have seen it said on the Atlas before that you can add seven or so percent to the approval numbers and get a good idea of how the election might look (in a two way race). For the national popular vote (no MoE given) this would show Trump tying a Democratic challenger 50% to 50%, which is already better than Trump did in 2016. If you apply this state by state, it looks like Trump is in a pretty good spot:

Donald Trump — 279
Dem. Challenger — 259

While the national poll does not have a MoE, the state by state numbers do. As such, if we look at the low end for Trump (approval-MoE+7%) we find this...

Dem. Challenger — 314
Donald Trump — 224

...and if we look at the high end for Trump (approval+MoE+7%) we find this:

Donald Trump — 333
Dem. Challenger — 205

On a side note, who in the world orders their states alphabetically by abbreviation? Apparently Morning Consult, that's who.

[Sorenroy]

It looks like you messed up the quote a little bit here. It should look like this:

Approval +7? That's pretty generous to Trump considering past occurrences.

In 2004 and 2012, the incumbent President got roughly 51% of the popular vote. Translating that to just the two-party vote, Obama got 52% in 2012 while Bush got around 51.5% in 2004.

In late October 2004, Bush's approval was 48%. In late october 2012, Obama's was 50-52%. Compare that with election day results, and Bush got around +3 from his approvals while Obama got jack squat. Thus we should take these approvals at face value for 2020 instead of giving Trump a bonus that previous Presidents running for re-election did not enjoy.

The model is by Nate Silver, and it refers to approvals early in the year. Electoral results in November will usually be very close to approval ratings in October.

Average incumbent facing an average challenger. Silver's list of incumbent Governors and Senators included George Allen, Jr. (2006), who began with an approval rating of 51% in the spring but still lost. He hitched his star to Dubya in 2006, which would prove a mistake. He ran an execrable campaign, culminating in an ethnic slur and in an incident in which his staffers beat up a critic. He also faced a challenger of extreme effectiveness. The usual advantages of incumbency and being connected to an incumbent President who did not yet have execrable approval numbers vanished. The usual challenger to an incumbent Senator or Governor is usually a weaker candidate, the incumbent usually showing why he won the last time.   

I agree that adding +7% is waaaay too generous, but I'm not so sure about going with face value on his approvals either.  I'd say it'd be fair (and cautious) to add 2-3%.

From October, 2-3% is the usual gain for an incumbent President from approval levels if a politician is still actively campaigning until Election Day. There was a difference between the campaigns of Dubya in 2004 and Obama in 2012: Obama practically shut down his campaigning because of Hurricane Sandy. He was going to win anyway. Dubya campaigned until the last hour because he had to, and there was nothing to stop him.

But we can't predict hurricanes, earthquakes, tornadoes,  or volcanic eruptions eight months earlier, can we?

Also, yeah. I can't find the original 538 article, but that number (+6 or +7) has been thrown around enough that I figured I'd at least try and make a map out of it. Plus I find that sort of thing fun. Really, it's a win-win all around.

[Sorenroy]

Wish we could get a good 50-state approval poll.

It's a Christmas miracle!

Here's the link (although I can't find one that breaks down the numbers state by state):
https://www.realclearpolitics.com/docs/2017_Reuters_Tracking_-_Core_Political_12_20_2017.pdf

[Sorenroy]

I heard something about a state by state Gallup poll that was supposed to be released in early 2018. Anyone know anything about that?