Trump approval ratings thread 1.1

[Mr. Morden]

The other thread got to 2000 replies, so time for a new one.  Here's the link to the old thread:

https://uselectionatlas.org/FORUM/index.php?topic=252808.2000

[Mr. Morden]

One useful feature on HuffPo's poll aggregation is the ability to look at the aggregate numbers for certain crosstabs.  E.g., here is the trendline for all polls on Trump's job approval among Republicans only:

link

Currently at 80.6% approve, 15.5% disapprove, for +65.1% approval.  Two weeks ago it was at +71%.

[Mr. Morden]

Morning Consult/Politico

https://morningconsult.com/2017/05/19/trumps-approval-sinks-amid-avalanche-bad-headlines/

Approve 41 (-1)
Disapprove 53 (+3)

crosstabs:

Trump job approval margin by region:
Midwest: -12
Northeast: -16
South: -5
West: -18

Trump job approval margin by race:
whites: -3
blacks: -62
Hispanics: -21

Trump job approval margin by income:
under $50k: -17
$50-100k: -4
over $100k: -8

14% of Trump voters disapprove of Trump’s job performance.  7% of Clinton voters approve of Trump’s job performance.

[Mr. Morden]

How low has a president gone and still recovered with no consequences? There was Truman, right? Even he lost Congress and only ran once.

Here's the answer according to Nate Silver:

https://fivethirtyeight.com/features/have-trumps-problems-hit-a-breaking-point/

You must be logged in to read this quote.

[Mr. Morden]

33%???

Wow.

It's hard to imagine how a president could get reflected with that kind of rating. Was Dewey just a really terrible candidate?

No, I think you misunderstand.  Truman's job approval by election day was bad, but not as bad as 33%.  33% is the lowest that he reached in his first term, but he reached it well before election day, and then made a comeback.  That's the lowest that anyone's ever reached before later being reelected.  E.g., I think Bill Clinton was in the high 30s for a while in 1993/1994, but then bounced back enough that by election day 1996, he was reasonably popular.

[Mr. Morden]

Trump down to 79.2% approve / 17.0% disapprove among Republicans in the HuffPo aggregate:

link

[Mr. Morden]

When do congressional Republicans begin backing away from Trump? Mid 30s approval is low but what does it need to be in order for mainstream republicans to distance themselves? I'm not referring to moderates like Ros-Lehtinen or Comstock who are already distancing themselves.

Look at the Republican approval numbers, if it gets below 75% or into the high 60's I think you'll see more jump ship.

Didn't Quiinnipiac had strongly approve down to 63% among GOP? Could be wrong.

They have his job approval among Republicans at 81%, of which 63% is "strongly approve".

[Mr. Morden]

Truman and Clinton were able to get their act enough soon enough to be reelected from this state but no one held on to Congress after something like this.

True, but the GOP victory in 1994 came from a 7 percentage point House popular vote margin.  And given current geographic distribution of Democratic and Republican voters, it's not at all a given that a 7 point margin of victory in the popular vote would actually result in the Democrats winning a majority of seats in the House.  It would probably be pretty close, with a margin like that.

[Mr. Morden]

Truman and Clinton were able to get their act enough soon enough to be reelected from this state but no one held on to Congress after something like this.

True, but the GOP victory in 1994 came from a 7 percentage point House popular vote margin.  And given current geographic distribution of Democratic and Republican voters, it's not at all a given that a 7 point margin of victory in the popular vote would actually result in the Democrats winning a majority of seats in the House.  It would probably be pretty close, with a margin like that.

I would think that it wouldn't translate directly because
1) Trump isn't running
2) As Reapportionment is aging, there might be regression back to natural composition of districts
3) Democrats will run a more targeted campaign and because they just have to go district by district, they will give up on certain issues as they need to stay competitive in that district they lost by 10 or 15 last time.

As I've noted before, if you actually just took the 2016 House results and assumed a uniform swing, then the Dems would need to win the popular vote by as much as ~12 points or so in order to win a House majority.  That's a markedly worse situation for them re: geographic distribution of their voters than, say, you would get from projecting 2006 off of the 2004 results.

But there won't be a uniform swing, and I would assume that voters in swing districts are swingier, and as you said, resources will be targeted to the competitive seats.  So a national victory margin of ~7 points would conceivably be enough to get them a House majority, but probably much less than that wouldn't be enough.

You must be logged in to read this quote.

I don't think there's been a recent House election in which the winner lost the popular vote by much more than about 1 percentage point.  But this is probably the all time worst period for Dems re: the geographic distribution of their voters.  It's only relatively recently (compared to the history of the country, that is) that the Dem-GOP split has become urban vs. rural much moreso than a regional divide.  So urban packing didn't used to be nearly as big of a problem.

[Mr. Morden]

For fun I just plugged in 538's approval ratings for every president since Truman into Excel along with the incumbent party change in seats in Congress. Using the trend line formula of Gains=(2.0188*Approval)-131.02, Trump's current approval would result in roughly a net loss of 54 seats for the GOP. In order for the GOP to hold even, they'd need Trump to have an approval rating of about 65%.

Is presidential job approval most strongly correlated with net gain/loss in # of seats, or with the total # of seats won, or with the national popular House vote, or with net change in national popular House vote since the last election?  And for that matter, is presidential job approval a better indicator than national House race polls?:

http://elections.huffingtonpost.com/pollster/2018-national-house-race

[Mr. Morden]

For fun I just plugged in 538's approval ratings for every president since Truman into Excel along with the incumbent party change in seats in Congress. Using the trend line formula of Gains=(2.0188*Approval)-131.02, Trump's current approval would result in roughly a net loss of 54 seats for the GOP. In order for the GOP to hold even, they'd need Trump to have an approval rating of about 65%.

Is presidential job approval most strongly correlated with net gain/loss in # of seats, or with the total # of seats won, or with the national popular House vote, or with net change in national popular House vote since the last election?  And for that matter, is presidential job approval a better indicator than national House race polls?:

http://elections.huffingtonpost.com/pollster/2018-national-house-race

At least using the very simplified method of "plug it into Excel and give it a trend line" approval is significantly more closely correlated with gains than with overall seats. The R² for gains is 0.6168 for gains, and just 0.0478 for seats for a linear relationship.

I didn't try generic house polls, since I don't have a comprehensive list of where the polls stood for more than the past few cycles. Do you know where I could find one?

No, sorry, I don't know.  Though I think this is something that the 538 crew should look into.  Someone should tweet them about it, because I think it's an interesting question.

[Mr. Morden]

Pew:

39% Approve (-/+)
55% Disapprove (+1)

Crosstabs:

http://www.people-press.org/2017/06/20/public-has-criticisms-of-both-parties-but-democrats-lead-on-empathy-for-middle-class/

[Mr. Morden]

Detailed cross tabs for Pew poll:

http://www.people-press.org/2017/06/20/presidential-approval-detailed-tables-june-2017/

Regional job approval margin:

Midwest: -11
Northeast: -23
South: -9
West: -27

[Mr. Morden]

NBC/WSJ:

http://www.nbcnews.com/politics/first-read/amid-controversy-trump-s-approval-remains-low-steady-n775736



Also, Republican voters now want Congressional Republicans to take the lead on policy rather than Trump:

[Mr. Morden]

Trump surge!

Great president or greatest president?

https://uselectionatlas.org/FORUM/index.php?topic=94046.0

[Mr. Morden]

NPR/PBS NewsHour/Marist poll:

37% Disapprove
51% Approve

You have that backwards.  It's 37% approve, 51% disapprove:

white with college degree: -22
white with no college degree: +14

[Mr. Morden]

Fox News (June 25 - 27):

Approve: 44%
Disapprove: 50%

http://www.foxnews.com/politics/interactive/2017/06/28/fox-news-poll-june-28-2017.html

From that poll:

fav/unfav %:
Melania Trump 51/28% for +23%
Mueller 30/19% for +11%
Obamacare 52/46% for +6%
Pence 47/42% for +5%
Comey 38/39% for -1%
Donald Trump 47/51% for -4%
Ryan 37/45% for -8%
Schumer 26/35% for -9%
McConnell 25/41% for -16%
Hillary Clinton 40/57% for -17%
Pelosi 33/50% for -17%

[Mr. Morden]

Quinnipiac national poll, conducted June 22-27:

https://poll.qu.edu/national/release-detail?ReleaseID=2471

[Mr. Morden]

Quinnipiac national poll, conducted June 22-27:

https://poll.qu.edu/national/release-detail?ReleaseID=2471

men: +1
women: -30
whites without college: +12
whites with college: -20
non-whites: -47

[Mr. Morden]

I said College Grads overall voted for Hillary 49-45% according to the "2016 Presidential Election Page" on Wikipedia. College Grads overall didn't vote for Hillary in the margins that Non-White Voters did.

49-45% is the margin for college graduates who don't also have a post-graduate degree.  College graduates overall voted Clinton 52% Trump 42%:

http://www.cnn.com/election/results/exit-polls

[Mr. Morden]

If Trump's current disapproval ratings in each of the states listed are any indication of his 2020 performance, he'll be absolutely destroyed in the Northeast.  The West isn't looking too good for him either.

The thing is, his favorability rating was already underwater in most of those states back in November (in many cases by more than it is now), yet he won the election anyway.  So really it comes down to the question of how much more of a factor is job approval rating as an incumbent president vs. favorability rating for a challenger.

There's also the question of whether major party presidential nominees with underwater favorability #s is now the "new normal", meaning that we can expect the 2020 Dem. nominee to be unpopular as well.  Even someone like Joe Biden, who's popular right now, could end up becoming unpopular once he wins the Dem. nomination for president, since he'd suddenly become the target of the GOP attack machine, and many GOP-leaning voters who are currently indifferent towards him would instantly discover reasons to hate him.

[Mr. Morden]

http://abcnews.go.com/Politics/months-record-low-trump-troubles-russia-health-care/story?id=48639490

[Mr. Morden]

Is there any other time other than war or recession where a Presidential approval rating has been so low?

Well, we are at war right now, in Afghanistan, Iraq, Syria, etc.  Post-9/11, the war never ends. 

[Mr. Morden]

Is there any other time other than war or recession where a Presidential approval rating has been so low?

Well, we are at war right now, in Afghanistan, Iraq, Syria, etc.  Post-9/11, the war never ends. 

Heh, true, but you get what I mean. I remember Clinton's first months were rocky because of DADT and a number of missteps but nothing like this.

According to this:

http://www.pewresearch.org/fact-tank/2016/01/12/presidential-job-approval-ratings-from-ike-to-obama/



Clinton's job approval bottomed out at 38% in October 1994, when there was no war or recession going on.  Also, I think some of the doldrums for Ford and Carter occurred at times when there was no actual recession (I don't remember the exact timing of the 70s recessions), though the economy was definitely in bad shape with out of control inflation and the like.

[Mr. Morden]

Look, I know it's hard for incumbents to actually lose reelection but is Trump really more than likely to do so seeing he has a 55% disapproval in a state like Utah?

Just to nitpick - Trump is uniquely unsuited for a state like Utah, and even with such a poor image among Mormons, he still did quite well. However, if by October 2020 he had the same approval ratings he does now and was up against a challenger that was about as popular as the average challenger would be, then he most likely would lose, and probably not by a hair either. In an election where his opponent wasn't almost as hated as he was, his "win" most likely would have been a loss with a margin in between Romney and McCain.

If he were to have today's job approval rating on Election Day 2020, then yes, he'd presumably lose.  But I don't think the historical record shows that there's any correlation between job approval rating in a president's first year and his job approval rating three years later.