More details on leaked internal polls from the Trump campaign

[7,052,770]

Not to say Trump is doing 4-D chess or anything, but I suspect trying to depress Democratic enthusiasm via trying to push the "inevitable Trump loss" will be a key strategy of the Trump reelection campaign, especially since it probably won't have the same effect on his supporters.

That strategy will never work again on Democrats, at least until people too young to remember 2016 make up a big chunk of the electorate.

It's the primary reason I feel really good about Democratic chances over the next few decades. Millions of Democratic leaners learned a agonizing, brutal lesson on 11/8/2016, and will not make the same mistake again.

[Annatar]

I doubt Biden will beat Trump by the margins these polls indicate but he certainly has the opportunity to beat Trump, in my opinion Biden's ceiling would be a 7% win in the NPV even were there to be an economic downturn because we know how well a generic dem would do in that scenario since we have 2008 as a reference point. In 2008 you had a collapsing economy, 2 unpopular foreign wars and an incumbent Republican President who was at 27% approval, these factors allowed Obama who basically was a generic dem to win by 7%, that's about Biden's ceiling as well in my opinion.

[LiberalDem19]

« In Wisconsin, Pennsylvania, Florida and Michigan — three states where Trump edged Democratic nominee Hillary Clinton by narrow margins that proved decisive in his victory — Trump trails Biden by double-digits. In three of those states — Wisconsin, Pennsylvania and Florida — Biden’s leads sit outside the poll’s margin of error.

Trump is also behind the former vice president in Iowa by 7 points, in North Carolina by 8 points, in Virginia by 17 points, in Ohio by 1 point, in Georgia by 6 points, in Minnesota by 14 points, and in Maine by 15 points. »

In other words Trump is losing
VA by 17
PA by 16
ME by 15
MN by 14
WI by 10
NC by 8
FL by 7
IA by 7
GA by 6
OH by 1

Some of this numbers are so bad that it’s just difficult to believe them. I means if these numbers are accurate Trump is losing the PV by 15

The PV is more like Biden +10 with those numbers probably. 15 would mean Texas's PVI has shifted to the right since 2016, and PA is gonna be more than 1% to the left of the country if Biden ends up being the nominee.

[junior chįmp]

If these are Trump's numbers now, I can only imagine what they'd look like with a recession

[Amenhotep Bakari-Sellers]

It wouldn't matter due to fact if Trump loses AZ, since McCain in no longer a factor in GOP politics,  the game is already over.

McCains have already endorsed Biden

[R.P. McM]

Ohio being more Republican than Georgia and North Carolina is also really hard to believe, though who knows what's going on in Ohio these days...

Huh? It already happened in 2016. GA-GOV was closer than OH-GOV in 2018....

Fair enough, but Georgia is still mostly a conservative state, except for Atlanta. Ohio, not so much. Ohio is more populist than conservative at least, and Trump is not exactly popular for the time being. Thus his entire populist shctick falls to the ground, vanishes into thin air. Georgia is static, Ohio is not. 2020 results will be very different from 2016 results in most states. If past results is what predicts the future, New Hampshire will be the state that changes the most in negative direction for Trump, Mississippi the state with the least movement.

Yep:
____
|GA|         R-votes   ┊   D-votes 
2008 ....  2,048,759 ┊ 1,844,123
2012 ....  2,078,688 ┊ 1,773,827
2016 ....  2,089,104 ┊ 1,877,963

The question for Democrats in GA is whether they can scrounge up 2.1 million votes. If they can't — and it's a big leap — Republicans will hold the state for another cycle. It's a different story in the Big Ten region. In WI/MI/PA, and, to a lesser extent, IA/OH, recent history puts the Democratic ceiling well above Trump's 2016 performance. There's a sizable pool of persuadable voters, it's just a matter of earning their support. In GA, close as it's been, strongly as it's trended, the universe of potential Democratic voters may still be too small to close the gap.       

[pbrower2a]

Growth in Georgia's one giant metro area, and it is coming mostly from people bringing their D-leaning Yankee votes with them. We see this to a large extent in Virginia and a lesser extent (so far, but approaching the tipping point) in Arizona, North Carolina, and Texas. Getting out the black and Hispanic vote will still be essential in Georgia.   

[pbrower2a]

Ohio being more Republican than Georgia and North Carolina is also really hard to believe, though who knows what's going on in Ohio these days...

Huh? It already happened in 2016. GA-GOV was closer than OH-GOV in 2018....

Fair enough, but Georgia is still mostly a conservative state, except for Atlanta. Ohio, not so much. Ohio is more populist than conservative at least, and Trump is not exactly popular for the time being. Thus his entire populist shctick falls to the ground, vanishes into thin air. Georgia is static, Ohio is not. 2020 results will be very different from 2016 results in most states. If past results is what predicts the future, New Hampshire will be the state that changes the most in negative direction for Trump, Mississippi the state with the least movement.

Yep:
____
|GA|         R-votes   ┊   D-votes 
2008 ....  2,048,759 ┊ 1,844,123
2012 ....  2,078,688 ┊ 1,773,827
2016 ....  2,089,104 ┊ 1,877,963

The question for Democrats in GA is whether they can scrounge up 2.1 million votes. If they can't — and it's a big leap — Republicans will hold the state for another cycle. It's a different story in the Big Ten region. In WI/MI/PA, and, to a lesser extent, IA/OH, recent history puts the Democratic ceiling well above Trump's 2016 performance. There's a sizable pool of persuadable voters, it's just a matter of earning their support. In GA, close as it's been, strongly as it's trended, the universe of potential Democratic voters may still be too small to close the gap.       

Here's one measure of a voting trend: how states voted in  the House elections of 2018. This could suggest a trend.

Alabama

Republican: 972,927 (58.8%) (1 uncontested race)
Democrats: 675,269 (40.8%)

Alaska

Republican: 128,516 (53.7%)
Democrats: 109,615 (45.8%)

Arizona

Democrats: 999,328 (49.8%)
Republicans: 989,802 (49.3%) (1 uncontested race)

Arkansas

Republicans: 553,536 (62.6%)
Democrats: 310,572 (35.1%)

California

Democrats: 5,041,566 (63.7%) (1 race with no candidate)
Republicans: 2,747,904 (34.7%) (8 races with no candidate)

Colorado

Democrats: 1,252,603 (52.4%)
Republicans: 1,050,938 (44.0%)

Connecticut

Democrats: 811,194 (61.0%)
Republicans: 508,669 (38.3%)

Delaware

Democrats: 227,353 (64.5%)
Republicans: 125,384 (35.5%)

(Florida -- votes are still being found and discovered, so no count. I am treating it as a tossup).

Georgia

Republicans: 1,981,713 (52.4%) (1 uncontested race)
Democrats: 1,802,475 (47.6%) (1 uncontested race)

Hawaii

Democrats: 287,735 (75.3%)
Republicans: 87,296 (22.8%)

Idaho

Republicans: 366,054 (62.0%)
Democrats: 204,020 (34.6%)

Illinois

Democrats: 2,651,012 (60.4%)
Republicans: 1,714,804 (39.1%)

Indiana

Republicans: 1,178,371 (56.6%)
Democrats: 897,632 (43.1%)

Iowa

Democrats: 656,986 (50.4%)
Republicans: 607,827 (46.6%)

Kansas

Republicans: 549,563 (53.9%)
Democrats: 447,134 (43.9%)

Kentucky

Republicans: 935,565 (59.6%)
Democrats: 613,070 (39.0%)

Louisiana

Republicans: 835,603 (57.2%) (1 uncontested race)
Democrats: 553,008 (37.9%)

Maine

Democrats: 328,409 (52.7%)
Republicans: 241,180 (38.7%)

Maryland

Democrats: 1,414,473 (64.9%)
Republicans: 717,945 (32.9%)

Massachusetts

Democrats: 1,529,641 (74.9%)
Republicans: 486,192 (23.8%) (4 uncontested races)

Michigan

Democrats: 2,108,119 (52.0%)
Republicans: 1,826,335 (45.1%) (1 uncontested race)

Minnesota

Democrats: 1,420,669 (55.2%)
Republicans: 1,125,569 (43.7%)

Mississippi

Republicans: 444,092 (50.6%) (1 uncontested race)
Democrats: 369,782 (42.1%)

Missouri

Republicans: 1,318,481 (55.1%)
Democrats: 1,016,096 (42.5%)

Montana

Republicans: 251,611 (51.1%)
Democrats: 227,036 (46.1%)

Nebraska

Republicans: 424,682 (62.5%)
Democrats: 255,053 (37.5%)

Nevada

Democrats: 491,004 (51.1%)
Republicans: 439,401 (45.8%)

New Hampshire

Democrats: 310,320 (54.4%)
Republicans: 249,714 (43.8%)

New Jersey

Democrats: 1,731,037 (59.6%)
Republicans: 1,139,101 (39.2%)

New Mexico

Democrats: 398,753 (58.2%)
Republicans: 262,138 (38.3%)

New York

Democrats: 3,664,970 (66.6%)
Republicans: 1,751,620 (31.8%) (6 uncontested races)

North Carolina

Republicans: 1,830,219 (50.5%)
Democrats: 1,748,018 (48.2%) (1 uncontested race)

North Dakota

Republicans: 192,733 (60.3%)
Democrats: 113,891 (35.6%)

Ohio

Republicans: 2,245,403 (52.3%)
Democrats: 2,019,120 (47.0%)

Oklahoma

Republicans: 730,531 (62.0%)
Democrats: 428,452 (36.3%)

Oregon

Democrats: 1,034,344 (57.4%)
Republicans: 686,952 (38.1%)

Pennsylvania

Democrats: 2,669,469 (54.9%)
Republicans: 2,179,246 (44.8%) (1 uncontested race)

Rhode Island

Democrats: 239,694 (65.0%)
Republicans: 128,831 (35.0%)

South Carolina

Republicans: 927,504 (54.3%)
Democrats: 757,499 (44.3%)

South Dakota

Republicans: 202,673 (60.3%)
Democrats: 121,002 (36.0%)

Tennessee

Republicans: 1,276,040 (59.2%)
Democrats: 843,658 (39.2%)

Texas

Republicans: 4,104,555 (50.4%) (4 uncontested races)
Democrats: 3,824,300 (47.0%)

Utah

Republicans: 510,244 (58.7%)
Democrats: 307,151 (35.4%)

Vermont

Democrats: 188,547 (69.2%)
Republicans: 70,705 (26.0%)

Virginia

Democrats: 1,864,483 (56.3%)
Republicans: 1,407,791 (42.5%) (1 uncontested race)

Washington

Democrats: 1,734,775 (62.8%)
Republicans: 947,374 (34.3%) (2 races with no candidate)

West Virginia

Republicans: 335,791 (58.4%)
Democrats: 232,856 (40.5%)

Wisconsin

Democrats: 1,358,156 (53.1%)
Republicans: 1,171,456 (45.8%) (1 uncontested race)

Wyoming

Republicans: 127,882 (63.7%)
Democrats: 59,929 (29.8%)[/quote]


Democrats actually gained some votes in the midterm election and Republicans lost some from 2016... it is not a good trend  for Republicans.

[R.P. McM]

Growth in Georgia's one giant metro area, and it is coming mostly from people bringing their D-leaning Yankee votes with them. We see this to a large extent in Virginia and a lesser extent (so far, but approaching the tipping point) in Arizona, North Carolina, and Texas. Getting out the black and Hispanic vote will still be essential in Georgia.  

Yeah, there's little doubt GA is trending Democratic. The only question is whether the 2020 nominee can generate enough votes to overcome the state's hardened Republican bloc. Two or three cycles from now, it'll be a foregone conclusion. But I'm less confident about next year.

Here's one measure of a voting trend: how states voted in  the House elections of 2018. This could suggest a trend.

I think it's a good indicator — less susceptible to candidate quality than a single gubernatorial or senatorial result.

[Stranger in a strange land]

« In Wisconsin, Pennsylvania, Florida and Michigan — three states where Trump edged Democratic nominee Hillary Clinton by narrow margins that proved decisive in his victory — Trump trails Biden by double-digits. In three of those states — Wisconsin, Pennsylvania and Florida — Biden’s leads sit outside the poll’s margin of error.

Trump is also behind the former vice president in Iowa by 7 points, in North Carolina by 8 points, in Virginia by 17 points, in Ohio by 1 point, in Georgia by 6 points, in Minnesota by 14 points, and in Maine by 15 points. »

In other words Trump is losing
VA by 17
PA by 16
ME by 15
MN by 14
WI by 10
NC by 8
FL by 7
IA by 7
GA by 6
OH by 1

Some of this numbers are so bad that it’s just difficult to believe them. I means if these numbers are accurate Trump is losing the PV by 15

Okay, Trump's gonna lose Virginia, but he's not gonna lose by 17 points. I take all these numbers with a whole shaker of salt.

[Catholics vs. Convicts]

Lol

https://twitter.com/MeetThePress/status/1140253188718829568

That graphic is beautiful and hilarious. Even if the numbers are off, its still doing a great job of pissing the orange guy off!

[Gracile]

I know these are Trump's own internals, but I find some of these numbers really hard to believe.

[pbrower2a]

I know these are Trump's own internals, but I find some of these numbers really hard to believe.

If they are systematically wrong -- then by how much? Pollsters are rarely off by more than 5% unless they are bungling, let alone fabricating, the poll.

[Old Man Willow]

I know these are Trump's own internals, but I find some of these numbers really hard to believe.

If they are systematically wrong -- then by how much? Pollsters are rarely off by more than 5% unless they are bungling, let alone fabricating, the poll.

And yet the psychopathic liar in the White House claims his real polls show him leading in "all 17 swing states", and brain dead conservatives believe him. When he inevitably loses, they will never accept the results. It will have been stolen by "illegals" and other nonsense.

[GeorgiaModerate]

I know these are Trump's own internals, but I find some of these numbers really hard to believe.

If they are systematically wrong -- then by how much? Pollsters are rarely off by more than 5% unless they are bungling, let alone fabricating, the poll.

We don't have any supporting data behind these results.  It's possible that they used a turnout model that's very unfavorable to Trump, in which case these topline results would be plausible.

[pbrower2a]

I know these are Trump's own internals, but I find some of these numbers really hard to believe.

If they are systematically wrong -- then by how much? Pollsters are rarely off by more than 5% unless they are bungling, let alone fabricating, the poll.

We don't have any supporting data behind these results.  It's possible that they used a turnout model that's very unfavorable to Trump, in which case these topline results would be plausible.

Using the right turnout model is often the difference between predicting the election well (thus Rasmussen was very right in 2004, 2010, 2014, and to an extent 2016, but terribly wrong in 2006, 2008, 2012, and 2018). Who turns out, and who doesn't?

It is not enough to question whether the politician is 'good' or 'bad'. Even 'objective standards' have some potential for bias and irrelevancy. Politicians can get elected with the aid of vile campaigns.

Trump still wins re-election with both the popular vote and the electoral college with an electorate much like those of 2004, 2010, and 2014. OK, it is no longer 2004... but basically the electorate has changed significantly from even 2014 or even 2016. The Millennial Generation is now most of the newer voters, and they are about 65-35 D in their voting while the voters dying off are about 55-45 R in voting. Add to this, the unusual participation in the midterm voting in 2018 suggests one thing and one thing only: that Millennial adults are taking voting more seriously than they used to. The 'Tea Party' electorate to the extent that it has not died off or gone so senile as to be unable to mark a ballot is still voting-- but its numbers are shrinking due to attrition from death and debility.

The Millennial Generation seems to have little stake in Movement Conservatism, and it can see right through Donald Trump.  The question yet to be answered is whether it can influence older voters who might change their minds on some issues of deep concern to Millennial adults, like heavy debt and low pay?       

[Dan the Roman]

I know these are Trump's own internals, but I find some of these numbers really hard to believe.

If they are systematically wrong -- then by how much? Pollsters are rarely off by more than 5% unless they are bungling, let alone fabricating, the poll.

We don't have any supporting data behind these results.  It's possible that they used a turnout model that's very unfavorable to Trump, in which case these topline results would be plausible.

Almost any Presidential campaign has multiple turnout dynamics for all of their internal polling. These look like a leak of a turnout dynamic, but probably not a very probable one, perhaps at the edge of worst case. It is also quite likely that there is another set of internal polling which shows Trump leading in everyone of these states, if only because creating a turnout model that would produce such a result is a key aspect of figuring out what you would need to do in order to win them, and then evaluate whether it is possible/ worth investing in.

The likelihood then is that

These leaks are real. But they are not what the Trump campaign expects to happen or thinks is likely.

Trump is telling the truth that his campaign has "polling" which shows him winning every one of these states. Again, however, it is highly unlikely the campaign puts much stock in this set of numbers either in terms of probability. Whether they have told the President that is an open question.

[Alben Barkley]

Trump down by 8 in North Carolina according to his own polls?

Clearly it's safe R!

[pbrower2a]

I know these are Trump's own internals, but I find some of these numbers really hard to believe.

If they are systematically wrong -- then by how much? Pollsters are rarely off by more than 5% unless they are bungling, let alone fabricating, the poll.

We don't have any supporting data behind these results.  It's possible that they used a turnout model that's very unfavorable to Trump, in which case these topline results would be plausible.

Almost any Presidential campaign has multiple turnout dynamics for all of their internal polling. These look like a leak of a turnout dynamic, but probably not a very probable one, perhaps at the edge of worst case. It is also quite likely that there is another set of internal polling which shows Trump leading in everyone of these states, if only because creating a turnout model that would produce such a result is a key aspect of figuring out what you would need to do in order to win them, and then evaluate whether it is possible/ worth investing in.

The likelihood then is that

These leaks are real. But they are not what the Trump campaign expects to happen or thinks is likely.

Trump is telling the truth that his campaign has "polling" which shows him winning every one of these states. Again, however, it is highly unlikely the campaign puts much stock in this set of numbers either in terms of probability. Whether they have told the President that is an open question.

As usual, the best corroboration of any poll is another poll.  There has been no shortage of polls of the states decided by 10% or less in 2016. Maybe I have yet to see enough polls for New Mexico, but other than that...that is seventeen states, roughly a third of all states.


If I saw numbers like these a year from now and I were handicapping the Presidential race I would be asking "how many nines?"


   

[pbrower2a]

Again, the 2020  Presidential election will be decided, barring weird events, in roughly one third  (17) of the 50 states and Dee Cee... and two Congressional districts that went for or against Trump by 10% or less. If Trump wins a state like Oregon, Delaware, New Jersey, or Connecticut that he lost  by 10% of more, then he is winning in a landslide. On the other side, should he lose such a state as  South Carolina or Alaska that he won by 10% or more, then he is losing in a landslide. (I mentioned only states decided by 10-15% in 2016 as examples of such.



2016 result among states decided by 10% or less:


 

8% or more -- saturation 7
4% to 7.99% -- saturation 5
1.5% to 2.99% -- saturation 3
under 1.5% -- saturation 2

OK, I dislike the style as a predictor of 2020. I expect Iowa to be closer to the result of Wisconsin because Iowa and Wisconsin have similar demographics, and I don;t expect Ohio to be a runaway. An election that gets the same state results as 2016 probably goes like this:

 

8% or more -- saturation 7
4% to 7.99% -- saturation 5
1.5% to 2.99% -- saturation 3
under 1.5% -- saturation 2

So what does a Trump landslide look like in these states? Trump gains about 5% on results like this.

 

8% or more -- saturation 7
4% to 7.99% -- saturation 5
1.5% to 2.99% -- saturation 3
under 1.5% -- saturation 2

The super-deep shade reflects that Texas is completely out of contention. In this scenario, President Trump has solved lots of problems, solidifying his 'shaky' support of 2016.   Trump is losing Connecticut, Delaware, New Jersey, and Oregon by single-digit margins, in case you are curious, and they appear on this map.

So what do I see in the polling results of the internal poll?

 

8% or more -- saturation 7
4% to 7.99% -- saturation 5
1.5% to 2.99% -- saturation 3
under 1.5% -- saturation 2

This is an electoral disaster for Trump. Alaska and South Carolina enter the zone of marginal contention... probably Indiana because it often swings wildly*, but my model does not allow me that much styling. This is about a 5% swing against Trump, which gives him only about 41% of the popular vote.

*A side note on Indiana: since 1920, no Republican has won the Presidency without winning Indiana by at least 10%. I expect this pattern to hold in 2020.