No, the other polls are not good

You can lead a horse to water, but you can't make it drink.

According to Donald Trump on Sunday, "[t]he @ABC poll sample is heavy on Democrats. Very dishonest - why would they do that? Other polls good!"

Well, news flash forthcoming.  The problems with the polling data have been discussed here at AT for at least a couple of weeks, so it is good to see Mr. Trump finally coming around.  The problem is that apparently, based on a number of detailed reports across conservative through liberal websites, Trump used that same massively flawed polling data as a rationale for firing his highly successful campaign manager.

The ABC News/Washington Post poll in question is indeed "heavy on Democrats."  According to the poll's demographic composition:

Partisan divisions are 36-24-33 percent, Democrats-Republicans-independents, in the full sample, 37-27-30 among registered voters.

Thus, among registered voters, there is a 10% liberal bias.  But Trump is down 12% to Clinton in the uncorrected poll.  Since Democrats outweigh Republicans nationally by at least 1%, this still looks as though Trump is trailing Clinton by 3% even after we correct for bias as best as possible.

The other national poll released Sunday was an NBC News/Wall Street Journal poll showing Trump behind Clinton by 5%.  The best I can tell from the information available about the poll's composition, there was about a 6-7% liberal bias.  Consequently, after correction, this poll looks to be a tie between the two candidates.

These are far cries from national polls released just before Corey Lewandowski was fired, which, when corrected for liberal bias, strongly suggested that Trump was out in front of Clinton by upwards of 10% or more.

Some "other polls" could include state-level races between Trump and Clinton released Sunday in Colorado, Florida, North Carolina, and Wisconsin.  Without correction, these polls all show Trump behind.  Accounting for potential bias, corrected data suggests a statistical tie in each state.

None of this is "good," unless Trump is playing for the national tie and hoping for hanging chads in Florida to make the difference one more time.  At best, the results could be construed as not very bad on their face, but that rosy view starts to fade when we consider that Trump had a healthy lead a week ago in corrected polling data that has now entirely disappeared.

And if Lewandowski's style really was a drag on the campaign, shouldn't his firing, especially when coupled with Trump's supposedly big Clinton attack speech last week and the successful Brexit vote, have given Trump a popularity boost?  It didn't.  In relative terms, Trump lost ground.

Lessons need to be learned, starting with the advice not to dump a winning campaign strategy and let the establishment take over.

Let us also hope that this Twitter account is a parody of Paul Manafort, because if it is the real deal, the Trump campaign is truly is deep trouble.

You can lead a horse to water, but you can't make it drink.

According to Donald Trump on Sunday, "[t]he @ABC poll sample is heavy on Democrats. Very dishonest - why would they do that? Other polls good!"

Well, news flash forthcoming.  The problems with the polling data have been discussed here at AT for at least a couple of weeks, so it is good to see Mr. Trump finally coming around.  The problem is that apparently, based on a number of detailed reports across conservative through liberal websites, Trump used that same massively flawed polling data as a rationale for firing his highly successful campaign manager.

The ABC News/Washington Post poll in question is indeed "heavy on Democrats."  According to the poll's demographic composition:

Partisan divisions are 36-24-33 percent, Democrats-Republicans-independents, in the full sample, 37-27-30 among registered voters.

Thus, among registered voters, there is a 10% liberal bias.  But Trump is down 12% to Clinton in the uncorrected poll.  Since Democrats outweigh Republicans nationally by at least 1%, this still looks as though Trump is trailing Clinton by 3% even after we correct for bias as best as possible.

The other national poll released Sunday was an NBC News/Wall Street Journal poll showing Trump behind Clinton by 5%.  The best I can tell from the information available about the poll's composition, there was about a 6-7% liberal bias.  Consequently, after correction, this poll looks to be a tie between the two candidates.

These are far cries from national polls released just before Corey Lewandowski was fired, which, when corrected for liberal bias, strongly suggested that Trump was out in front of Clinton by upwards of 10% or more.

Some "other polls" could include state-level races between Trump and Clinton released Sunday in Colorado, Florida, North Carolina, and Wisconsin.  Without correction, these polls all show Trump behind.  Accounting for potential bias, corrected data suggests a statistical tie in each state.

None of this is "good," unless Trump is playing for the national tie and hoping for hanging chads in Florida to make the difference one more time.  At best, the results could be construed as not very bad on their face, but that rosy view starts to fade when we consider that Trump had a healthy lead a week ago in corrected polling data that has now entirely disappeared.

And if Lewandowski's style really was a drag on the campaign, shouldn't his firing, especially when coupled with Trump's supposedly big Clinton attack speech last week and the successful Brexit vote, have given Trump a popularity boost?  It didn't.  In relative terms, Trump lost ground.

Lessons need to be learned, starting with the advice not to dump a winning campaign strategy and let the establishment take over.

Let us also hope that this Twitter account is a parody of Paul Manafort, because if it is the real deal, the Trump campaign is truly is deep trouble.