Pages

Monday, August 29, 2016

Trump gains five points in Monmouth poll

Been waiting for this poll (and the Marist) to come in...  as it had shown a twelve point Hillary lead three weeks ago. The results still show a healthy seven point lead for Clinton, but the outlier has sort of moved somewhat back in line.

Clinton lost support among Democrats (from 92% down to 85%)... as questions about her emails and the Clinton Foundation continue to hound her. She dropped four points overall from 50% in Early August. Trump gained two points from early August, moving from 37% to 39%.

The poll sampled more Republicans (239) than Democrats (226).  It appears that Republicans were weighed down by 2%, while Democrats were weighted up by 4%. The poll sample also included 78% Non-Hispanic-White voters, but was weighed down to 71%

Because of the change in this poll (from twelve points to seven)... Trump is officially under 5 points in the overall average and down to four points in the Projection. Haven't seen that for some time.

(Note: It is the opinion of some that a Presidential Candidate cannot make up five points in 10 weeks, much less five points in three weeks. Meaning there must be something fishy going on). 

8 comments:

James said...

Ch, I hear Trump is in BIG trouble with Catholics who make up about one fourth of the voters. His feuding with the Pope hasn't gone over well with them.

Romney got about 6% of the Catholic vote, but Hillary is presently polling at pulling in 23 to 27% of the Catholic vote.

Doomsday for Trump?

James said...

I guess the fact that Nate Silver (538) now gives Trump an astronomically high 21.2% chance of winning brings warmth to the cockles of Ch's heart.

C.H. Truth said...

James - since I have historically been more accurate than Nate Silver (by a long shot) - what he has to say generally has little bearing on my feelings towards a race.

But I will say this... as WP pointed out (someone else from a math and science background) - Nate Silver has lost all semblance of an objective statistician by literally "adjusting" the survey results based on his own bias.

I say bias, because that is (by definition) what it is. There is no way to know at this point which polls will be accurate and which polls will be wrong. There is no objective manner to determine that polling is biased in favor or Trump (as Silver suggests). Especially considering how much recent polling has been the opposite (2014 had a significant bias to the left).

It's purely his best guess, based on a hunch, or based on his own viewing of where he believes the race "should" be.

I thought he would have learned his lesson:

http://fivethirtyeight.com/features/how-i-acted-like-a-pundit-and-screwed-up-on-donald-trump/

But he is basically doing the exact same thing. Whether it is a repeat of his earlier admitted mistake, time will tell. But it's hard to imagine how someone would objectively look back and see themselves as overly subjective about someone in the Primary (downplaying his polling). Admit it was a mistake. Then turn around and assume 70% of all polling is wrongly showing too much Trump support in the General election as well.

James said...

I guess you're referring to Silver's BIG mistake of thinking Trump would not be the nominee.

Any reader of this blog can go to Wikipedia and look up Nate Silver and get a more accurate sense of his track record than you will ever admit.

C.H. Truth said...

No James - what Silver is stating is that he used subjective measures to adjust the polling (in the primary) against Trump. He admitted to playing pundit (rather than being objective).

He's doing exactly the same thing in the general election by adjusting 70% of the polling in Hillary's favor. In some cases by upwards of 4-6 percent.

He was wrong to do so in the Primary. I offer it is still wrong to do so in the General.

James said...

He admitted a mistake. Why would he turn around an repeat another mistake that he will/would have to admit?

James said...

But Trump was an anomaly that practically EVERYONE got wrong.

Again, go to Wiki and look at Silver's presidential election track record, his electoral votes track record, his Senate and House predictions, etc.

C.H. Truth said...

James - I have been following Silver before the NY Times picked up his 538 website. He got extra attention because he was already a fairly well known baseball statistician.

He got famous as a political pundit for predicting 49 of 50 states in the 2008 Election, an election where almost every prognosticator got 49 states or all 50 states correct. In fact, I don't know any of the main projection sites that got less than 49 wrong (other than probably ElectoralVote.com).

In 2010 he missed three Senate Seats alone in one year. All of them incorrectly predicting Democrats to win races they lost. He was off (also favoring the Democrats) by nine seats in the House. This was a disaster, and was worse than any other prognosticator that year.

In 2012 he got all 50 states correct. However, with the exception of Florida (and possibly Ohio) there was not a single state that was really in doubt. Since he always picks the Democrats in close races, and Obama won both Florida and Ohio, he got those right. But then again, so did almost all of the projection websites in 2012. It simply was not a tough election to judge. In 2014 he missed two Senate Seats, incorrectly predicting that two more Republicans would lose (Pat Roberts and Thom Tillis).

Bottom line: He's not that good. And he predictably leans left in his projections. Always has, and probably always will.