Of course, this was on the heels of some very generous and well timed live interview polling showed Clinton jumping to a big lead right after the Democratic National Convention...
I have read a couple of arguments recently (from these pundits) that the narrowing of the polls has been taking place because of the online polls (which they believe favor Trump) and because of the lack of live interview polls (many of the major outlets only provide this polling monthly). The suggestion was that all of this nonsense about a Trump comeback and a narrowing of the polls would come to a screeching halt as soon as these live phone interview polls started coming out.
Slowly, quietly, but surely these live interview polls have been coming out in the past few days. If I narrow my search back to polls that surveyed within the past 10 days, this is what I have found (four way race):
- FOX: Live Phone - Clinton +2
- IBD/TIPP: Live Phone - Even
- Franklin Pierce: Live Phone - Clinton +3
- GWU/Battleground: Live Phone - Clinton +2
- CNN: Live Phone - Trump +2
- Average of these five live phone polls: Clinton +1
Now I would also expect to see the ABC/WashPost, NBC/WSJ, CBS/NYT, Bloomberg, and the Marist/McClatchy polls very shortly. In fact I would have expected to see some of them already. For instance, ABC/WashPost released portions of a poll last week - but no Presidential horse race results were released? Did they spend the time and effort to get a poll that measured approvals and other typical questions of the candidates, but didn't ask who the survey participants were voting for?
These MSM commissioned polls have been traditionally the most favorable polls for Clinton, which is what I am suspecting Nate Silver and gang to be looking forward to. But if these other pollsters are any indication, it doesn't appear that the live interview polling is the difference. In fact the current live interview polls don't look any different than the robo-call or on-line polling. The last thirteen polls released (four way race) show no leads for either candidate over four points (and that four point lead was an online poll).
What we can state for certain is this: If for some unexplained reason... the other major MSM polls come out with results different than FOX, CNN, TIPP, Battleground, and Franklin Pierce, it will not have anything to do with live phone interview polling.
17 comments:
MSNBC didn't like the results of the CNN poll, so they 'adjusted" the poll but reducing the number of white high-school respondents to give Clinton a 4 point lead.
You can expect the same "adjustment" for the ABC/WashPost, NBC/WSJ, CBS/NYT polls.
I plugged the crosstabs of the CNN poll into my weighted spreadsheet (weighted by an average Demographic over the past several election cycles... and Trump led by three points (not two). For the records I give the Democrats an advantage over Republicans - but I don't see this being in the range of 2008. At best, it may be around where it was in 2012. If it gets back fairly close to where it was in 2004, Clinton is doomed.
There is an assumption on the left that the Demographics for Hillary will be better than they were for Obama (because of an overall change in national demographics). The assumption is that white vote dropped and minority vote went up based purely on demographic changes. They are ignoring the "possibility" that some of that minority voting outreach (eg: Blacks turning out at a number higher than Whites) had to do with President Obama.
I believe this is a dangerous assumption to make. I wouldn't pin my election hopes on Hillary improving the minority vote turnout over what Obama brought out (thus decreasing the White vote turnout percentage significantly more than it was in 2008/2012).
Ch, I understand we just now have 91 presidential polls, including every state in the union. And what do they tell us?
That it looks like we'll be getting a repeat of the 2012 election except that NC will this time go blue.
And that's what a lot of the experts were predicting even before those 91 polls.
Here James - I thought you were a fan of Nate Silver? He uses "all" of the polls, both state and national, and right now... he sees his simulations showing Hillary winning an average of 296 ECV and Trump 241.
And shows the odds at around 2-1.
Suddenly you are questioning Nate Silver?
Seems to me, James... you simply hold onto whatever it is that tells you what you want to hear, and ignore the rest.
Silver's "polls only" forecast gives Hillary NC and 304.3 electoral votes.
Well Obama got 332 electoral college votes in 2012 and 365 in 2008. I guess Silver does not see the election as Hillary being one state better than Obama was. I am guessing he's in a general disagreement with your tweedledee and tweedledumb over at ElectoralVote.com on this?
We've got a long ways to go. A few weeks can be an eternity in politics. Obviously Hillary can no longer "run out the clock" and expect to win. Let's see how she holds up under pressure.
uh...
you forgot the survey monkey / mother jones / think progress / daily kos / occupy democrats!!!11! / talking points memo / MSDNC poll.
Florida (PPP): Trump 44, Clinton 43, Johnson 5, Stein 1, McMullin 1
average Demographic over the past several election cycles... and Trump led by three points (not two).
Your analysis is just more bs, CH. Your previously denying you had spread sheets and your average demographic is whatever you want it to be lets me believe your analysis is just making stuff up to fit your bias. I'm sure you can provide the data and calculations that led you to your conclusion. I won't hold my breath.
Yeah Opie, I can supply the data:
Take a look at the cross tabs of the CNN poll -
http://elections.huffingtonpost.com/pollster/polls/cnn-25340
You will see that Hillary's problem is that she is losing Independents by 20 points. Not that there are too many white working class males (aka: too many Republicans).
Btw: this is a problem for her with Franklin Pierce (down 15) ARG (down 18) TIP (down 12) FOX (down 12) Ipsos (down 14) Morning Consult (down 13) Zogby (down 13)... etc...
To put it in perspective, Romney won independents by 5 (or Obama lost by 5). Clinton has been consistently polling large double digit deficits to Trump with independents.
Yeah Opie, I can supply the data:
Take a look at the cross tabs of the CNN poll -
http://elections.huffingtonpost.com/pollster/polls/cnn-25340
_________________________________________________________________________________________________________________________________________
i don't know why he's having such a hard time with this. the cross-tabs data is self explanatory.
trump is even less unfavorable than hillary: 56% vs. 57%.
it's either sheer stupidity or outright panic. or perhaps a combination of both.
BTW a 20 point lead for Trump among independents is rather devastating for Clinton.
While there's a gender gap for Trump among women it's not as big as Clinton's gender gap among men.
Trump is starting to make inroads with "married women". Still losing single women badly, but is not actually taken a lead with "married women". That's probably what is closing that gap a little.
Still losing single women badly
_________________________________________________________________________________________________________________________________________
and the GOP always will. remember "Julia?"
she was the poster child emblematic of america's single women who need uncle sugar to be their daddy and their husband all rolled into one.
Whites without a college degree appear to make up nearly half of their sample. In 2012, by the way, whites without a college degree was slightly more than a third of all voters,” Todd said. “The point is, your numbers may not be wrong but your weighting may be, your assumptions. So the CNN folks assumed an electorate that is not an impossible scenario for Trump, but it would be an historic shift if it occurred.”
You don't have the ballz to post this since it makes your argument moot!
Well first of all Opie - CNN doesn't provide a breakdown of the "sample" which is why Chuck Todd is suggesting that it "appears" that nearly half their sample is whites without college degrees. That's a guess (at best). Then Todd follows up with the statistic that it was "slightly more than a third" in 2012. Which is also a guess because the actual breakdown of education with race wasn't part of the exit polls in 2012 either.
I can only assume what you might say to describe "my" analysis if instead of using actual numbers I just say "appears that nearly half" and "slightly more than a third". You would call me on pushing b.s. (instead of real numbers) and be correct.
So rather than guess, why not look at what we can analyze without suggesting that anything "appears nearly" or that it is "slightly more" than anything.
The bigger problems are the gaps. Even in the 2014 midterms the Democrats received 34% of Whites without college degrees. The CNN polls shows Hillary garnering 24%. Obama won over 50% of non-college educated and Democrats won 45% in 2014 (when they got clobbered). Clinton is getting 37%.
That's a 10% drop in non-college grads white support from 2014
And a 13% drop overall in non-college grads support from 2012
So let's just focus on the second stat - in 2012 people without college degrees made up 53% of the electorate. Shouldn't Democrats be more concerned that Clinton dropped 13 points with a demographic that makes up 53% of the electorate... and not so much on whether or not that demographic was over sampled in this one poll?
Just curious...
But the bottom line here gang:
There are 63 pages of CNN details on the poll. Most of those pages are various cross tabs. In some of those cross tabs the margin of error is up to 8% (making parsing the cross tabs almost fruitless). I can guarantee you that if you are looking for something "not" to like about the poll, you can find it within the 63 pages.
That being said, everyone's liberal polling expert (Nate Silver) grades the CNN poll as an A- and in his humble opinion, he leaves the +2 at +2 for his adjustment.
Post a Comment