01 December 2015

Polling wrap-up, 16-30.11.2015

With Thanksgiving firmly in the rear-view mirror, it's time to recap the latter-half of November, 2015. Precisely because of Thanksgiving, there was considerably less polling in the 4th week of the month, so we are looking at more like a solid 10-day window at the end of November.  For this reason, I am just doing one large set of aggregate numbers (averages) instead of dividing the national polling into "week 3" and "week 4". There was also some very interesting state pollling.


GOP NOMINATION: NATIONAL POLLING


1.) Morning Consult (R), released 17.11.2015: Trump +19
2.) IPSOS / Reuters, released 19.11.2015Trump +19 (RRV + IRV, R-leaning), Trump +23 (RRV only), Trump +18 (IRV, R-leaning only)
3.) PPP (D), released 19.11.2015: Trump +7
4.) Bloomberg / Selzer, released 19.11.2015: Trump +4
5.) NBC News / Survey Monkey, released 20.11.2015: Trump +10
6.) ABC / WAPO, released 21.11.2015: Trump +10
7.) FOX NEWS poll, released 22.11.2015: Trump +10

-------------------------------------------------------------------
PRRI (Public Religion Relations Institute), released 23.11.2015: Margin: TIE (Rs), Carson +2 (RRV), Carson +7 (RLV)
------------------------------------------------------------------

8.) The Economist / YouGov poll, released 25.11.2015: Trump +22
9,) Gravis (R) / OANN, released 25.11.2015: Trump +22
10,) IPSOS/Reuters, released 25.11.2015Trump +24 (RRV + IRV, R-leaning), Trump +26 (RRV only), Trump +18 (IRV, R-leaning only)
x.)  IPSOS/Reuters, released 27.11.2015: Trump +16.1


The three IPSOS/Reuters polls need some explaining. Generally, IPSOS/Reuters does what is called 5-day rollling polling, which means that the firm is polling every day and logs all of the results HERE. If you click on a link related to recent polling values, it will take you to an interactive graph, where you can see the toplines, published in this kind of form:



That's the graph from the poll that IPSOS/Reuters released, surprisingly, on 27.11.2015. But every ten days or so, like clockwork, they release what they call their "Core approval polling" results, also based on rolling polling. So, why IPSOS/Reuters decided to release two national polls right next to each other like this is a mystery to me. It sure seems very partisan to me, for IPSOS/Reuters is trying to say that Trump's support has drastically dropped. Well, no, it has not. One day, he peaked at 43%, one or two days later, he was at 32%. But that's the ENTIRE reason for rollling polling: to smooth out the rough edges from day to day and therefore reduce the margin of error. For this reason, I have listed the last poll as "x.)" and not given it a number. The general numbers from the 25.11.2015 Core Approval poll are much more consistent.

Also, the PRRI poll has no number, not because Carson led in it, but because it is only of religious people, in this case, religious Republicans, and therefore not representative of the entire GOP electorate. Including this poll in the calculations would be like an apples to oranges comparison and simply not right.

Now, these polls were input chronologically, even daily, which means that a poll that came in earlier in the day is higher on the list than a poll that came in later on that same day. Yes, I am that picky.

We see some amazing consistency in these national polls.

First off, we see two polls right next to each other from two incredibly different-from-each-other pollsters (Morning Consult - a Republican firm, and IPSOS - a French firm), showing an identical Trump +19 value. IPSOS publishes three values, as you can see, but I still think that value no. 1 is the most representative of the overall GOP electorate. The same thing applies, of course, to the overall Democratic electorate.

Then, we see two polls (PPP, Bloomberg/Selzer) showing Trump in single digits. Those polls both came in later in the day on 19.11.2015.

Then, we see three polls in a row (NBC / ABC / FOX - three major news outlets), all showing exactly Trump +10. I think we all know how often those three news outlets can produce wildly differing results, so the consistency here is truly a factor worth noticing.

Then, after the weekend and right before the Thanksiving holiday, from three very independent-from-each-other pollsters (YouGov, Gravis-R, Ipsos), his margin more than doubled to +22-24.

Removing the first and last IPSOS leaves us with nine national polls to average in a release-date time-frame of 19-25.11.2015 (6 days), making for an aggregate of:

Trump +14.22

There is no doubt that in terms of national polling, Donald Trump is still very firmly in the lead. People can make all the arguments that they want that much of this polling, when asked, also shows that the highest % of the lot would never vote for Trump, but in a crowded field, that means nothing. With just 30%, Trump can win the nomination as long as the field is full. Mitt Romney was literally capped at around 23% on the aggregate for 2 solid years and he still won the GOP nomination in 2012.  I see no statistical sign that Trump is fading.

However, we do see that both Rubio and Cruz are rising. In fact, Cruz, although he is more often than not still in third place, is rising faster than Rubio. Also, in all of the national polling, Jeb Bush is mired at between 4-6%.

You can see all of the toplines and lots of internal screenshots HERE.


GOP NOMINATION: STATE POLLING

Colorado:
Quinnipiac, released 18.11.2015: Carson +6 (over Rubio, Trump in 3rd)

Florida:
Florida Atlantic University (FAU), released 18.11.2015: Trump +18

Iowa:
Morning Consult / Campaign for sustainable RX pricing, released 19.11.2015: Trump +12
CBS Battleground Poll / YouGov, released 22.11.2015: Trump +9 (over Cruz, Carson in 3rd)
Quinnipiac, released 24.11.2015: Trump +2 (over Cruz - statistical tie)
Iowa State Poll, released 26.11.2015 (conducted 01-15.11.2015): Carson +10.5 (over Rubio, Trump in 4th)

There is a problem with the Iowa State poll. Notice the very large time frame in which it was conducted, and then they waited 9 full days to publish. Usually, a poll is conducted over a maximum of 3 days if it is an individual poll, and over more than 5 for the 5 day rolling effect. But a two-week polling time window is just too much.

With the Iowa State poll in the aggregate, it's: Trump +3.13
And without the Iowa State poll in the aggregat, it's: Trump +7.67

Either way, Trump still leads in Iowa.

Massachusetts:
Suffolk University, released 24.11.2015: Trump +14.2 (over Rubio)

Nevada:
Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: Trump +20

New Hampshire:
FOX News poll, released 19.11.2015: Trump +14
Suffolk Univ. / Boston Globe, released 21.11.2015: Full poll HERE: Trump +11*
Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: Trump +18
CBS Battleground Poll / YouGov, released 22.11.2015: Trump +19

*The Suffolk poll also did a theoretical matchup with Mitt Romney in the mix. In that case, Romney won by +16 over Trump. Only, it's moot since the filing date for the New Hampshire primaries has already passed.

New Hampshire aggretate: Trump +15.5

Trump is solidly, verifiably ahead in New Hampshire.


New Jersey:
Farleigh-Dickinson, released 17.11.2015: Trump +13

Oklahoma: (yes, Oklahoma!)
Sooner Poll / News on 6 / News 9, released 20.11.2015: Trump +8.8 (over Cruz)

Considering that Oklahoma polls are about are rare as snowstorms in the Sahara Desert, it's good to see one from the gold-standard for that state: Sooner.

South Carolina:
Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: Trump +2
CBS Battleground Poll / YouGov, released 22.11.2015: Trump +16

South Carolina aggregate: Trump +9

Virginia:
Princeton / UMW, released 16.11.2015: Carson +5

Wisconsin:
Marquette University, released 20.11.2015: Carson +2.7 (statistical tie)


That makes for 18 state polls out of 11 states, of which Trump easily won 14. And in the four first-in-the-nation states (IA, NH, SC, NV), he is ahead. There are rays of sunshine in all of this for Carson, in Colorado, Virginia and Wisconsin. But the two candidates who are moving upward the most are Cruz and Rubio. Now, whether this is "flavor of the month" stuff or a long-term shift, only time will tell.

But the long and short of all of this is that Donald Trump is firmly in the driver's seat at both the national and the state level. If these numbers hold, then Trump is currently cruising toward the GOP nomination. And with just 60 days now before the Iowa Caucuses, it is soon going to be "make or break" time for a number of other candidates.


You can see all of the toplines and lots of internal screenshots HERE.




DEM NOMINATION: NATIONAL POLLING


1.) Morning Consult (R), released 17.11.2015: Clinton +31
2.) IPSOS / Reuters, released 19.11.2015Clinton +12 (DRV +IRV, D-leaning), Clinton +21 (DRV only), Sanders +2 (IRV, D-leaning only)
3.) PPP (D), released 19.11.2015: Clinton +33
4.) Bloomberg / Selzer, released 20.11.2015: Clinton +25
5.) NBC News / Survey Monkey, released 20.11.2015: Clinton +16
6.) ABC / WAPO, released 21.11.2015: Clinton +26
7.) FOX NEWS poll, released 22.11.2015: Clinton +23

-----------------------------------------------------------------------
PRRI (Public Religion Relations Institute), released 23.11.2015Clinton +39 (Ds), Clinton +36 (DRV), Clinton +33 (DLV)
-----------------------------------------------------------------------

8.) The Economist / YouGov poll, released 25.11.2015: Clinton +20
9.) Gravis (R) / OANN, released 25.11.2015: Clinton +27
10.) IPSOS/Reuters, released 25.11.2015Clinton +20 (DRV +IRV, D-leaning), Clinton +28 (DRV only), Clinton +20 (IRV, D-leaning only)


The PPRI poll is separated from the rest and not in the aggregate for the same reason that I listed for the Republicans (apples to oranges comparison). Removing the first IPSOS/Reuters, which would now be a repeated, then the Democratic national aggregate is:

Clinton +24.56

That is an enormous landslide margin over Bernie Sanders nationally and in 7 of those 9 polls, Clinton is at +20 or above. This kind of margin is what I have sometimes called the "Vice-President" margin, for we have seen these kinds of aggregates before (although I don't have all the numbers at hand and am relying on memory, so be kind to me, here): Mondale in late 1983 over the Democratic field, Bush 41 in late 1987 over the Republican field, Gore in late 1999 over the Democratic field (Gore was actually doing better than this). My sense is very much that Democrats have quite willingly settled on Hillary Clinton as their nominee and see her position as former Secretary of State in a similar vein to that of a Vice-President. Now, whether this brings good or bad luck in the GE, only time will tell, because history has been mostly very unkind to Vice-Presidents who have tried for the presidency on their own after serving VP. Wait and see. There is no doubt that Hillary is way in front, and also in state polling. See:


DEM NOMINATION: STATE POLLING

Arizona: 
BRC / Rocky Mountain Poll, released 19.11.2015: Clinton +28

Colorado:
Quinnipiac, released 18.11.2015: Clinton +27

Florida:
Florida Atlantic University (FAU), released 18.11.2015: Clinton +44
(the delegate haul from Florida to the DNC is enormous)

Iowa:
Morning Consult / Campaign for sustainable RX pricing, released 19.11.2015: Clinton +22
CBS Battleground Poll /YouGov, released 23.11.2015: Clinton +5
Quinnipiac, released 25.11.2015: Clinton +9
Iowa State Poll, released 26.11.2015 (conducted 01-15.11.2015): Clinton +21.7

I already listed the reasons for why the Iowa State poll is so problematic above, among the Republican numbers. That same reason exists here. The Iowa aggregate, with the Iowa State poll, is:

Clinton +14.43

And without the Iowa State poll, the aggregate is:

Clinton +12

Either way, Clinton is in landslide terrority over Sanders in Iowa. Clinton lost Iowa to both Obama and Edwards in 2008. So, these numbers are good news for the Clinton team in overcoming ghosts of the past.

Massachusetts:
Suffolk University, released 24.11.2015: Clinton +24.5

Nevada:
Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: Clinton +29

New Hampshire:
FOX News poll, released 19.11.2015: Sanders +1 (statistical tie)
Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: Clinton +4 (statistical tie)
CBS Battleground Poll /YouGov, released 23.11.2015: Sanders +7

New Hampshire aggregate: Sanders +1.33 (statistical tie)

New Jersey:
Farleigh-Dickinson, released 17.11.2015: Clinton +36

Oklahoma: (yes, Oklahoma!)
Sooner Poll / News on 6 / News 9, released 20.11.2015: Clinton +34.4

Considering that Oklahoma polls are as rare as getting heat-stroke at the North Poll, it's good to see a poll from Oklahoma's gold-standard, Sooner. There are tons of undecideds in this poll, but that doesn't mean they are leaning toward Sanders. In 2008, without ever stepping a foot in the state, Clinton trounced Obama by +23.6 here. Oklahoma is a sure bet for Clinton, but brings very few delegates to the Convention, being such a deep red state nationally.

South Carolina:

Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: Clinton +44
CBS Battleground Poll /YouGov, released 23.11.2015: Clinton +47

South Carolina aggregate: Clinton +45.5

Virginia:
Princeton / UMW, released 16.11.2015: Clinton +36

Wisconsin:
Marquette University, released 20.11.2015: Clinton +8.6

That makes for 18 polls from 12 states, of which Clinton has won 16 and only 4 of those 16 are margins under +10. In fact, 9 of those margins are at close to +30 (absolute blowout margin) and above.

Of the four first-in-the-nation states (IA, NH, SC, NV), Clinton is well ahead in 3 of those four, while New Hampshire is a statistical tie.

Facit:  numerically, it is looking quite grim for Vermont Senator Bernie Sanders. Hillary Clinton is coasting practically everywhere.


The complete toplines and lots of internal data for all of the Democratic polling can be found HERE.


D vs. R matchups: NATIONAL POLLING


Morning Consult (R), released 17.11.2015: 6 Clinton vs. GOP matchups, she wins 5 of 6, her losing margin to Carson is drastically reduced over the last MC poll.

PPP (D), released 19.11.2015: 12 matchups, 6 Clinton vs. GOP and 6 Sanders vs. GOP. Sanders loses all six. Clinton wins 5 of 6, but most of the margins are lean and within the MoE. She does the best against Fiorina (+12), who has slipped all over the place in polling.

FOX NEWS poll, released 22.11.2015: 8 matchups, 7 of them Hillary vs. GOP, one Sanders vs. Trump. There is one tie, the GOP wins the rest. The FOX poll reflects a voter screen that shows partisan affliation at R+1, with Republicans being 40% of the overall electorate. This is wishful thinking. This is the reason why the FOX poll is so radically different from the rest. (Screenshot at the big link below)

D vs. R matchups: STATE POLLING

Colorado:
Quinnipiac, released 18.11.2015: 8 matchups, 4 Clinton vs. GOP, 4 Sanders vs. GOP. According to Quinnipiac, both Democrats lose very badly to the GOP in this state. I did an extended analysis of the poll at the big link below. There are lots of mathematical reasons for why this poll is a major outlier.

Florida:
Florida Atlantic University (FAU), released 18.11.2015: 5 Clinton vs. GOP matchups. She loses all 5.

Iowa:
Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: 4 Clinton vs. GOP matchups, she wins 1, loses 3, everything but Clinton vs. Carson is within the MoE. I also did an extended analysis of this poll as well. The voter screen is also not in line with reality. The analysis is worth the read, believe me.

Nevada:
Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: 4 Clinton vs. GOP matchups: 1 tie, Trump wins, Clinton wins the other two.


New Hampshire:
FOX News poll, released 19.11.2015: 8 Clinton vs. GOP matchups: 1 tie, 3 Clinton wins, 4 GOP wins, most within the MoE. The partisan identification of the poll (R+3) is indeed possible, but the I component is far too small.

Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: 4 Clinton vs. GOP matchups, she wins all 4, most within the MoE. And that is with a voter screen (partisan identification) that is far removed from reality. See: the big link below.

South Carolina:

PPP (D), released 17.11.2015: 13 matchups, 8 Clinton vs. GOP, 5 Sanders vs. GOP. The GOP wins all matchups, but only 1 in double digits over Hillary. We do not expect a Democrat to win this state, one that Obama lost by -9 in 2008 and by -11 in 2012. And the much leaner margins for the other GOPers also mean nothing. In the one or two polls (Obama vs. Romney) in 2011/2012, Romney was up by only 6 points. So, the results here are unsurprising and there are no tea-leaves to be read from it. At the big link below, you will see that I screenshotted and marked the black vote statistics. Trump, who claims to be doing so well among Black voters, gets 7% of the black vote in the Palmetto State, pretty much right in line with electoral history for this state. Carson, however, gets 14% of the AA vote against Clinton and 19% of the AA vote against Sanders. Were Carson to somehow win the GOP nomination, these numbers could perhaps also play out nationally. Wait and see.

Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015: 4 Clinton vs. GOP matchups. The GOP wins all four, 1 within the MoE, 3 outside of the MoE. Carson does the best, Trump does the worst. Again, these results are unsurprising.

But the next poll is:

Virginia:

Princeton / UMW, released 16.11.2015: 3 Clinton vs. GOP vs either Webb or Trump as (I), 2 of the same sort for Sanders. In  a three-way, Clinton loses to Carson by 6 but beats Trump as an R by 6 and Trump as an I by 16. Sanders loses both matchups. There is really no way to quantify this, as the Princeton poll is the first of it's kind to put out these kind of three way matchups with Webb on board as an Independent.

Roanoke, released 19.11.2015: 6 matchups: one tie, Clinton wins the other 5 and does the best against Trump (+14). Roanoke was the pollster who claimed on October 31st, 2012, that Romney was ahead of Obama by +6 (Obama won by +4), so Roanoke was way off in 2012, to the Right, just to note. In fact, it was the most off of all of the end-pollsters for this state. So, I don't think anyone can accuse Roanoke of being liberally biased.

Wisconsin:
Marquette University, released 20.11.2015: 6 matchups, 3 Hillary vs. GOP, 3 Sanders vs. GOP.  Hillary loses to Rubio and Carson by -1 but beats Trump by +10. Sanders wins all three, and trumps Trump by +17. According to Marquette (which had a mathematical bias of 1 point to the LEFT in 2012 - within acceptable parameters), Sanders does better here than Clinton.


The long and short of the state polling is less about Clinton and more about Carson. Carson, who is losing in national polling to Trump in terms of the GOP nomination, generally does considerably better than Trump in D-vs-R polling, both at the state and national level. So, the guy less likely to win the GOP nomination is more likely to have a chance to flip some blue-wall states, were he to get the GOP nomination. This is a paradigm in polling we have seen for about 6 weeks now, maybe a little longer.

You can see all the D-vs-R polling toplines and tons of internals (including screenshots) HERE.


SENATORIAL / GUBERNATORIAL polling

The big link is HERE (toplines, margins, some interesting internals).

The big event was of course the Louisiana Runoff, where Democrat Jon Bel Edwards essentially ended Sen. David Vitter's career, and his landslide +12.22% winning margin was very much in line with the end polling aggregate, which showed (the final four polls): Edwards +10.5. So, where the claim can be made that polling was off in Kentucky for the Gubernatorial in early November, it was right on the money for the LA-runoff election, no doubt about it.

Also at the link, you will find polling from Arizona, Maryland and South Carolina, none of it all that surprising.



And finally, presidential approval/disapproval, some pol approval/disapproval and some issues polling, the big link for which is HERE.

This is a category that I am not too sure I will continue. Sometimes, there is just too much data and it is hard to know what is really relevant and what is not. Plus, there are tons of aggregators out there that do the presidential approval numbers.

There were, however, some interesting numbers about the immigration debate and the debate over accepting or not accepting Syrian refugees.

Because of the long awaited Christmas/New Years holidays, it is more likely that December will look like a three week month in terms of polling.

However, with the next Republican presidential debate scheduled for December 15th and the next Democratic presidential debate set for December 19th, I suspect that we will see a huge flurry of polling between today, December 1, 2015 and December 14th.

Ok, those were the polling numbers for the second half of November, 2015, realizing that 2016 is now just 31 days away.

No comments:

Post a Comment

Constructive comments and critique are always welcome. Please keep it polite and respectful.