16 November 2015

2016 polling round-up, 15-30.11.2015: Presidential D vs. R matchups

D vs. R matchups: NATIONAL POLLING


Morning Consult (R), released 17.11.2015:
2,001 RV, MoE = +/-2.0

Clinton 42 / Carson 43, Margin = Carson +1
Clinton 44 / Trump 43, Margin = Clinton +1
Clinton 44 / Bush 37, Margin = Clinton +7
Clinton 44 / Rubio 36, Margin = Clinton +8
Clinton 45 / Fiorina 33, Margin = Clinton +11
Clinton 46 / Paul 34, Margin = Clinton +12

PPP (D), released 19.11.2015:
1,360 LV, MoE = +/-2.7

Clinton 43 / Rubio 45, Margin = Rubio +2
Clinton 45 / Trump 44, Margin = Clinton +1
Clinton 46 / Carson 45, Margin = Clinton +1
Clinton 43 / Bush 41, Margin = Clinton +2
Clinton 46 / Cruz 44, Margin = Clinton +2
Clinton 46 / Fiorina 41, Margin = Clinton +5


Sanders 40 / Fiorina 42, Margin = Fiorina +2
Sanders 39 / Bush 42, Margin = Bush +3
Sanders 41 / Trump 44, Margin = Trump +3
Sanders 39 / Cruz 44, Margin = Cruz +5
Sanders 38 / Rubio 44, Margin = Rubio +6
Sanders 39 / Carson 46, Margin = Carson +7



FOX NEWS poll, released 22.11.2015:
1,016 RV, MoE = +/-3.0

Clinton 42 / Rubio 50, Margin = Rubio +8
Clinton 39 / Bush 45, Margin = Bush +6
Clinton 42 / Carson 47, Margin = Carson +5
Clinton 41 / Trump 46, Margin = Trump +5
Clinton 41 / Cruz 45, Margin = Cruz +4
Clinton 43 / Christie 46, Margin = Christie +3
Clinton 42 / Fiorina 42, Margin = TIE

Sanders 41 / Trump 46, Margin = Trump +5

Once again, this:



In 2008, it was D+8, in 2012, it was D+7 and it has not been R at 40% since 1972.

D vs. R matchups: STATE POLLING

Colorado:

Quinnipiac, released 18.11.2015:
1,262 RV, MoE = +/-2.8

Clinton 36 / Rubio 52, Margin = Rubio +16
Clinton 38 / Carson 52, Margin = Carson +14
Clinton 38 / Cruz 51, Margin = Cruz +13
Clinton 37 / Trump 48, Margin = Trump +11

Sanders 39 / Rubio 52, Margin = Rubio +13
Sanders 40 / Carson 52, Margin = Carson +12
Sanders 42 / Cruz 49, Margin = Cruz +7
Sanders 44 / Trump 46, Margin = Trump +2

I kept looking at these numbers and kept asking myself what could be so possibly off about them, so I checked the internals. This is what Quinnipiac claims is the partisan identification of the poll, from the methodology pdf:




So, Quinnipiac claims an R +7 advantage in this state and also claims that the white, non, hispanic population is 79%.  So, I checked the data.

Here are the very latest VR statistics from Colorado, from October 2015 (.pdf).
Add up the figures, and it comes to this:

Colorado VR 2015 – October


D: 1,074,723 (30.34%)
R: 1,101,253 (31.08%)
I/O/U: 1,366,680 (38.58%)
Total: 3,542,656 (Active, Inactive, New)

Margin: I/O/U +7.50%

Margin D/R: R +0.74% (statistical tie)

So, the partisan identification of the poll is 6 points to the right, compared to the ACTUAL, factual VR statistics for this state.

We go deeper. The EXIT POLL for Colorado 2012 showed:





In November 2012, at the General Election, the partisan breakdown in Colorado was actually D+5.

So, a Quinnnipiac poll showing R+7 now is therefore 12 points away from where the actual electorate in Colorado was in 2012.

Oh, and BTW, the actual VR statistics in CO right before the November election were:

Colorado VR 2012 – End of October

D: 1,150,527 (31.57%)
R: 1,157,083 (31.75%)
I/O/U: 1,336,734 (36.68%)
Total: 3,644,344 (Active, Inactive)

Margin: I/O/U +4.93%

Margin D/R: R +0.18% (statistical tie)

Right before the election in 2012, the R's had a similar very slight advantage over the Ds.

So, a 33 R / 26 D / 33 I model is WAY off.

Let's look at the Latino and White vote component.

Quinnipiac claims that the demographics of the poll are 79% White, 10% Latino.

Ok, let's fact check that as well. As of 2014:




So, the ACTUAL, factual demographics of Colorado are:

69 White / 20 Black. This means that this Quinnipiac, in relation to actual demographics, is 10% too rich on the White vote and 10% too poor on the Latino vote.

Let's also compare that to the actual VR statistics in CO right before the November election, which were:



According to the exit polls, 14% of Colorado's voters in 2012 were Latinos, tendency: rising. There is no reason in the world to think that the Latino vote is going to be less than 14% in 2016.

So, the basis for this Quinnipiac poll is so unbelievably off, for the first time ever, I am forced to call a Quinnipiac poll a "trash" poll. Why they were unwilling to weight the poll for actual, real, likely statistics is just beyond me.

Now, I have been writing for two years now that Hillary has been struggling in Colorado, but there is really no way for her to be losing to Ted Cruz (R-TX) by 13 points while yet strongly beating him in most national polling.

Florida:

Florida Atlantic University (FAU), released 18.11.2015:
829 RV, MoE = +/-3.3

Clinton 41 / Carson 50, Margin = Carson +9
Clinton 41 / Trump 49, Margin = Trump +8
Clinton 43 / Rubio 50, Margin = Rubio +7
Clinton 40 / Bush 44, Margin = Bush +4
Clinton 45 / Cruz 48, Margin = Cruz +3


Iowa:

Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015:
641 RV, MoE = +/-4.0

Clinton 40 / Carson 46, Margin = Carson +6
Clinton 40 / Rubio 43, Margin = Rubio +3
Clinton 40 / Bush 41, Margin = Bush +1
Clinton 41 / Trump 40, Margin = Clinton +1

Within the poll is a party-identification internal that is highly unlikely:


The internal shows an R+7 party affiliation advantage. That is way, way off from reality.

Nevada:

Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015:
628 RV MoE = +/-4.0

Clinton 41 Trump 44, Margin =  Trump +3
Clinton 42 / Rubio 42, Margin = TIE
Clinton 44 / Carson 41, Margin = Clinton +3
Clinton 44 / Bush 38, Margin = Clinton +6




New Hampshire:

FOX News poll, released 19.11.2015:
804 RV, MoE = +/-3.5

Clinton 40 / Rubio 47, Margin = Rubio +7
Clinton 40 / Kasich 43, Margin = Kasich +3
Clinton 42 / Bush 45, Margin = Bush +3
Clinton 43 / Carson 45, Margin = Carson +2
Clinton 41 / Fiorina 43, Margin = TIE
Clinton 44 / Christie 43, Margin = Clinton +1
Clinton 44 / Cruz 41, Margin = Clinton +3
Clinton 47 / Trump 40, Margin = Clinton +7

This internal is interesting and possible, but unlikely:


Political affliation: R+3. In NH, this is indeed possible, but there are many more I's than the 23% listed.


Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015:
530 RV, MoE = +/-4.0

Clinton 42 / Bush 40, Margin = Clinton +2
Clinton 42 / Rubio 40, Margin = Clinton +2
Clinton 43 / Carson 41, Margin = Clinton +2
Clinton 45 / Trump 38, Margin = Clinton +7

The political affiliation internals in this poll are also way, way, way off from reality:


This poll shows an R+9 advantage in political affiliation, way off from reality. And yet, Clinton is still winning.

South Carolina:

PPP (D), released 17.11.2015:
1,290 RV, MoE = +/-2.7

Clinton 39 / Carson 51, Margin = Carson +12
Clinton 41 / Bush 47, Margin = Bush +6
Clinton 42 / Trump 47, Margin = Trump +5
Clinton 42 / Rubio 47, Margin = Rubio +5
Clinton 43 / Huckabee 47, Margin = Huckabee +4
Clinton 41 / Fiorina 45, Margin = Fiorina +4
Clinton 43 / Cruz 46, Margin = Cruz +3
Clinton 41 / Kasich 43, Margin = Kasich +2

Sanders 33 / Carson 51, Margin = Carson +18
Sanders 33 / Rubio 46, Margin = Rubio +13
Sanders 35 / Bush 48, Margin = Bush +13
Sanders 35 / Cruz 45, Margin = Cruz +10
Sanders 38 / Trump 48, Margin = Trump +10

The internals by race are interesting. According to the PPP (D) poll, Dr. Ben Carson gets a considerably higher chunk of the Black Vote in SC than any of his other GOP counterparts.






Morning Consult (R) / Campaign for sustainable RX pricing, released 19.11.2015:
627 RV, MoE = +/-4.0

Clinton 37 / Carson 48, Margin = Carson +11
Clinton 38 / Rubio 46, Margin = Rubio +8
Clinton 39 / Bush 44, Margin = Bush +5
Clinton 41 / Trump 44, Margin = Trump +3


Virginia:

Princeton / UMW, released 16.11.2015:
656 LV, MoE = +/-4.3

Clinton 39 / Carson 44 / Webb 12, Margin = Carson +5
Clinton 42 / Trump 36 / Webb 16, Margin = Clinton +6
Clinton 42 / Trump 27 / Bush 24, Margin = Clinton +15

Sanders 33 / Carson 42 / Webb 17, Margin = Trump +5
Sanders 33 / Trump 38 / Webb 20, Margin = Carson +9

Roanoke, released 19.11.2015:
601 "Residents", MoE = +/-4.0

Clinton 44 / Carson 44, Margin = TIE
Clinton 45 / Rubio 41, Margin = Clinton +4
Clinton 46 / Bush 39, Margin = Clinton +7
Clinton 46 / Fiorina 39, Margin = Clinton +7
Clinton 47 / Cruz 39, Margin = Clinton +8
Clinton 50 / Trump 36, Margin = Clinton +14

Wisconsin:

Marquette University, released 20.11.2015:
803 RV, MoE = +/-4.2

Clinton 43.7 / Rubio 45.2, Margin = Rubio +1.5
Clinton 44.2 / Carson 45.3, Margin = Carson +1.1
Clinton 47.5 / Trump 38.3, Margin = Clinton +9.8


Sanders 46.0 / Rubio 41.9, Margin = Sanders +4.1
Sanders 47.1 / Carson 41.1, Margin = Sanders +6.0
Sanders 51.9 / Trump 35.3, Margin = Sanders +16.6



No comments:

Post a Comment

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