20 September 2012

09/19 was a record setting day in polling for 2012 / Battleground outlook

On Wednesday, 09/19, 17 state polling results for a total of 12 states were released, plus one composite "swing state" poll, bringing the total to 18 individual polls released on one day:
09/19: CO, VA (3), WI (2), MO, CA, „Swing States“, MA (2), ME (2), MI, NH, OH, FL, GA09/19: CO, VA (3), WI (2), MO, CA, „Swing States“, MA (2), ME (2), MI, NH, OH, FL, GA

That is more polls on any one day thus far in the entire election season and a sure sign that things are truly heating up going into election day.

On the two days before, 14 state polling results for a total of 12 states were also released:
09/16: KY, VA
09/17: WI, IN, MA, OR
09/18: FL, MI, PA, MA (2), NH, VA, CO
The one new state in the mix that had not been polled at all for 2012 was Kentucky.

Because some key states have been polled multiple times in the last 4 days, I would like to take a look at them and their averages. Not all of these states are battlegrounds, but it is still important to see what is going on in "core" territory for both parties.

5 Polling results for MASSACHUSETTS have been released  within 3 days (09/17-09/19). This is mostly because of the contested race between challenger Elizabeth Warren (D) and incumbent Scott Brown (R) in contention for a seat that Teddy Kennedy (D) held for almost 47 years. But as a side-result of the polling, the presidential race has also been polled, and the numbers show that President Obama is closing in on his winning margin from 2008:



19UN Mass / Lowell / Herald19.09.12497 LV+/-5.55936523
18MassINC / WBUR19.09.12512 LV+/-4.459311028
17WNEC (WNEU)18.09.12444 LV+/-4.66038222
16PPP (D)18.09.12876 LV+/-3.35739618
15Suffolk17.09.12600 LV+/-4.06431733
14Kimball Consulting ( R )10.09.12756 LV+/-3.55640716

As you can see from the "EVERYTHING TABLE", Obama won MA in 2008 by +25.81% over John McCain. But his end-polling average for the Bay State in 2008 was +21.33%, so Obama outperformed his polling average in MA by +4.48% in '08.

Just one week ago, before all this new polling, Obama's average in MA was +16.00%, it is now +23.33%. This means that at this current time, Obama's polling average in MA is slightly above his end polling from 2008. There is absolutely no guarantee that he will again outperform the polls in 2012, but it is telling indeed when five completely independent-from-each-other pollsters all have Obama with crushing margins. It is also very telling when PPP (D), the pollster that many on the Right like to accuse of being too "democratic-friendly", is showing a result very close to Kimball Consulting's result. PPP's results are the second smallest of the bunch, and Kimball is a Republican polling firm. No one has ever really thought that MA would ever be in contention, but since Mitt Romney is the former Governor of Massachusetts and it is therefore considered his second "home" state, then a sign of a Republican wave nationally would be lowered Democratic margins in this state. But that is not happening here.


Four polling results for Virginia have been released from 09/18-09/19, plus there was both a PPP and a Rasmussen poll from a couple of days before, so we have 6 very fresh polls of the Old Dominion to look at. Here is the average, going back two weeks:



37WAA19.09.121238 LV+/-2.848,5045,705,802,80
36Fox19.09.121006 LV+/-3.0494297
35Quinnipiac / NYT / CBS19.09.121474 LV+/-3.0504644
34WAPO18.09.121104 RV+/-4.0524448
33PPP (D)16.09.121021 LV+/-3.1514635
32Rasmussen14.09.12500 LV+/-4.5494831
31NBC / WSJ / Marist13.09.12996 LV+/-3.1494475
30Gravis (R)11.09.122238 LV+/-2.244,0449,396,57-5,35

Note that the Gravis poll from 09/11 will soon fall out of the statistic, and it is the only poll since before 08/21 to show Romney ahead. When this poll leaves the statistic, then Obama's average, which is currently +3.43%, will rise. We can see from the EVERYTHING TABLE that in 2008, Obama won VA by +6.30% and his end polling average was +4.93%. This means that he also outperformed his polling average in VA, by  +1.37%. It also means that at this time Obama is 1.50% UNDER his polling average from 2008, but +3.43% is a vast improvement since Electoral Landscape number 3, from 09/15. A lot has changed in this state in just 5 days. This means that Virginia has already left the tossup category but is still very much a battleground state. That could, however, change again.


There were two polls released for the state of MAINE on 09/19. ME is now considered core DEM territory, but ME, like NE, does elector splitting, and ME's Democratic voting history is a short one: until 1992 it was considered a very, very Republican state. 



8bMPRC – CD 219.09.12993 RV+/-3.1147,8041,0011,206,80
8aMPRC – CD119.09.12993 RV+/-3.1158,7033,907,4024,80
8MPRC19.09.12856 RV+/-3.3553,5037,309,2016,20
7PPP (D)19.09.12804 LV+/-3.55539616

Note that these two pollsters, completely independent from each other, have practically identical margins for President Obama in the Pine Tree State. His average is +16.10%. We see from the EVERYTHING TABLE that Obama won ME by +17.32% in '08, his end polling average was +17.15%. This means that while his average and end results for 2008 were practically indentical to each other, he is currently 1.05% UNDER his polling average from '08, which is minimal erosion for him, but the MPRC poll also puts out it's results by congressional district and as we can see, ME-02 is safe DEM in 2012, just as it was in 2008. THAT is the data from ME that is so important at this time.


In 2008, I made a strong argument that New Hampshire, sandwiched in-between VT (current polling for Obama: +37%) and ME (current polling for Obama: +16.10%), that the blue "pull" must be very hard to overcome in a state that has practically the same demographics as both VT and ME.

And in most of the polling for NH, this appears to be so, but limited. Apart from one Rasmussen that came out 09/19, showing Romney ahead by +3 in the "Granite State":
Nr.NEW HAMPSHIREDateSampleMoEObamaRomneyUnd.Mar.


16Rasmussen19.09.12500 LV+/-4.545487-3
15ARG18.09.12463 LV+/-3.2484662
14WMUR / UNH13.09.12592 LV+/-3.94540155

This Rasmussen poll moves NH back into the tossups. In Electoral Landscape number 3, I wrote of NH:

More importantly, in NH polling, the President is now hitting 50%, an important mark. In the case of a very close election, however, it is still conceivable that Governor Romney can overtake the lead in New Hampshire once again."

So, for now, NH is defying my argument from 2008 and the shift in voter registration in this state, which I will publish in a couple of days, shows why this is happening. In other words, NH is demonstrably moving back to a lean win state as it was in both 2000 and 2004. Flip a coin...

There have been three polls released for WISCONSIN (09/17, 09/19) and the results are very clear:


33Marquette19.09.12705 RV+/-4.15440614
32Quinnipiac / NYT / CBS19.09.121485 LV+/-3.0514546
31PPP (D) / DFA17.09.12959 LV+/-3.16494831

WI has moved demonstrably out of the battleground territory. Please also note that once again, PPP (D), whom the Right loves to say is biased, is showing the leanest margin of the three new polls. The Marquette Poll is considered the "Gold Standard" for Wisconsin, just as Selzer is for Iowa and Indiana.
We see from the EVERYTHING TABLE that Obama won WI by +13.90% in '08, his end polling margin was +11.34%, which means he outperformed the end polling in WI by +2.56%. Obama is currently demonstrable UNDER the end polling for 2008, by -4.34%, so Paul Ryan (R- VP nominee, WI) is making a difference here, but probably not enough to make the state truly competitive for the GOP. I expect Rasmussen, Mason-Dixon, Quinnipiac and maybe Marist to put out a Wisconsin poll soon. CNN / ORC just put out a Michigan poll, so maybe they will also poll the Badger State. Wait and see. At any rate, after practically a month without any fresh data for WI, we now have three fresh post-convention polls, and that is data worth looking at.


Ohio has only been polled once since Electoral Landscape number 3, but the FOX NEWS poll just released is identical in margin to the NBC margin from a week ago:


36Fox19.09.121009 LV+/-3.1494297
35NBC / WSJ / Marist13.09.12979 LV+/-3.1504377
34ARG13.09.12600 LV+/-4.0484751
36Rasmussen13.09.12500 LV+/-4.5474671
35Gravis (R)10.09.121548 LV+/-3.047,2743,199,524,08
34PPP (D)09.09.121072 LV+/-3.0504555

It is easy for partisans to claim that one pollster is "in the tank" for a certain candidate, but when two completely different polling organizations - one considered left leaning, the other, right leaning - produce identical margins, then there is certainly some validation in the polling. Not only that, we can see a verifiable shift over a week: on 09/13, both polls had the race at +1 for the President. The last two both have him at +7. And the oldest two in the two-week average have him at +4 and +5 respectively.  How does a +4.18% margin compare?
We see from the EVERYTHING TABLE that Obama won OH by +4.58% in '08, his end polling margin was +2.30%, which means he outperformed the end polling in OH by +2.28%. His current +4.18% margin is 1.80% above the end polling from '08. There can be no doubt that President Obama is more solid in the Buckeye State than he was in 2008, and this statistic must the the most alarming signal to the Romney campaign - for they cannot win the White House without Ohio, truly, they cannot.


There have been two new polls published for Florida, keeping the state just as "tossupy" as ever:


45Fox News19.09.12829 LV+/-3.0494475
44Gravis Marketing (R)18.09.121728 LV+/-2.547,1047,705,20-0,6
43NBC News / WSJ / Marist13.09.12980 LV+/-3.1494475
42Concensus Communications (R)13.09.12606 LV+/-4.04242160
41McLaughlin (R) for AIF13.09.12

40Rasmussen13.09.12500 LV+/-4.5484662
39SUSA11.09.12596 LV+/-4.1484484

What to make of this? Well, we have two Republican pollsters calling it for Romney by +3 and +0.6, one Republican pollster as a tie in the very low fourties, with 16% undecided- which is practically impossible this late in the game- plus two Republican leaning pollsters who have Obama ahead. The one pollster considered the most neutral of all of these, SUSA, has it at Obama +4, but that poll will soon be old news.
We see from the EVERYTHING TABLE that Obama won FL by +2.81% in '08, his end polling margin was +1.79%, which means he outperformed the end polling in FL by +1.02%. This also means that his current polling average of +1.77% is practically identical to his end polling from 2008, in spite of the fact that four of the six polls in the current average are from decidedly Republican pollsters. This should be the second warning sign to the Romney campaign - that they have not moved ahead in FL following the convention in Tampa. What the numbers DO show is a similar resiliency for the incumbent as they showed for Pres. Bush in the 2004 campaign.


Colorado is showing signs that it may even be a closer race than Florida at the end of the day: there are two new polls from the "Million Dollar Highway" State (09/18-09-19), plus 3 from 09/13, all pretty fresh polling, and the state looks very, very competitive:


21Quinnipiac / NYT / CBS19.09.121497 LV+/-3.0484751
20Rasmussen18.09.12500 LV+/-4.545478-2
19Denver Post / SUSA14.09.12615 LV+/-4.0474671
18Project New America13.09.12503 LV+/-4.4494475
17ARG13.09.12600 LV+/-4.0494742

This keeps CO in the tossup category and the Rasmussen +2 projection for Romney is every bit as feasible as the Quinnipiac +1 for Obama.

In Part III
 of my long, long, long analysis of Colorado from 2009, I graphed out the statistics for the pick-up counties and the tipping-point counties in this state. If the election becomes close so that Colorado's 9 EV really look as if they will be the EV to make the difference, then seeing which candidate visits which counties will tell us alot about what their respective teams think about their chances of winning:

If we see Obama spending just tons of time in Arapahoe and  Jefferson counties, both big counties with large military installations in them, and traditionally very RED counties, then you know that the Obama team really thinks it is going to retain CO. If Romney campaigns there but also hits Garfield county (which was the closest county in the entire USA in 2008, McCain by 2 votes. Yes, 2 votes) and campaigns in Denver, then you will know that the Romney team thinks it can lock up CO, for the only way for him to win the state is to turn around the situation in Arapahoe County, sweep a little county like Garfield county and try to keep Obama's margin down in Denver (similar story in Ohio in Cuyahoga County). Also, Broomfield County, just created before the 2004 election, which went for Kerry by a hair in '04 and Obama by a landslide in '08, will be interesting to watch, for this county was carved out by REPUBLICANS at the time and they were sure that it would be like the Silicon Valley of CO and be very red. Well, it didn't turn out that way. But Broomfield has a very upscale demographic, one that Romney should be able to attract.

One last state: California. No one is expecting Romney to win the "Golden State", but the goal would be to hold Obama's feet to the fire and to keep the margin down. But this is not happening:


17Field19.09.12891 LV+/-3.45834824
16CBRT / Pepperdine13.09.12900 LV?+/-4.055,4033,4011,2022,00
15SUSA12.09.12524 LV+/-4.35735822
14USC / LA Times23.08.121041 LV+/-3.055,6032,7011,7022,90

The fourth poll on this list is NOT in the average - it is too old, but notice how consistent the margins have been. Currently, it is Obama +22.67% in California, with it's 55 EV.
We see from the EVERYTHING TABLE that Obama won CA by +24.03% in '08, his end polling margin was +24.00%, which means his '08 result was IDENTICAL to his end polling. This also means that his current polling average of +22.67% is 1.36% UNDER his end polling average from '08. That is statistically small and will not change the outcome of the race in this state, but it is very slight erosion for Obama. What IS important about all of this is that this +24.03% margin from 2008 meant a raw vote margin of 3,262,692, which was larger than Bush's entire national margin of 3,012,171 from 2004. What does that mean? It means that in the case of a very close national election in 2012, crushing margins in both CA and NY pretty much guarantee that Obama will win in the NPV. Also, once again, a truly close national election would also mean that Romney should be gaining everywhere, even some in California. Were Romney way ahead right now, we should be seeing demonstrably reduced margins in California, but we are not.


There is some good news for Mitt Romney in this sea of (mainly) blue polling numbers over the last 3-4 days:

In Missouri, his average is now +8.75% over Obama. It looks like Romney has a lock on the state.

Also, a new Insider Advantage poll of Georgia has Romney up by +19. Insider Advantage is the "Gold Standard" for the Peach State and it's number from 2008 were spot on for that race. This would be a vast improvement over McCain's +5.20% margin in '08 and even larger than Bush's win in '04, if the margin holds.

The first and up till now, only poll of Kentucky shows Romney ahead by +14%
We see from the EVERYTHING TABLE that McCain won KY by +16.22% in '08, his end polling margin was +12.40%, which means that McCain outperformed his end polling average by +3.82. With only one poll, it is impossible to build a viable polling "gene-pool", but a +14 margin in the "Bluegrass State" is a good landslide margin and also has ramifications for neighboring TN, which as a very similar demographic to KY and was won by a similar margin by McCain in 2008. The older polling of TN is way too stale now and I suspect that Romney will easily win that state on 2012.


FACIT: the last 4 days of polling, making a total of 32 state polls, points to the race moving DEMONSTRABLY in Obama's direction.

No comments:

Post a Comment

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