Wednesday, October 31, 2012

Obama / Romney Ceilings in Virginia using past turnout

Gov. Bob McDonnell and Michele Bachmann join Mitt Romney on stage at a rally in Portsmouth, Virginia last May. Photo courtesy of Mark Makela/Reuters
Quinnipiac University has released their final round of swing state polling in Virginia, Ohio, and Florida, and the top lines aren't very pretty for Mitt Romney. Specifically, in the state of Virginia, Obama leads Romney by two points, 49-47%. But a glance at the internals should raise some eye brows. First of all, how does Obama manage to beat Romney in the state of Virginia while losing Independent voters by a 57-36% margin, or 21 points? Simple: by finding a Virginia electorate that is even MORE Democratic than the electorate that voted in 2008. Quinnipiac's party ID finding for their final Virginia poll is 35% Democrat, 27% Republican, 35% Independent, or D+8. In 2008, turnout was D+6 (39D/33R/27I).

Who are we to argue with poll findings? But I would propose that Quinnipiac's party ID findings certainly represent Obama's ceiling in Virginia, and probably Romney's floor.

What would Mitt Romney's ceiling look like in Virginia, using the Quinnipiac poll findings? Given that no statewide election was held in Virginia in 2010, I'll look to the 2009 Governor's race and the 2004 Presidential race for a example of Mitt Romney's ceiling in Virginia. In both years, Virginia turnout was R+4.
Here is what the Quinnipiac poll would look like reweighted to 2009/2004 Virginia turnout.

Quinnipiac poll weighted to 2009 Virginia turnout:

Party ID        Obama                    Romney

33D              31.68  (96%)            0.99    (3%)
37R               2.22    (6%)             34.41  (93%)
30I                10.8    (36%)           17.1    (57%)
                      44.7%                        52.5%

Friday, October 26, 2012

Another D+9 party ID finding in Ohio, this time from ARG

Large Romney rally in Cleveland, Ohio, Nov. 4, 2012. Photo found at
What's with all the polls out of Ohio showing a D+9 electorate? Earlier this week, Time/CNN released a poll showing Obama leading in th Buckeye state by 5 points, with a D+9 sample. Today, the American Research Group releases an Ohio poll showing Obama leading Romney 49-47%, with likely voters self-identifying as 43% Democrat, 34% Republican, and 23% Independent. That's significantly MORE Democratic than the Ohio electorate in the wave election year of 2008, when turnout was D+5. Even more startling than the ARG party ID is Romney's numbers with Independents: he's leading Obama by 21 points yet still trails by 2.

Assuming ARG's party ID finding is NOT an accurate representation of the Ohio electorate this November, what would the ARG poll look like reweighted to 2008 partisan ID?

ARG poll reweighted to 2008 Ohio turnout:

Party ID        Obama                    Romney

37.5D            32.625 (87%)         3.375     (9%)
32.5R            3.25     (10%)          28.6      (88%)
30I                10.8     (36%)          17.1      (57%)
                     46.675%                  49.075%

If Obama is fortunate enough to have the 2012 electorate resemble the 2008 electorate, Romney would LEAD the Ohio ARG poll 49-47%, a reversal of Obama's actual lead in the poll.

In 2010, a strong Republican year, the Ohio electorate was TIED between Democrats and Republicans at 36D/36R/28I. Here is what the ARG poll would look like if Romney is able to bring out Republicans in full force, as in 2010:

Wednesday, October 24, 2012

CNN/Time confirms Obama's Ohio lead, but is it really 5 points?

Obama campaign kick-off in Columbus, Ohio, May 5, 2012. Photo courtesy of Scott Olson/Getty Images

The CNN/Time poll released this afternoon showing Obama ahead 49-44% also finds Ohio likely voters identifying as Democrats at a historically high net rate when compared to past exit poll numbers. Per the poll's findings, 37% of respondents were Democrats, 28% Republican, and 29% Independent, or D+9. Just for fun, lets reweight the poll to 2008 turnout (D+5), per exit polls and Josh Jordan of NRO.

Time/CNN poll reweighted to 2008 Ohio turnout per CNN exit polls:

Party ID        Obama                    Romney

37.5D            34.5    (92%)           2.25      (6%)
32.5R            3.575   (11%)          27.625  (85%)
30I                11.4     (38%)          15.9      (53%)
                       49.475%                   45.775%     

Tuesday, October 23, 2012

Romney +2 with a survey sample similar to 2008 turnout

PPP tweet from Sep. 30, 2012.
PPP (D) recently joined the swarm of pollsters releasing daily presidential tracking polls, and today's numbers showed Romney moving ahead of Obama among likely voters nationally, 49-47%. Obama and Romney were tied on Monday, 48-48%. A quick glance at the poll's internals reveal that Romney's 2 point national lead comes despite poll respondents having a strong Democratic lean (D+6, 41D/35R/24I%). That's only one point less Democratic than national turnout in 2008, according to CNN Exit polls.

So how does Romney get to a 2 point lead when Democrats outnumber Republicans by roughly the same margin as 2008? By winning Independents (50-41%), and by keeping Obama at a surprisingly low number with his own base (81%). Romney captures 87% of Republicans, as well as 16% of Democrats. Obama only attracts 11% of Republicans.

Since most can agree that D+6 is on the high end of the spectrum in terms of Democratic turnout, here's a look at the PPP survey reweighted to a likely best-case turnout scenario for Romney: 37% Dem, 37% Rep, 26% Ind - the turnout numbers from the 2004 election, per CNN.
PPP daily tracker weighted to 2004 CNN exit poll party ID:

Party ID        Obama                    Romney

37D               29.97   (81%)          5.92    (16%)
37R               4.07     (11%)          32.19   (87%)
26I                10.4   (40%)            13.26   (51%)
                      44.44%                     51.37%

Monday, October 22, 2012

New Hampshire: UNH/WMUR poll shows Big Obama lead and Big Democratic sample

President Obama addresses a Manchester, New Hampshire crowd with home-state Senator Jeanne Shaheen (D) and Gov. John Lynch (D). Photo found here.
University of New Hampshire / WMUR poll, taken October 17-21st, 2012 of 773 likely New Hampshire voters. MoE: 3.5%.

Obama - 51%  (+1 from last poll)
Romney - 42%   (-2 from last poll)

Partisan identification of likely voters:
Democrat - 46.1%
Republican - 38.9%
Independent - 15.0%

Among Democrats:
Obama - 94%
Romney - 2%

Among Republicans:
Obama - 4%
Romney - 91%

Among Independents:
Obama - 39%
Romney - 37%

UNH / WMUR poll weighted to 2008 New Hampshire exit poll party ID:

Party ID        Obama                    Romney

29D               27.26   (94%)          0.58     (2%)
27R               1.08     (4%)            24.57   (91%)
45I                17.55   (39%)          16.65   (37%)
                      45.89%                     41.8% 

UNH/ WMUR poll weighted to 2004 New Hampshire exit poll party ID: 

Party ID        Obama                    Romney

25D               23.5   (94%)          0.5     (2%)
32R               1.28     (4%)            29.12   (91%)
44I                17.16   (39%)          16.28   (37%)
                      41.94%                     45.9%           

OHIO: Obama leads Romney 49-47% if Quinnipiac Poll is re-weighted to 2008 turnout numbers, leads 51-45% w/ 2004 turnout.

Rural Ohio, a region of the state Obama-Biden lost to McCain-Palin in 2008, 54-43%. Photo courtesy of
Quinnipiac is out with a new poll of Ohio showing Obama with a strong 5 point lead, 50-45%. There are two potential bright spots for Romney, however: 1) Despite Obama's strong lead in the state, it is only half as strong as it was in Quinnipiac's pre-debate poll. Then, Obama led Romney 53-43%. 2) Quinnipiac finds likely voters identifying as 35% Democrat, 26% Republican, and 34% Independent, for a party ID of D+9. That finding indicates that Democratic turnout in 2012 would be near double the turnout in 2008, a wave Democratic year in which Democrats shattered turnout records. Party ID in 2008 was D+5. What happens to the Quinnipiac poll when it is reweighted to 2008 turnout?

Quinnipiac poll weighted to 2008 Ohio turnout:

Party ID        Obama                    Romney

37.5D            34.875 (93%)          1.875    (5%)
32.5R            1.95     (6%)            30.225  (93%)
30I                12.6     (42%)           14.7     (49%)
                       49.425%                  46.8%

If the Quinnipiac party ID finding of D+9 happens to be incorrect this November, and turnout reverts back to 2008 levels, Obama would lead Romney by just 49-47% in the Q-poll. Again, that is the case simply by reweighting to 2008, which will be no easy feat for the President the second time around. On the other end, here is what happens to the Q-poll if you re-weight to 2004 Ohio turnout numbers, when the electorate was  R+5 (35D/40R/25I)?  

Quinnipiac poll weighted to 2004 Ohio turnout:

Party ID        Obama                    Romney

35D            32.55  (93%)            1.75    (5%)
40R            2.4      (6%)              37.2    (93%)
25I             10.5   (42%)             12.25   (49%)
                   45.45%                     51.2%

Sunday, October 21, 2012

PPP IOWA poll better news for Romney than it appears.

Obama campaigns at a wind energy farm in Oskaloosa Iowa, Aug. 14, 2012. Photo courtesy of Carolyn Kaster, Associated Press.
PPP is out with a brand new poll of Iowa showing Obama leading Romney by 1 point, 49-48%. More fascinating than the top line itself is the fact that PPP finds Iowa likely voters identifying as 41% Democrat, 34% Republican, and 25% Independent, or D+7. This means Iowa voters are currently identifying as Democrats in a much higher proportion than they did four years ago, when Democratic enthusiasm and fervor greatly outmatched Republicans.

Why, then, is the Iowa electorate looking so significantly more Democratic in a year when most polling suggests enthusiasm lies with the GOP? The answer is far from clear, but a good argument can be made that D+7 in Iowa this November is unlikely. Look no further than 2008 Iowa exit polls, when party ID was 34D/33R/33I, or D+1. And that was in a strong Democratic year. In 2004, when George W. Bush won nationally by 2.5 points and carried Iowa 49.9 - 49.2%, turnout was R+2 (34D/36R/30I). Also, while midterms are not the best barometer for presidential turnout, the 2010 Iowa Governor's race saw an R+4 electorate (31D/35R/34I). So unless Democratic enthusiasm spikes well above 2008 levels, the D+7 electorate PPP finds in Iowa seems very unlikely.

What would the PPP poll look like if it were re-weighted to 2008 CNN exit polling?

PPP poll weighted to 2008 Iowa turnout:
Party ID        Obama                    Romney

34D              29.24  (86%)            4.42    (13%)
33R              3.3      (10%)             28.71  (87%)
33I               14.19   (43%)            16.83   (51%)
                      46.73%                    49.96%

Now you can see why this poll is better news for Mitt Romney than it might appear. If Democrats are able to repeat their 2008 turnout performance in Iowa, Romney would lead the PPP poll 50-47%. On the other end of the spectrum, assume November turnout resembles the 2010 Iowa electorate (R+4). What happens to the PPP poll when re-weighted to GOP-friendly turnout numbers?

PPP poll weighted to 2010 Iowa turnout (Governor's race):

Wednesday, October 17, 2012

SurveyUSA shows Ohio is no "firewall" for Obama. Romney behind only 1 pt in 2008-scenario

Paul Ryan, his wife and mother campaign in Owensville, Ohio with Governor John Kasich (R), Sen. Rob Portman, and his wife, on Sep. 12, 2012. Photo found here.
Survey USA is out with a new poll showing Obama leading Romney in Ohio, the President's "firewall," by a 45-42% margin. 3% choose someone other than Obama or Romney, and 9% are undecided. Poll findings indicate that 39% of Ohio likely voters self-identify as Democrats, 32% identify as Republican, and 26% identify as Independent. Both Obama and Romney are attracting equal portions of their base, while Romney captures Independent voters by 8 points, 41-33%. As a result, it's easy to see that we're dealing with a Democratic friendly poll sample, given that Romney leads strongly with Independents, yet still trails the President overall. In the Democratic year of 2008, Ohio partisan identification was D+8 according to CNN exit polls. But poll-nerd/analyst @NumbersMuncher points out that the actual Ohio partisan ID in 2008 was closer to D+5, or 37.5D/32.5R/30I. He explains it here:

Monday, October 15, 2012

Romney performing well in Pennsylvania in 2008-style electorate

Obama, Romney at the Alfred E. Smith dinner, Oct. 18, 2012. Photo courtesy of AP Photo/Charles Dharapak
PPP just released a Pennsylvania poll showing Obama leading Romney by 7 points, 51-44%. The President led 52-40% last month. PPP also finds 48% of Pennsylvania likely voters identifying as Democrats, 37% identifying as Republican, and 15% identifying as Independent. If that finding is correct, then Pa. voters are identifying as Democrats (D+11) by the highest margin of the last 4 national elections (D+3 in 2010, D+7 in 2008, D+5 in 2006, D+2 in 2004).

Putting aside that 48% of likely Pa. voters are currently identifying as Democrats, what happens to the PPP poll if election day turnout looks more like 2008, when Democrats broke turnout records throughout the country (44D/37R/18I)?

PPP poll weighted to 2008 CNN Pennsylvania exit poll:
Party ID        Obama                    Romney

44D               34.32 (78%)             7.04   (16%)
37R               5.18  (14%)               29.97 (81%)
18I                9.18    (51%)            7.38   (41%)
                       48.68%                       44.39%

Obama's 7 point lead becomes a 49-44% lead.

Now, assume Republicans get lucky in Pa. and have an election night that resembles 2010 turnout (40D/37R/23I). Here's how the PPP poll would look:   

PPP poll weighted to 2010 CBS Pa. Gov. exit poll:

Party ID        Obama                    Romney

40D               31.2 (78%)              6.4     (16%)
37R               5.18  (14%)             29.97 (81%)
23I                11.73 (51%)             9.43   (41%)
                       48.11%                     45.8%

ARG Virginia poll is better news for Romney than it appears

The American Research Group just released a Virginia poll showing Romney leading the President 48-47%. ARG finds likely voters self-identifying as 40% Democrat, 35% Republican, and 25% Independent (D+5). Those party ID numbers closely mirror Virginia's D+6 party ID from 2008.

But what happens to the ARG poll if Virginia turnout more closely resembles that of the 2009 Governor's race? (Virginia had no major statewide race in 2010). Partisan identification in November of 2009 was 33D/37R/30I, or R+4. Statewide turnout in 2004 was also R+4.

ARG poll weighted to 2009 CNN Va. Gov exit poll:

Party ID        Obama                    Romney

33D               29.7 (90%)              2.31   (7%)
37R               1.11  (3%)               33.67 (91%)
30I                12     (40%)             15.6   (52%)
                         42.81%                   51.58%  

UNSKEWING The Washington Post / ABC poll

WashingtonPost / ABC is out with a new poll showing Obama leading Romney nationwide by 3 points, 49-46%, among likely voters. Nothing seems too peculiar about the top line, other than the fact that it's the first national poll to show Obama with a modest lead post Denver-debate. But what is peculiar is the surprisingly high number of respondents identifying as Democrats as compared to Republicans. The poll's partisan identification is 35% Democrat, 26% Republican, and 33% Independent, or D+9.

If turnout actually reflects the Washington Post poll's findings, then we can expect to see the most Democratic electorate in modern history this November, even more so than the record breaking Democratic turnout we saw just 4 years ago (which was D+7).

In light of this difficult to trust finding, let's "unskew" the poll, and reweight it to 2010 and 2004 turnout (which was TIED among the two major political parties). To be specific, 2010 party ID broke down like this: 35D/35R/30I. Here's what the ABC numbers would look like reweighted to this party ID:

ABC/WashingtonPost poll weighted to 2010 CNN national exit poll:

Party ID        Obama                    Romney

35D              31.85 (91%)            2.8    (8%)
35R               2.45   (7%)             32.55 (93%)
30I                12.6  (42%)             14.4  (48%)
                       46.9%                    49.75%  

Monday, October 8, 2012

Pew result has pollsters scratching heads, Republicans ecstatic, and Democrats livid.

Pew Research just released a poll taken Oct. 4-7th (entirely AFTER the Denver debate) showing Mitt Romney surging into a 49-45% lead among likely voters, and a 46-46% TIE among registered voters. The top line seems even more miraculous when you consider where the race stood at the time of Pew's last national poll on Sept. 12-16 (Obama led Romney 51-43% among likely voters, and 51-42% among registered voters).

Even beyond the top line result, the Pew poll is receiving a lot of attention for their internal finding that likely voters nationwide identified as 36% Republican, 31% Democrat, and 30% Independent, or R+5. This number may seem unrealistic to a lot of exit poll watchers, given that it indicates an electorate 5 points more Republican than 2004 AND 2010 turnout, both thought to be strong Republican years. Though, to be fair, Pew's September poll found an electorate that was 39D/29R/30I, or D+10. In other words, Pew found a NET 15 POINT SWING in partisan identification over the course of just three weeks.

For the sake of argument, assume for a moment that turnout in November looks more like 2008 than Pew's post-debate findings. In that case, the Pew poll results would look like this:

Pew poll weighted to 2008 CNN national exit poll:

Party ID        Obama                    Romney

39D               36.66 (94%)           1.95   (5%)
32R               2.24   (7%)             29.12 (91%)
29I                12.18 (42%)           13.34  (46%)
                       51.08%                   44.41%

Pair of Rasmussen swing-state polls show that despite debate victory, Obama maintains small lead.

The poll results for Mitt Romney have been a bit mixed in the aftermath of the Denver debate. Surveys are nearly unanimous in their agreement that Romney won the debate, and in a big way. But polling since last Wednesday has sent mixed signals regarding the size of his debate bounce. After an initial bump on Rasmussen of about 4 points, the two are tied again today amongst leaners (48-48%). The President was ahead 50-45% before the debate, per Gallup, but is tied in today's 3 day rolling average.  PPP has released two state polls (in WI and Va) that show Romney gaining on the President, though still behind.

And then you have Rasmussen's Colorado and Iowa polls out today. The president leads Romney in both states, 49-48% in the former, 49-47% in the latter. The last Rasmussen Colorado poll from September 17th showed Romney ahead of Obama, 47-45%, while the September 19th Iowa poll found Romney ahead 47-44%. If you buy Rasmussen's numbers, the poll results indicate Romney has actually done worse in Colorado and Iowa since the debate.

Sunday, October 7, 2012

Post Denver Debate, Virginia remains in Obama's favor, per PPP (50-47%).

Despite saying overwhelmingly that Mitt Romney won the debate last Wednesday (61-28%), Virginia continues to give President Obama the lead by a 50-47% margin. That's a slight improvement from PPP's last poll in mid-September that showed the President ahead by 5 points, 51-46%.

The poll finds likely voters in Virginia identifying as 38% Democrat, 32% Republican, and 30% Independent (D+6). That matches the Democrats turnout advantage in the state in 2008, which was measured  at 39D/33R/27I. But in 2004, CNN exit polls found a much more Republican electorate than 2008 (35D/39R/26I).  And even more recently, in the 2009 Governor's race, the Virginia electorate was much more reminiscent of 2004 than 2008 (33D/37R/30I, or R+4).

So what would the PPP poll have looked like had they weighted their partisan sample to reflect the 2009 Virginia Governor's race exit poll numbers?

Monday, October 1, 2012

Despite survey respondents being 8pts more Democratic than GOP, Obama only leads 50-47% in CNN poll

The new CNN poll becomes the fourth of the day to show the Obama/Romney race w/in 3 points. CNN has Obama ahead among likely voters 50-47% (ARG has Obama at 49-46%, ABC/WashPo has Obama at 49-47%, and Politico/Battleground has Obama at 49-47% - not to mention today's Rasmussen Tracker has Obama ahead 50-47%, and Gallup shows a 49-45% Obama lead). In other words, it seems like a pretty safe bet to say that Obama is heading into the first debate with a 3 point lead, and is awfully close to 50% nationally.

Like the PPP NC poll released last night, and the ARG NC poll released this morning, a large Democratic advantage in partisan identification neutralizes Mitt Romney's strong lead with Independent voters (49-41%). CNN's survey finds voters identifying as 37% Democrat, 29% Republican, and 34% Independent (D+8), the same net Democratic partisan advantage from 2008. But if the electorate in November resembles the 2010 national electorate, the CNN numbers flip from a 3 point Obama advantage to a 4 point Romney advantage (50-46%):

In contrast to nearly every pollster since the DNC, today's Politico / Battleground poll finds electorate closer to 2010 than 2008

Today's  POLITICO/George Washington University Battleground Poll of likely voters has Barack Obama leading Mitt Romney 49-47%. The poll was taken September 24-27.

Among Democrats, Obama wins 94% to Mitt Romney's 5%. Among Republicans, Romney wins 91% to Obama's 5%. Among Independents, Romney leads 47-43%. Partisan identification for the poll is closer to 2010 party ID than 2008, at D+2 (43D/41R/15I). If turnout looks closer to 2008, as most polls have found since the DNC, the Battleground poll would look like this:

Obama leads Romney nationally 49-46% in survey that finds a D+6 electorate.

ARG just released their national poll, and it finds Mitt Romney and Paul Ryan dropping one point from the previous ARG poll 10 days ago. They now trail Obama-Biden 49-46%. The poll was taken September 27-30.

Much like their NC poll, ARG finds that likely voter party ID is very similar to 2008 turnout (38D/32R/30I; D+6). Here are the ARG numbers, re-weighted to 2008's D+7 partisan identification.

ARG has Romney leading 50-46% in NC, despite finding a D+10 electorate

The American Research Group is out with a new poll of North Carolina likely voters, and the news is good for Mitt Romney, relatively speaking. In a survey that was conducted Sep 28-30, Mitt Romney leads Barack Obama by 4 points, 50-46%. But like the PPP NC poll released last night, a closer look at the numbers reveals even more good news for Romney: Mitt leads Obama by 4 points and hits 50% in a survey that finds NC likely voters self identifying as 43% Democrat, 33% Republican, and 24% Independent (D+10). That's an awfully Democratic electorate for Mitt to be leading by four points. 2008 CNN exit polls in North Carolina showed party ID to be D+11 in a year of historic Democratic turnout. Still, Obama carried the state by just 0.4%. Here is what the ARG poll would look like if the partisan identification were re-weighted to match 2008 NC turnout.

Even with a D+14 electorate, Obama and Romney are TIED w/ NC LVs, 48-48%. Romney leads by 18 w/ Indys.

Public Policy Polling finds Barack Obama and Mitt Romney tied among North Carolina likely voters in their new poll taken Sept 27-30, 2012. Obama led 49-48% in their poll from earlier this month. But a look at the internals sheds light on why NC is still such a tough state for Barack Obama to carry.

The President can only manage a tie with Mitt Romney, despite the fact that PPP finds the NC likely voter pool to be 14 points MORE Democratic than Republican. In 2008, when Barack Obama carried the state by less than half a point, Democrats outnumbered Republican voters by just 11 points, for a party ID of 42D/31R/27I. PPP, however, finds the NC electorate to be 48% Democrat, 34% Republican, and 19% Independent, or D+14.