Final Monmouth Poll - November 5, 2012

Likely Voter, 1417 sample size, Tied 48-48, 2% undecided.

Monmouth just released their final poll, and it matches every other poll, showing a tie.  The pollsters really don't want to make a call this election.  It is very similar to the last poll where they showed a 3 point Romney lead.  The D/R/I is D+4 at 35/31/34 which matches the last time this poll was conducted.  The Independent preference also matches the 16 point Romney lead we saw two weeks ago.

Notice however that Obama is getting more crossover support than Romney, which I find to be a little hard to believe.  The VERY odd thing in this poll though can be seen in the following table:

Of the undecideds that made up their minds, ALL of them went to Obama.  This strains credulity, and is the ONLY reason why this poll shows a tie.  Even a pessimistic, but believable 50/50 split would have resulted in a 49.5-46.5 lead for Romney.

Putting these results through the models, we get the following:

Tied - Current result
O+2.0 - 2008 turnout
R+0.9 - D+3 turnout
R+3.3 - 2010 turnout
R+3.4 - 2004 turnout
R+6.7 - Rasmussen Party ID

Like others, this poll does not make a prediction on who will win the race.  However, it indicates the following possible results:

Using a D+3 model Romney wins 50-49
Using an Even turnout Romney wins 51-48
Using an R+5.8 turnout Romney wins 54-47

Gallup

I'm seeing lots of discussion about Gallup down in the comments, so let me address it.

First of all, I am guessing with this poll.  They don't release internals, and I am basing everything off a partisan split that they published when they were taking heat for  showing Romney so far ahead.

The top line number is Romney leading 49-48 with 2% still undecided.  They give no indication of Independent preference.  However, being at 48 is not a good place for Obama to be.  My suspicion is that Gallup has been hit by the same "early voter" problem that all the other polls have.  We never got to see this play out, because we didn't see the daily tracking poll trend down over the last week.  Gallup uses a very tight likely voter screen, but if someone says "I voted" they go right in, regardless of whether it is true or not.

But, let's take the numbers at face value.  Putting this poll into my models using the same D+0.2 partisan split that they were using before, we get the following results.

If turnout is D+3, Obama leads by 0.5%
If turnout matches 2010, Romney leads by 1.8%
If turnout matches the Rasmussen Party ID poll, Romney leads by 5.0%

This poll only trails Pew right now as the best case scenario for Obama, and he still loses the election by almost 2 points if our turnout matches 2010.  And keep in mind, I am using every pessimistic assumption in this calculation, such as Independents have no preference between the candidates.

So quit worrying, and go vote.

STOP THE PRESSES!

I'm stunned.  I really am not sure what to make of these numbers.  If you thought Rasmussen's September poll of partisan ID was beyond belief, well the October numbers are incredible.  To refresh your memory, this poll is conducted monthly and asks 15,000 people what party they are affiliated with.  In my Rasmussen Party ID model, I use a 3 month average of these numbers.

The new October number is way beyond anything I was expecting.  The new numbers are:

Democrat 33.3%
Republican 39.1%
Independent 27.5%

The already amazingly good September number of R+2.6 went to R+5.8!

With the new October number, the 3 month average changes:

Old  -  New
33.8    33.6   Democrat
36.4    37.8   Republican
29.8    28.6   Independent

I will be now using an R+4.2 sample for my Rasmussen Party ID model.  Expect to see that number move much higher.  My analysis below changes to Romney leading by 6.9% in that model.

November 5, 2012 - The Penultimate Analysis

One more day.

Polls included:

Tracking Polls
Even - Rasmussen Daily Track
R+5 - Gallup
O+1.3 - IBD/Tipp
O+1 - ABC/WaPo

Others
R+1 - Battleground
O+3 - Pew
R+1 - NPR
O+1 - CBS/NYT
Even - Fox
O+5 - National Journal
Even - CNN/ORG
Even - NBC/WSJ

O+0.36% - Current RCP Average
O+0.64% - Average using the 2008 turnout model
R+2.27% - Average using the D+3 turnout model
R+4.57% - Average using the 2010 turnout model
R+4.64% - Average using the 2004 turnout model
R+6.39% - Average using the Rasmussen Party ID turnout model

Average Democrat advantage across all polls: D+5.22

The RCP shows a small drift toward Obama on the next to last day, driven my the inclusion of 4 new polls from CNN, Pew, NBC/WSJ, and Battleground, with the average Democrat advantage being driven largely by the D+11 CNN/ORG poll.  I have posted detailed analysis of each of those polls below.  These should all be final polls, and I will not be looking at them again.  Today we should see final polls from most of the other polling firms, and I will provide analysis of those as I am able.

In the tracking polls we continue to have no updates from IBD/Tipp and Gallup.  Drudge teased a Galup results showing Romney up 3 yesterday, but it was removed.  We are expecting them to provide final results today from their final days of polling.  IBD/Tipp I expect to not return, since they were severely affected by Sandy.

Rasmussen continues to show a tie, despite moving to a slightly more Republican baseline (D+2) and Independent preference moving back to a 9 point preference for Romney.  I believe that in the end, Rasmussen will probably be within a point or two, and will not sacrifice his reputation, however he is having some problems with his methodology.  A 9 point lead with Independents in a D+2 poll should be showing a 1 or 2 point Romney lead.  My suspicion is he continues to be plagued by too many Democrats making it past his likely voter screen.  This has been a systemic problem this cycle, which has cropped up especially in the last week while in the middle of the early voting season.  All of the pollsters will need to find a way to correct this problem next cycle if they wish to be accurate.

ABC/WaPo moved out to an Obama +1 lead while maintaining a D+4 sample.  Independent support for Romney increased to 2 points from the Even support they were reporting yesterday.  All in all, ABC/WaPo is looking to be fairly close in its results, once you account for the Democrat over sample.  They are indicating a Romney win of 2.5% if the turnout is Even.  I view this as a relatively conservative result, driven by the low preference for Romney by Independents.  Their over sample also appears to be caused by too many likely voters.  The results of this poll look a lot like a registered voter result.

Since several of the polls are getting very old, and in the case of Tipp may not even be updated, I'm offering an alternative set of results.  In this case I am throwing out the Gallup, Tipp, National Journal, NPR, and CBS/NYT polls because of their age.  I am also throwing out the CNN/ORG poll because it is an obvious outlier that is too good for Romney.  This leaves us with the 6 recent polls, 2 of which are tracking polls.

Polls included:

Tracking Polls
Even - Rasmussen Daily Track
O+1 - ABC/WaPo

Others
R+1 - Battleground
O+3 - Pew
Even - Fox
Even - NBC/WSJ

O+0.50% - Current RCP Average
O+1.60% - Average using the 2008 turnout model
R+1.30% - Average using the D+3 turnout model
R+3.60% - Average using the 2010 turnout model
R+3.66% - Average using the 2004 turnout model
R+4.17% - Average using the Rasmussen Party ID turnout model

This result is looking much more believable.  With an Even turnout, Romney will win by 3 to 4 points.  When Gallup releases their tracking poll I expect them to confirm this with a Romney lead of 3, as was teased yesterday.

Final CNN/Opinion Research Poll - November 4, 2012

Likely Voter, 693 sample size, Tied 49-49, 0% undecided.

Everyone has been waiting all night for me to look at this one.  I figured I should go through the credible polls before this one.

So the race is tied.  Right.

Incredibly, this poll is even more over sampled than the one they did on 10/1.  I'm amazed they published it at all, it is so ridiculous.   The D/R/I of the poll... Oh why don't I just quote directly from the methodology?

Among those likely voters, 41% described themselves as Democrats, 29% described themselves as Independents, and 30% described themselves as Republicans.

You read that right.  D+11.

D+11!!

And RCP just puts it right into the average with all the others.

So how can a D+11 poll show an even race?  Why isn't it showing Obama ahead by 8 or 9, at least?

I'm glad you asked.  Independents favor Romney in this poll by 24 points.  Oh, don't take my word for it, it is right there in the crosstabs:

So for laughs, let's put these numbers into the models:


Even - Current result
R+2.8 - 2008 turnout
R+5.7 - D+3 turnout
R+7.9 - 2010 turnout
R+8.0 - 2004 turnout
R+9.6 - Rasmussen Party ID

When I include this poll into the averages, it will push the numbers up about .25% even with 12 polls.  I will also being doing a version of the averages where I throw out the obvious outliers.  This poll will have to be thrown away, it favors Romney beyond reasonable expectations.

Anyway, this poll is saying that it is too close to call, and the final result will be:

Tied race 49-49

The models instead indicate that this poll is projecting the following:

Using a D+3 model Romney wins 53-47
Using an Even turnout Romney wins 54-46
Using an R+2.6 turnout Romney wins 55-45

Final NBC/WSJ Poll - November 4, 2012

Likely Voter, 1475 sample size, Obama leads 47.5025-47.4975, 3% undecided.

If you haven't noticed my irritation with RCP and poll reporting, the lead numbers above should give you a clue.  The top line numbers are reported as "Obama +1" then you read into the article and find out that

Mr. Obama led his rival by a whisker, 48% to 47%—a difference of seven voters among a pool of 1,475 surveyed.

1% out of a pool of 1475 is 147 voters.  7 out of a pool of 1475 is 0.005%.  Honest reporting of this poll would declare it tied.

As I did with the Battleground poll, I am changing their method of sample determination.  They use a metric of "leaning" Independents, which obviates the concept of partisan turnout.  The last time I looked at this poll the D/R/I was 44/38/18 or D+6.  The following is their sample data:

The proper identification of the partisan ID for this poll is 31/26/41 or D+5.  Additionally they report that Independents favor Romney by 7, however the "leaning" independents favor Romney by 3.  This leads to an overall Independent support of slightly over 2 points (remember that was 7% of a 14% subsample).

When we adjust this result to the different models, we get the following results:

O+0.005 - Current result
O+1.5 - 2008 turnout
R+1.9 - D+3 turnout
R+4.5 - 2010 turnout
R+4.6 - 2004 turnout
R+6.6 - Rasmussen Party ID

This represents about a 1 point shift toward Obama since the last time the poll was run, but Romney retains a comfortable lead in all possible turnout scenarios.

Going through the crosstabs, one omission stands out when looking at specific issues of importance to the electorate:

The voters are not asked who would be better dealing with the economy or jobs.  These two issues are 3 times more important to the voters than all other issues, when asked in other polls.  The fact that the pollster doesn't even ask is telling.

In summary, this poll does not attempt to predict who will win the race.  Leaving the result at 47.5-47.5.  However, using the models, it is predicting the following results:

If turnout is D+3 Romney wins 50.5-48.5
If turnout is Even Romney wins 52.5-48
If turnout is R+2.6 Romney wins 54-47.5

Final Pew Poll - November 4, 2012

Likely Voter, 2709 sample size, Obama leads 48-45, 4% undecided.

Again, RCP chooses to select among two numbers provided by the poll and chooses the one more favorable to Obama.  RCP is reporting Obama leading 50-47, but that number is only acheived by allocating the undecided voters 50/50.  This is not a very good approach, since undecideds at this point are likely to either support Romney or not vote.  It is another example of RCP cherry picking to show a result more favorable to Obama.

The last time this poll was run the D/R/I was 35/34/30 or D+1.  This time they have increased the Democrats in the sample to D+4 with a D/R/I of 36/32/29.  Independents have moved from favoring Romney by 8 to favoring him by 2.

Putting these results through the models, we get the following:

O+3 - Current result
O+4.0 - 2008 turnout
O+1.1 - D+3 turnout
R+1.2 - 2010 turnout
R+1.3 - 2004 turnout
R+3.1 - Rasmussen Party ID

Even with the partisan split, this poll is still showing a shift toward Obama of about 1.5 points.  Having said that, consider the following table included in their release:

Notice that the enthusiasm number for Republicans has increased since the last poll and they maintain a significant advantage over the Democrats.  This leads to the conclusion that a D+4 turnout is very unlikely to occur, and that the Even or better turnout I expect is more likely.

So why is this poll so skewed?  The obvious answer can be found in the following table:

The like voter screen is way to loose, and is certainly driven by the fact that 34% stated they have already voted.  The total of early votes nationwide is barely over 20%.  Whatever the reason, the likely voter screen indicates that we will have an 85.9% participation rate.  Here are the participation rates of all Presidential elections going back to 1988:

Clearly, Pew has too many Likely Voters in their screen.  About 27.7% too many.  History has shown that Likely Voters trend toward Republicans more than Registered Voters.

In the end, this is Pew's final opinion on this race.

Pew's final prediction is Obama winning 50-47.

However, based on the evidence within their poll, I believe this poll properly points to a Romney victory.

Pew's poll supports a prediction of Romney winning between 50.2-49 and 51-48.

Final? Battleground Poll - November 4, 2012

Likely Voter, 1000 sample size, Romney leads 49-48, 3% undecided.

Let's start with the top line number.  The last time this poll was run, one week ago before Sandy, Obama was ahead 49-48.  So this poll has moved 2 in Romney's direction, so much for that Sandy bump.  Now if you look on RCP and the Politico article, you see this listed as tied.  However, look at the following from inside the poll:

By my math that adds up to a 49-48 Romney lead.

Next I want to look at the partisan sample.  This has been a particular problem with this poll, since they push people to identify as partisan due to their "leaning".  They then provide a preference of the very few remaining Independents.  Here is the results of this poll:

In the past I have used the numbers as is, and used the models to reweight.  However, since they have forced Independents into this artificial construct this time, I am putting them back where they belong and using the "lean" to determine Independent preference.  The D/R/I for this poll is really 36/31/30, with Independents favoring Romney by 2 points.  It is a D+5 poll.


Putting these results through the models, we get the following:

R+1 - Current result
R+0.3 - 2008 turnout
R+3.3 - D+3 turnout
R+5.6 - 2010 turnout
R+5.7 - 2004 turnout
R+7.4 - Rasmussen Party ID

This poll indicates a very strong Romney lead by at least 3 points when the votes are counted.  It isn't just me saying that, it is the Terrance Group's opinion as well.  With their final poll, both Terrance Group and Lake offer a final prediction.  Here is the Terrance one:

In sum, this data indicates this election continues to remain very close on the surface, but the political environment and the composition of the likely electorate strongly favor Governor Romney. These factors come into play with our “vote election model” – which takes into account variables like vote intensity, voters who say they are definite in their vote, and demographics like age and education. In that snapshot of today’s vote model, Mitt Romney has majority support (51%). This is an election that will be decided by voters’ focus on the economy, and by turnout and intensity – as we enter the last 72 hours of the campaign, both of those factors favor Mitt Romney.

 Now before you think I am just cherry picking the Republican version, let's also look at the Lake analysis:

In the final days of the campaign, a majority of voters expects and awaits the President’s second term—a metric that has historically had a better record of predicting Presidential electoral outcomes. Cementing voters’ expectations of an Obama victory requires that the President and his Party drive Democratic turnout and maintain their edge among independents. Obama and Democrats have already managed to lock in more support than their Republican counterparts among those who have already voted and have critical advantages on which to build to help propel turnout. Sustaining Democrats’ early vote leads will necessitate they maintain, if not redouble, the strength of their ground game in these final days, particularly in battleground states that remain divided in their support for Obama and Romney. 

"a majority of voters expects and awaits the President’s second term—a metric that has historically had a better record of predicting Presidential electoral outcomes." - What does this even mean?  That because people expect Obama to win, he likely will win?  Why do we bother with elections then?

"President and his Party drive Democratic turnout and maintain their edge among independents." - Didn't the poll already show that Independents favor Romney?  And what about the increased enthusiasm that is visible to us all and mentioned by Terrance Group?

"Democrats have already managed to lock in more support than their Republican counterparts among those who have already voted" - Other polls show this to not be true, and actual votes submitted so far shows the Democrats far behind their goals.

"necessitate they maintain, if not redouble, the strength of their ground game in these final days" - This seems to indicate that the current efforts are insufficient to achieve victory.

To sum up, the Republican pollster from this partnership is offering a specific prediction of Romney's final vote results.  The Democrat is offering platitudes and advice, without making a prediction of Obama winning.

The Battleground final poll predicts Romney will win 51-48.

A couple videos to get you pumped up

When the story of this campaign is written there will be two points where the tide turned and Romney became the 45th president.

The first was in Denver, in the first debate.  That was when we learned that we could trust him, we could believe in him, that he had what it takes to be President.

The second was Red Rocks.  That was where he learned he could believe in us.  That we would be the movement he needs to win.

RDBrewer at Ace reworked the official Red Rocks video with a more uplifting soundtrack

Here is the original

Finally, a new one the campaign released today from the Friday rally in Ohio

Two more days until it is our turn.

Battleground Poll - November 4, 2012

Likely Voter, 1000 sample size, Tied 48-48, 3% undecided.

This is a place holder for a later update once I can find the internals.  I am currently only able to work off the article at Politico, which has picked numbers out of the poll to report, without linking to the full data.

Despite the discussion in the article that "Obama has gotten a boost from Hurricane Sandy", please remember that this poll was Obama +1 last time it was run.  Obama has lost a point of support in the last week. 

They also report that Independents now favor Obama by 1 point, but note that the change is because voters that previously identified as Independent are now reporting as lean Republican.  I've avoided playing with their sample numbers in the past to correct their "lean toward a party" sub-sample, but it looks like I will have to finally, just to get a consistent sample for my models.

Once I can get the internals, I will post the model results.