WaPo/ABC Poll: Obama 49-Romney 46

[Adam Griffin]

More evidence of the Obama field operation...

You must be logged in to read this quote.

I'm guessing the margin of error would be pretty high on this for either side?  Wouldn't it consist of smaller state by state subsamples?

If you attempt to draw lessons from a particular state with this sample, then yes. When comparing as an aggregate, though, I think it would be as accurate as any other variable in the poll.

[Cliffy]

I guess it's really kind of a duh stat, it's 9% higher in Dems and a 10% spread. 

More evidence of the Obama field operation...

You must be logged in to read this quote.

I'm guessing the margin of error would be pretty high on this for either side?  Wouldn't it consist of smaller state by state subsamples?

If you attempt to draw lessons from a particular state with this sample, then yes. When comparing as an aggregate, though, I think it would be as accurate as any other variable in the poll.

[The Vorlon]

If you throw this poll into the mix, we get, give or take fractions of a %, a tied race.


Partisan divisions in this survey, Democrats-Republicans-independents, are 33-22-37 percent
among the general population, 34-25-36 percent among registered voters and 35-26-33 percent
among likely voters.

Partisan divisions in the 2008 exit poll were 39-32-29 percent

The whole issue of partisan "weighting" is more complex than it is made out to be by many.

Firstly, a great many folks who "on the first pass" identify as independents will, if pushed, say they lean to one party or the other. 

This poll did not "push" the independents for a soft party affiliation.  Republicans "typically" pick up a few % on the party ID question if you "push" the independents a bit.

My general rule of thumb is to look at the exit polls for 2004 and 2008.

In 2004 it was pretty much a dead heat GOP to Dems, in 2008 the Dems had about a 7% advantage.

This 0 to D+7 range seems to me to be the "rational" playing field.  Clearly the nation is less enthusiastic than the were in 2008 for Obama, but on the other side of the coin, historically the GOP and Dems being even is also pretty rare. - I think this common sense approach pretty much sets the boundries of a reasonable party ID.

[wan]

I guess it's really kind of a duh stat, it's 9% higher in Dems and a 10% spread. 

More evidence of the Obama field operation...

You must be logged in to read this quote.

I'm guessing the margin of error would be pretty high on this for either side?  Wouldn't it consist of smaller state by state subsamples?

If you attempt to draw lessons from a particular state with this sample, then yes. When comparing as an aggregate, though, I think it would be as accurate as any other variable in the poll.

I guess you forgot your not perfect either didn't you embarrass yourself by posting an old and not accurate Florida poll and then admitted it was a mistake.

[Cliffy]

Umm no I didn't embarass myself, the poll was posted with a different name.  Keep reaching though and stalking obviously.

[Eraserhead]

It's nice to see a non-tracking national poll for a change. Hopefully there will be more to come in the next day or two.

[The Vorlon]

It's nice to see a non-tracking national poll for a change. Hopefully there will be more to come in the next day or two.

I suspect everybody and their dog will have one out in the next two days as pre-debate snapshots.

[Cliffy]

Weren't we at 0 in 10'? Wasn't Walker supposed to be recalled?   I thought 04 we were R+2 or 3?    

I'd say D+3 is realistic, above that not so much, I highly doubt it will be above 0 for the Dems but I look at it as D+3.  The GOP is not staying home this time and you don't have hope and change.    

If you throw this poll into the mix, we get, give or take fractions of a %, a tied race.


Partisan divisions in this survey, Democrats-Republicans-independents, are 33-22-37 percent
among the general population, 34-25-36 percent among registered voters and 35-26-33 percent
among likely voters.

Partisan divisions in the 2008 exit poll were 39-32-29 percent

The whole issue of partisan "weighting" is more complex than it is made out to be by many.

Firstly, a great many folks who "on the first pass" identify as independents will, if pushed, say they lean to one party or the other. 

This poll did not "push" the independents for a soft party affiliation.  Republicans "typically" pick up a few % on the party ID question if you "push" the independents a bit.

My general rule of thumb is to look at the exit polls for 2004 and 2008.

In 2004 it was pretty much a dead heat GOP to Dems, in 2008 the Dems had about a 7% advantage.

This 0 to D+7 range seems to me to be the "rational" playing field.  Clearly the nation is less enthusiastic than the were in 2008 for Obama, but on the other side of the coin, historically the GOP and Dems being even is also pretty rare. - I think this common sense approach pretty much sets the boundries of a reasonable party ID.

[King]

I'd say D+3 is realistic, above that not so much, I highly doubt it will be above 0 for the Dems but I look at it as D+3.  The GOP is not staying home this time and you don't have hope and change.   

https://uselectionatlas.org/FORUM/index.php?topic=161683.0
https://uselectionatlas.org/FORUM/index.php?topic=161683.0
https://uselectionatlas.org/FORUM/index.php?topic=161683.0
https://uselectionatlas.org/FORUM/index.php?topic=161683.0
https://uselectionatlas.org/FORUM/index.php?topic=161683.0

Please respond.

[timothyinMD]

Sad that they had to jack the sample from D+3 to D+9 in order to still eek out an Obama lead.  Romney +3

[Fmr President & Senator Polnut]

http://www.youtube.com/watch?v=siW43SXk2Ac

[Cliffy]

I'd say D+3 is realistic, above that not so much, I highly doubt it will be above 0 for the Dems but I look at it as D+3.  The GOP is not staying home this time and you don't have hope and change.   

https://uselectionatlas.org/FORUM/index.php?topic=161683.0
https://uselectionatlas.org/FORUM/index.php?topic=161683.0
https://uselectionatlas.org/FORUM/index.php?topic=161683.0
https://uselectionatlas.org/FORUM/index.php?topic=161683.0
https://uselectionatlas.org/FORUM/index.php?topic=161683.0

Please respond.

Wow missed a lot, sorry went to bed.  I'll try to catch up but working right now. 

[Eraserhead]

Sad that they had to jack the sample from D+3 to D+9 in order to still eek out an Obama lead.  Romney +3

Not this sh!t again!

[Cliffy]

Sad that they had to jack the sample from D+3 to D+9 in order to still eek out an Obama lead.  Romney +3

Not this sh!t again!

the basic question is do you believe that Democrats are going to turn out 9 pts higher than Republicans in exit polls this time around?

[Brittain33]

Sad that they had to jack the sample from D+3 to D+9 in order to still eek out an Obama lead.  Romney +3

Not this sh!t again!

the basic question is do you believe that Democrats are going to turn out 9 pts higher than Republicans in exit polls this time around?

Lots of former Republicans are now identifying as conservatives.

[ajb]

Sad that they had to jack the sample from D+3 to D+9 in order to still eek out an Obama lead.  Romney +3

Not this sh!t again!

the basic question is do you believe that Democrats are going to turn out 9 pts higher than Republicans in exit polls this time around?

Lots of former Republicans are now identifying as conservatives.

A helpful graph to this effect:

http://www.people-press.org/2012/06/01/trend-in-party-identification-1939-2012/

[Cliffy]

Cool thing is we'll find out in a few weeks

[Speed of Sound]

Sad that they had to jack the sample from D+3 to D+9 in order to still eek out an Obama lead.  Romney +3

Not this sh!t again!

Just do it, Eraserhead. Just take that little ol' mouse, and move that cursor over that nice shiny red button, and just click it. Just click that inviting "ignore" button. You'll never have felt so good in your whole life. I promise you.

[The Vorlon]

Sad that they had to jack the sample from D+3 to D+9 in order to still eek out an Obama lead.  Romney +3

Not this sh!t again!

the basic question is do you believe that Democrats are going to turn out 9 pts higher than Republicans in exit polls this time around?

Lots of former Republicans are now identifying as conservatives.

A helpful graph to this effect:

http://www.people-press.org/2012/06/01/trend-in-party-identification-1939-2012/

So what you are saying is that since PEW showed partisan ID advantages of 3% and 11% in 2004 and 2008 respectively for the Dems, while the actualy exit poll results were 0% and 7%, that PEW pulls 3 or 4% to the Dem side?

So based upon this, PEW current 8% advantage is actually 4 or 5% then?

[ajb]

Sad that they had to jack the sample from D+3 to D+9 in order to still eek out an Obama lead.  Romney +3

Not this sh!t again!

the basic question is do you believe that Democrats are going to turn out 9 pts higher than Republicans in exit polls this time around?

Lots of former Republicans are now identifying as conservatives.

A helpful graph to this effect:

http://www.people-press.org/2012/06/01/trend-in-party-identification-1939-2012/

So what you are saying is that since PEW showed partisan ID advantages of 3% and 11% in 2004 and 2008 respectively for the Dems, while the actualy exit poll results were 0% and 7%, that PEW pulls 3 or 4% to the Dem side?

So based upon this, PEW current 8% advantage is actually 4 or 5% then?

Exit polls are polls, and as such they're not much more "actual" than the Pew numbers. Pew is already polling thousands of people, so their MoE is about as low as you can get. And exit polls and the Pew surveys ask partisan ID questions in slightly different ways, and at different points in the surveys, yadda yadda, so I wouldn't lose a whole lot of sleep over the finer details.
The trends in the Pew data are clear, and as you point out they're backed up by the exit polls. D and R numbers were close to even in 2004, then D numbers rose, and R numbers fell, over the second Bush term, with the result that in 2008 there was a record gap in partisan ID.
Since then, numbers for both D and R have dropped, roughly in tandem, while numbers of independents have increased. That's why it's not surprising to see partisan ID margins this year which closely resemble those in 2008, even though obviously Democratic numbers have declined over the last four years -- Republican numbers have declined, too, roughly in tandem.

[The Vorlon]

Exit polls are polls, and as such they're not much more "actual" than the Pew numbers. Pew is already polling thousands of people, so their MoE is about as low as you can get. And exit polls and the Pew surveys ask partisan ID questions in slightly different ways, and at different points in the surveys, yadda yadda, so I wouldn't lose a whole lot of sleep over the finer details.

Actually, PEW and Exit polls are different..

The exit polls are reweighted after the fact to exactly match the results of a real election.

They are not perfect, but after they have been reweighed, they are awfully darn close.

But returning to PEW...

PEW said a D +11 in 2008, and now PEW says D +8, ie a 3% R improvement

The trend line identified by Pew is the gap has narrowed by about 3% since 2008, so this would suggest 2012 exit polls of Dem +45or so? based on a 3% change from 2008, as per the PEW trendline?

[ajb]

Exit polls are polls, and as such they're not much more "actual" than the Pew numbers. Pew is already polling thousands of people, so their MoE is about as low as you can get. And exit polls and the Pew surveys ask partisan ID questions in slightly different ways, and at different points in the surveys, yadda yadda, so I wouldn't lose a whole lot of sleep over the finer details.

Actually, PEW and Exit polls are different..

The exit polls are reweighted after the fact to exactly match the results of a real election.

They are not perfect, but after they have been reweighed, they are awfully darn close.

But returning to PEW...

PEW said a D +11 in 2008, and now PEW says D +8, ie a 3% R improvement

The trend line identified by Pew is the gap has narrowed by about 3% since 2008, so this would suggest 2012 exit polls of Dem +45or so? based on a 3% change from 2008, as per the PEW trendline?

Yes, exit polls are reweighted after the fact, so that their final results are in accord with the actual election results. Doesn't necessarily mean that they correctly figure out which demographics to rejig. There's particularly clear evidence, for example, that exit polls fairly consistently underestimate the proportion of the Hispanic vote going to Democrats (see  the discussion at http://www.latinodecisions.com/blog/2010/11/04/how-the-national-exit-poll-badly-missed-the-latino-vote-in-2010/, for example). So it's not in fact clear how close to perfect exit polls are on demographic subsets.
Overall, the trend seems to be for a slightly smaller D plurality this time round, if present trends continue to election day. But given that we're talking about a small difference between two larger and uncertainly defined numbers, I'd hesitate to assume a 1-1 correspondence between Pew and the exits.
Bottom line is, there's certainly nothing weird about a D+9 sample right now.

[krazen1211]

Do you know what's happened to national self-ID of Republicans v Independents since 2009?

There's more than one reason why Romney is doing so much better among Independents than McCain did....

Right, right,
You didn't answer my question?  Do you think Democrats are going to turnout 9pts higher than republicans nationally??  Doesn't seem even WaPo does....

Polls from pollsters like PPP do not count. Scroll above.