Quinnipiac/NYT/CBS Swing State Poll: Obama with big leads in PA, OH & FL

[Tender Branson]

A snapshot of the race, taken during a burst of summer campaigning, found that Mr. Obama holds an advantage of 6 percentage points over Mr. Romney in Florida and Ohio. The president is in a stronger position in Pennsylvania, leading by 11 percentage points, the poll found. The margin of sampling of error is plus or minus 3 percentage points in each state.

The New York Times, in collaboration with Quinnipiac and CBS News, is tracking the presidential contest in six states through polls over the next three months. Ohio, Florida and Pennsylvania have a combined 67 electoral votes. Surveys will also be taken in Colorado, Wisconsin and Virginia, which have 32 electoral votes. Four years ago, Mr. Obama won all six states.

Read more: http://www.post-gazette.com/stories/news/us/-obama-leading-romney-in-3-pivotal-states-647041/#ixzz22GsSWBzW

Detailed numbers will be released later today on the Quinnipiac page:

http://www.quinnipiac.edu/institutes-and-centers/polling-institute

[Tender Branson]

Good to see.

I'd like to see the Pro-Romney camp invest 50 Mio. $ into Pennsylvania, just like McCain did - with absolutely no effect (maybe even the opposite, pushing the state even further into the Obama-column).

[Tender Branson]

BTW:

The article says that this is Quinnipiac's first set of "likely voters".

[Tender Branson]

Are these Quinnipiac polls or NYT/CBS polls? Pretty huge difference in quality...

No, NYT and CBS pay Quinnipiac to conduct the polls for them. But that doesn't change the quality. Quinnipiac has been a very good pollster in recent years, so don't shoot the messenger. The problem is Mitt Romney after all, not the pollster.

[Tender Branson]

Are these Quinnipiac polls or NYT/CBS polls? Pretty huge difference in quality...

No, NYT and CBS pay Quinnipiac to conduct the polls for them. But that doesn't change the quality. Quinnipiac has been a very good pollster in recent years, so don't shoot the messenger. The problem is Mitt Romney after all, not the pollster.

Oh yeah, as fine it's Quinnipiac doing the polls, that's good. These are pretty good numbers then, though the Florida lead is kind of hard to believe.

Maybe. But remember that SurveyUSA also showed Obama+5 in their latest FL poll.

And Quinnipiac and SurveyUSA basically had the best FL polls in 2008:

https://uselectionatlas.org/POLLS/PRESIDENT/2008/pollsa.php?fips=12

[Tender Branson]

I'm actually more interested what Quinnipiac will find in their Colorado and Wisconsin polls.

[Tender Branson]

Full results:

FL: 51-45 Obama
OH: 50-44 Obama
PA: 53-42 Obama

July 24-30, 2012

Florida Likely voters N = 1,177
Ohio Likely voters N = 1,193
Pennsylvania Likely voters N = 1,168

You must be logged in to read this quote.

http://s3.documentcloud.org/documents/405617/the-quinnipiac-university-new-york-times-cbs-poll.pdf

[Tender Branson]

The NYT also has a new site with crosstabs:

http://elections.nytimes.com/2012/swing-state-polls

[Tender Branson]

Florida D+9! Ohio D+ 8!

Rightttt....

GIGO. Shrug.

I already thought you would complain about the sample. But Quinnipiac asked both questions in this poll: partisan ID and how respondents are registered.

How those "likely voters" are currently registered:

FL: 42% DEM, 36% GOP, 22% IND/OTH
OH: 42% DEM, 35% GOP, 22% IND/OTH
PA: 46% DEM, 40% GOP, 14% IND/OTH

Very similar to the actual voter registration numbers in each state.

But of course you don't like the numbers ...

[Tender Branson]

Florida D+9! Ohio D+ 8!

Rightttt....

GIGO. Shrug.

I already thought you would complain about the sample. But Quinnipiac asked both questions in this poll: partisan ID and how respondents are registered.

How those "likely voters" are currently registered:

FL: 42% DEM, 36% GOP, 22% IND/OTH
OH: 42% DEM, 35% GOP, 22% IND/OTH
PA: 46% DEM, 40% GOP, 14% IND/OTH

Very similar to the actual voter registration numbers in each state.

But of course you don't like the numbers ...

Well, that's interesting, because Florida is actually only 40% D by registration and most expect a dropoff for the Democrats of 2 points or so when using a likely voter screen. It's very curious to flip that in the opposite direction.

And where precisely did you get your actual voter registration for Ohio?

It doesn't really matter what "most believe". The poll tells us something different.

There is no party registration in OH, but for the other 2.

[Tender Branson]

Florida D+9! Ohio D+ 8!

Rightttt....

GIGO. Shrug.

I already thought you would complain about the sample. But Quinnipiac asked both questions in this poll: partisan ID and how respondents are registered.

How those "likely voters" are currently registered:

FL: 42% DEM, 36% GOP, 22% IND/OTH
OH: 42% DEM, 35% GOP, 22% IND/OTH
PA: 46% DEM, 40% GOP, 14% IND/OTH

Very similar to the actual voter registration numbers in each state.

But of course you don't like the numbers ...

Well, that's interesting, because Florida is actually only 40% D by registration and most expect a dropoff for the Democrats of 2 points or so when using a likely voter screen. It's very curious to flip that in the opposite direction.

And where precisely did you get your actual voter registration for Ohio?

It doesn't really matter what "most believe". The poll tells us something different.

There is no party registration in OH, but for the other 2.

'The poll' doesn't tell us; rather, those doing the party weighting of the poll do.

Actual calls by 'the poll' show a mere 1% edge for the Democrats in the unweighted sample in Florida and a 4% edge for the Democrats in Ohio. Certainly you and others are entitled to figure that such should translate into a 6/7% edge for the Democrats, but that's on you.

Well, I guess the folks "weighting" the polls at Quinnipiac know what they are doing.

Otherwise they wouldn't have correctly predicted the Romney+14 win in FL in the primaries and the Romney+1 win in OH. And also their FL 2008 poll was spot on, OH was only off by 2 points and PA was also spot on.

[Tender Branson]

This is also fitting for this thread (coming from Scotty Rasmussen):

You must be logged in to read this quote.

[Tender Branson]

The rejection of polls you dont like and embracing of those you do thing is becoming an embarrassing cliche.

Believe me, I wish these numbers were true.  I think having Democrats with a +9 party ID in Florida and a +8 party ID in Ohio is a little too optimistic.  Nothing wrong with looking at the actual registration numbers and adjusting to get a clearer view of where the race is.  

The poll actually uses registration numbers that are similar to the current registration in the states:

https://uselectionatlas.org/FORUM/index.php?topic=156770.msg3370217#msg3370217

Party ID is just something very different, many are IDing with Independents in polls because they don't like policies of both candidates, but on election day when the exit pollster comes around they are more likely to be motivated to check a partisan label.

[Tender Branson]

Well, I guess the folks "weighting" the polls at Quinnipiac know what they are doing.

Otherwise they wouldn't have correctly predicted the Romney+14 win in FL in the primaries and the Romney+1 win in OH. And also their FL 2008 poll was spot on, OH was only off by 2 points and PA was also spot on.

That's a possibility. That would of course discount that they 'knew what they were doing' a mere 3 months ago when the sample was GOP + 3, and, as to be expected, Romney had a 6 point lead.

At the time, interestingly enough, those results were disputed by Democrats.

Of course on October 23, 2008, Quinnipiac released a poll showing Obama up 14 in Ohio over McCain. Thus, history shows at best that they sometimes know what they are doing.

McCain was basically imploding in the first half of October 2008 ...

And there have been other polls during this period, even in OH, that had Obama up by 10+

Maybe he was really ahead by that much at this time.