Races tight, Pennsylvania reclaims "swing state" status: poll

[cali]

(Reuters) - The U.S. Senate race in Pennsylvania has become "too close to call" three weeks before the election, while President Barack Obama's lead in the state has slipped from 12 to 4 percentage points after a disappointing debate performance, a Quinnipiac poll said on Tuesday.

Pennsylvania, seen until recently as secure for the Democratic president, now appears to be competitive, with Obama leading Republican Mitt Romney 50 percent to 46 percent. Just 7 percent of the state's likely voters say they might change their minds before November 6, the poll found.

"It's no secret that the president underperformed in the debate, including by his own account," Tim Malloy, assistant director of the Quinnipiac University Polling Institute, said in an interview.

While Pennsylvania-born Vice President Joe Biden's debate performance last week against Romney's running mate, U.S. Representative Paul Ryan, played well with voters in the state, it "did not move the ticket at all," Malloy said.

<snip>

http://www.reuters.com/article/2012/10/16/us-usa-campaign-pennsylvania-poll-idUSBRE89F0GT20121016

[Shivering Jemmy]

1. Poll's wrong.

PA not a swing state.

[cali]

2. why is the poll wrong?

Or is it just because you don't like it?

[Cosmocat]

3. pretty much sums it up

really tiresome that the same people laughing at republicans for their whinning about the polls 2 weeks ago are now whinning about the polls, and anyone who dares say the numbers are the numbers are hating scumbags looking to drag people down.

I live in PA, and I still feel pretty confident that BO and Casey will win, and there has been polling since that shows a larger margin for both.

But, there is no doubt that whatever way it was spun, the debate tightened things up substancially in PA like everywhere else in the country.

Gotta hope our guy rings the bell tonight, and be ready to do what we can as this election rounds to the final drive.

[Shivering Jemmy]

8. See my post above

For Cali

[ananda]

12. That is my perception also..

.. that PA is moving even more strongly into the Dem camp!

[BumRushDaShow]

4. PA is not a swing state

We go through this cycle after cycle. It is appearing that there are demographics such as women and minorities (blacks and hispanics) that have suddenly been dropped out of "likely voter" polling. And all of this is being attributed to one debate, which is nonsense.

[justiceischeap]

5. I know yesterday's polls are nothing but there's a 3 point difference between

Quinni and PPP (PPP has Obama @ +7). Also, 1988 was the last time PA went Republican. I can't imagine it'll go this time around.

[tishaLA]

6. they used to get weird polls like this in NJ too

and republicons thought they might be able to take the state, but never do. Or, in 2004, there was a poll suggesting HI might be winnable for the repubs.

I'm not worried about one poll saying the state is close.

[Shivering Jemmy]

7. I don't like or dislike polls. They are numbers.

It's at the far end of a distribution. Other polls have it at 5 or 7. Median around 5.5. So it is data about the distribution. But it is not characteristic of the distribution.
But in the write up the pollster pretends it is.

Put more succinctly: polls wrong

[OldDem2012]

15. When the pollsters switched to "Likely Voters (LV)", they seem to have completely disregarded....

....the huge lead the President has with women and minorities. The problem is how the pollsters view Likely Voters as being much more enthusiastic on the GOP side and a lot less enthusiastic on the Democratic Party side.

Why is that?

How did that happen virtually overnight?

Think about it.

Come on, cali...the evidence of a bias against Dem Likely Voters is right there in front of your face.

And yes, I don't like pollsters who prove themselves to be dishonest.

[JoePhilly]

9. Real Clear average has Obama +5.

And if you look at the last 6 polls taken ... check the ranges.

Obama's lowest is 47% and his highest is 51%. Obama is 50% or better in 3 of the 6 polls.
Romney's lowest is 42% and his highest is 46%.

Romney's ceiling is below Obama's floor.

[fugop]

10. Thank you!

Putting it in those terms soothed my mind. Well done!

[JoePhilly]

14. The last time Romney beat Obama in ANY PA poll was 1st week of February.

I pay less attention to the averages and much more attention to the ranges.

[Yo_Mama_Been_Loggin]

20. Yet their map shows PA as a toss up.

Go figure

[JoePhilly]

21. Because they base the map primarily on the averages. Which only tell you so much.

What I'm trying to describe, and I'll admit, in a rather simplistic way, is one of the ways one looks for statistical significance.

The important question is not simply the difference between the averages for Obama and Romney. The important question is whether or not the difference in the averages is persistent and reliable.

And examining the RANGES is one, relatively simplistic approach to getting a sense of that. The lead could be TINY and still be a persistent and reliable finding.

And the fact that RC's map shows PA as a toss up is irrelevant to the question of how reliable and persistent Obama's lead is. For instance, 538.com has their own model showing PA leaning Obama (same 5% lead), and yet they say the probability that Obama wins PA is 92%.

Why? Because they are looking beyond the simple average, and examining other aspects that help determine how persistent and reliable that difference is.

As you said "go figure", but you'll need to move beyond a discussion of simple averages if you want to know what the difference between the averages indicates.

[JackN415]

11. After being curious about how polls are done, and thought about it...

I no longer care or take them seriously.

http://www.democraticunderground.com/10021541900

http://www.democraticunderground.com/10021533669

The errors of skewed samples are likely much bigger than these pollsters' claim.

[morningfog]

13. I am not at all worried about PA. It is squarely in Obama's win column.

[OldDem2012]

18. Bingo! nt.

[PRETZEL]

16. Where I'm at in PA

I'm still pretty confident about the President being the winner here. It's true that there are many more Rmoney signs but as an indication of strength, I'd still say Rmoney's support is still fairly weak. There just doesn't seem like there's as many signs around as in past elections.

Casey, on the other hand, I think is running a pretty poor campaign. He's only running one ad around here and he's been running the same one all season. It's getting old. He needs to start getting a bit more serious.

One good week of campaigning by both the President and Casey here will stop any sort of talk of a Rmoney/Tom Smith for Senate talk.

[Cosmocat]

17. yep

I live in a pretty conservative area, and there are far more signs for Romney than BO, but it should be even more than it is.

I would say half the signs for Romney that there were for Bush.

Really, no one, LITERALLY mentions Romney. They just hate president Obama. That is it, they don't even try to advocate for Romney.

[HappyMe]

19. Meh.

In 20 minutes there will be another poll going the other way. The only thing that matters are the numbers on election day.