ZOGBY/Inisder Advatage showing late Obama shift in PA yesterday

[AnarchoFreeThinker]

Two polls this morning for polling done yesterday only both show a last pro-Obama shift. That said, Clinton has countered each pro-Obama shift in the past in PA. Bin Laden. So take this with a grain o' salt.

Insider Advantage's polling on Sunday only: Clinton 49, Obama 39
Insider Advatage's polling on Monday only: Clinton 49 (even), Obama 42 (up 3)
NOTE where the Undecideds went here.

Zogby's polling on Sunday only (not to be confused with the Saturday-Sunday total): Clinton 53, Obama 38
Zogby's polling on Monday only (extrapolated from two-day Sunday-Monday total): roughly Clinton 48 (down 5), roughly Obama 43 (up 5)

So it's looking like a 5-7 point Clinton victory IF TURNOUT IS GOOD FOR:
1. The rural vote.
2. The women's vote.
3. The over-35 vote.
4. The white vote.
5. The Catholic vote.
6. The non-college educated vote.
Any indications turnout is slight for these groups can cause her some trouble.

Crosstabs from PPV's poll from the weekend and exit polls suggest that Obama will do very well, though, IF TURNOUT IS GOOD FOR:
1. The under-35 vote.
2. The college-educated vote.
3. The newly registered vote.
4. The African-American vote.
5. The liberal vote.
6. The urban vote.
Of these, the biggest mystery is how accurately pollsters have captured the under-35 vote. PPV weighted this group in line with turnout in some earlier primaries in PA's neighbors and were alone in showing Obama ahead in PA during Saturday-Sunday polling. Then there's the thousands of newly registered voters who politico.com says will break 60-40 for Obama. Again, though, just how accurately pollsters have captured that group--due in no small part to a large number of its constituents being under-35--is another issue. Polling is inaccurate when a candidate changes the electorate, but at this point you've got to figure pollsters might have a little more of a clue how Obama changes the electorate.

In short, I have no idea. It would be easy to take a median and go from there, and, true, that would maximize my chance of being least wrong. But it certainly wouldn't guarantee I was right.

My GUESS is without Clinton's bin Laden ad, Obama was headed for a 2-3 point loss today (if we can believe the trend lines from the only two polls from yesterday exclusively). That ad is sure to rouse a few thousand conservatives, though, and make her starting margin more like 5-6. If she gets all her peeps to the booth, it could be 10-12. If he gets his to the booth and she doesn't, it could be a +1 to -3 margin for him. If they both turn out their voters en mass equally, then we're back to 5-6.

So, to summarize, after reading about sixty polls over the last six weeks, plus exit polls, my final prediction for PA is....

that it will depend on who votes. It'll be a 1-point loss to a 12-point win for Clinton. How's that for a cop-out?

[SheilaT]

1. I am not making any predictions

especially as I'm not following the polling in Pennsylvania closely, but just from what I see around me, especially here in Kansas, Obama has pulled in new voters, especially young ones, and also older (middle-aged) ones who rarely if ever voted before. Which isn't to imply that Clinton supporters aren't equally dedicated and motivated.

[Coexist]

2. wow, great post

http://ruralvotes.com/thefield/

the above link shows some county breakdowns, too.

I will be referring to your post when they discuss exit polls, and that link as I follow the state map on CNN.

(I love MSNBC, but often switch to CNN, they have 'smarter' coverage - more details, etc.)

[SteppingRazor]

3. That's about where I've got it.

You said total of 1-point loszs to 12-point victory, but your most likely outcome looks like 5-6 point victory for Clinton. I've been saying 5-9 or 5-10 point victory for Clinton for a couple days now, though if I were a betting man (and I am) I wouldn't put money on the Penn. primary (I haven't).

[Eurobabe]

4. Obama by squeaker +2. Rinse.Lather.Repeat if necessary.

[PBS Poll-435]

5. Huh?

Clinton performed the best since Zogby started daily PA tracking, and Obama did the worst.

She broke 50%(51)

Obama fell to 41%.

And, his support among all polls in the last month (except for yesterday's PPP outlier) is from 37-45%. He will not break that number.

[thewiseguy]

11. Zogby takes the two day average and Hillary was up 15 for Sunday and up 5 for Monday

So he took the average and her lead came to be 10.

Also for Sunday night's poll. Hillary was as I said up 15 for Sunday and down by 3 on Saturday. So he took the average again and Hillary ended being up by 6 in that poll.

Bottom line Zogby's numbers are so inconsitent that I would not pay attention to them.

On Saturday Obama was up by 3!
On Sunday Hillary went up by 15!
On Monday Hillary was up by 5!

So his numbers are fluctuating by 10 points on a daily basis. Which is insane and unreliable.

[AnarchoFreeThinker]

17. that's an excellent point.

[Teaser]

12. logged

.

[PBS Poll-435]

15. For what?

[Teaser]

28. just tagging it

.

[grantcart]

6. two other factors

1) no Limbaugh effect

2) Those who support Clinton but are tired of the war

[MaineDem]

7. As usual, it comes down to GOTV

Which camp does better at getting its voters to the polls. That's what it's all about.

Good luck to everyone involved. It's been a long campaign.

[ginnyinWI]

8. well, at least you're honest!

:) The fact is that there are so many variables that nobody knows.

[Teaser]

9. HRC by 8

an inconclusive result.

[book_worm]

10. all bets are off if turnout is strong in Philly and its suburbs and Obama gets

a strong vote.

[AnarchoFreeThinker]

25. yep

[Cosmocat]

13. Conservatives ???

what "conservatives" ...

Pa has a closed primary ...

All the "variables" people have been wanting to speculate about, it comes down to one thing ... PA is an older state ... Bottom line, given the choice between the Clinton brand and a younger, upstart, african american from Chicago, they are going with the Clinton brand ...

The last few weeks the polls have been anywhere between -5 to -teens for Obama ... Average it out and it is -7/-8 ... Maybe give BO a slight bump for the under polled younger vote and having a better ground game, and the best he is looking at is -5 or so ... A bad day probably puts him in the -9 range ...

[AnarchoFreeThinker]

18. lots of people who I.D. themselves as conservative vote D

It's not a majority, by any stretch, but enough that it'd matter if, for instance, they vote b/c of the bin Laden ad. that's why they ran it. They weren't playing to progressives.

[totodeinhere]

14. Didn't the pre-election polls in New Hampshire show Obama winning by better than 10 points?

I don't believe the polls anymore. IMO it's anyone's guess who will win in PA. But one thing is clear. Anything short of a blowout for Clinton is a loss for Clinton.

[ORDem]

16. Those shifts are within the MOE. nt

[AnarchoFreeThinker]

19. true...but looks more like a shift when both polls head the same way.

Although how much of a shift and how permanent it is are, to say the least, inconclusive.

[GoesTo11]

20. Zogby sure likes getting attention

Trust me, there was not a real 5% shift in sentiment one way one day and back the next day. Just sampling error.

[high density]

21. Zogby seems to be saying everything in the hopes that one

thing he says turns out to be right.

[windbreeze]

22. I almost hate to say this...but...being from Pa...

I think it's a toss up...gonna be just about even...and that's w/o looking at any polls..I'll eat my words if I have too...but this one is a real nail biter for either candidate...I won't say why...I just hope there's NO shenanigans with the votes...wb

[Laura PourMeADrink]

23. I've been crunching numbers too - my range - close to yours is a .9375

Obama win to a 11.03 Clinton win - Average: 6 Pt Clinton win.

But what's interesting is how PPP captured the young vote:
PPV weighted this group in line with turnout in some earlier primaries in PA's

Do you know how the other pollsters weighted them?

[AnarchoFreeThinker]

24. on the other weighting of young voters....

for Zogby and Rasmussen, I don't know. You have to subscribe for that breakdown. For Suffolk (Clinton by 10), they had the under-35s at at 8% of the sample, when 24-25% is more likely. For Survey USA (Clinton by 6) and for Insider Advantage (Clinton by 7), they each were closer, at about 21% of the sample from under-35s (which generously assumes in each case an even breakdown within one of their age subgroups). Then there was PPP (Obama by 3), which had the under-35s at about 24-25% (again, after extrapolating). It looks, not surprisingly, that the larger the under-35 group in the sample, the better Obama does.





ed for typo

[futureliveshere]

26. Great post except that I am nervous when anyone quotes ZOGBY. He sucks bigtime!!!

[AnarchoFreeThinker]

27. yeah, he's been ALL over the map this cycle

[stevenleser]

29. Interesting. I think the Bin Laden ad will have the opposite effect.

Perhaps I am giving people too much credit, but I think people who see it will be offended.

[Political Heretic]

30. Nice analysis, friend. Well done.

[AnarchoFreeThinker]

31. UPDATE: did Clinton going negative supress the vote of her own leaners?

Found this to suggest the rural Undecideds might actually decide to stay home, among other scenarios discussed:

http://www.fivethirtyeight.com/2008/04/its-turnout-stupid.html