Sweet Poll Goodness from Zogby:

[Teaser]

Kickin Ass and takin' names in the battleground states, baybee.

http://online.wsj.com/public/resources/documents/info-battleground04.html


Damn, Zogby gives good poll.

[Eurobabe]

1. Fvck Bu$h in OHIO! I'm moving there

one more vote in Southeast for KERRY/EDWARDS!

[irancontra]

2. very nice..

he does give good poll. ;)

[featherman]

3. and THERE's your electoral majority

Gore states plus NH = 264
add in FL to win but keep workin' NV, OH, CO, WV, AR, MO

[cindyw]

4. Arkansa, Florida and New Hampshire!

That's fine with me. Looks like these stupid polls showing Wisconsin a Bush state are off base. I would trust Zogby over them any day. I also don't have as low of an opinion of online interactive polling as others do. Voting itself is interactive. A person answering their phone and saying they are likely to vote is not as important as someone who seeks out to join a Zogby interactive poll. Isn't that closer to actual voting anyway. You must seek out the polling booth. It does not come to you.

[Teaser]

9. Zogby has extensively tested his interactive polls

I remember when they were being tested. He asked all sorts of questions, not just political, and tested them against phone poll results. Things like shopping habits, etc.

Zogby is a scientist and I trust that he wouldn't go into production with a product he hadn't adequately tested.

There's nothing wrong with doing an interactive poll so long as the poller adequately weights for self-selection bias. I assume that in his multiple tests, Zogby managed to refine a formula (which he'll never reveal, I'm sure) that compensates for self-selection bias.

Really, all one has to do is find a statistical fudge factor to multiply by the interactive poll results. What happened, over the years, is that Z noticed a correlation between self-report and random report polls. He likely then did a regression analysis on the phenomenon and wound up with this methodology.

If it works well, Gallup will be out of business and who wouldn't love that?

[cindyw]

16. I can't believe Gallup is not out of business yet.

I have been watching them for the last four years. Before now, when the only polls to watch were presidential popularity polls, I noticed that gallup was consistently 10 points higher than anyone else. Always in Bush's favor.

[Gloria]

39. Gallup is a name only...it was sold years ago to some outfit in

Nebraska or Kansas, I forget which. Their "home office" in Princeton is there because of "image." I know of only one person there who was working with me at the "real" Gallup back in 1979-1980. And, there are only a few people there...when I visited, it was like a "shell" company...

When I was there, Gallup was full of academic types and it was an honor to be there...heck, it paid less than most research houses in the Princeton area. The old man still walked the halls of Bank Street and bigshots like Sander Vanocuor (remember him? I can't spell his name anymore) would pay visits.

The phone room was on site...literally, about 6 lines, until they opened a slightly bigger phone operation in another office at Research Park off Rt. 206. The gal running the show was absolutely a stickler, as was our coding dept. and field operations. At that time, the Presidential preference question was put on the Omnibus survey, which was a hodge podge of questions from in house and clients studies, and was done in IN PERSON interviewing!

Andy Kohut, who does the analysis for PBS at election time and now heads Pew (as well as founding Princeton Survey Research) was the President when I was there (alternating position with another guy). He was pretty intense even then and very serious. So, I would respect his analysis.

But as far as Gallup goes......it's not Gallup anymore and hasn't been for years!!

[tritsofme]

21. Does he claim Interactive are scientific?

I don't believe he does, and I take this polls with a grain of salt.

[Teaser]

23. It's as scientific as any other poll.

A telephone poll is no more "scientific" than any other kind of data sampling technique. And I speak as a scientist.

in the past, telephone polling has worked because the distribution of telephones was approximately uniform across the population. So one was as like to get a conservative R or a liberal D or whatever when a random set of digits was dialed.

As has been extensively written about, this relationship is falling apart between call screening, non-response, and cell phone only households. So one might ask if a telephone poll can claim to be any more "scientific" than an internet poll.

In any event, this isn't some web-based MSNBC poll question of the night. Zogby's interactive surveys are meticulously researched matters. He collects extensive information from respondees on all sorts of demographic information and then assembles a set of respondents who match, at ever demographic category, the demographics of Z's likely voter pool. He also uses a screen to downweight the self-response bias in the data. I don't know anything about the nature of the screen, Z doesn't release it's details. The fact that it's proprietary would seem to indicate that it does work. He wouldn't waste time hiding the details of something that didn't work.

We'll see if it's all it's cracked up to be, but I guarantee you that it's been tested out the wazoo.

[tritsofme]

26. I just don't understand how he could possibly get a respresentative

sample of voters from a pool of respondents that are political junkies.

[Teaser]

36. Those ain't the only respondents.

And as I said, he has a formula to correct for the self-report bias of political junkies. If you have such a formula, you know how error propagates into the outcome percentages and can set confidence intervals (MOEs) on the outcomes.

That said, all of this hinges on Zogby's formula for correct for the "junkee factor" being correct. I have no idea if it is or not. I know he's tested it extensively, but we'll have to wait and see. I just trust that he wouldn't stake the reputation of his business on some half baked idea he hasn't tested adequately.

[Mr_Spock]

5. Um, are you saying that polls can change over time?

I'm shocked!

[iwantmycountryback]

6. I wonder if we should just give up on W.Virginia

It seems we've been pretty consistently down there so I wonder if the 5 EV's are worth spending money there.

[featherman]

10. disagree... Mason Dixon (a GOP leaning poll) has

Kerry down just 1 point as of yesterday. Bush's reelect and approval ratings there are in the toilet (trailing national). WV is a DEM state that went GOP in 2000. I think it is worth working hard.

[tritsofme]

32. Mason/Dixon is GOP leaning?

They have a long track record of getting good results for state polls.

I'd trust them over Zogby Interactive any day of the week.

[featherman]

42. Interesting comment. I've been watching them for years

in PA in particular and found them to be off by +5 or more on the GOP side as compared with the final vote tally. But whatever... if you like 'em, you like 'em and must give credibility to the notion that Bush is up by +9 in NH, a prima facie absurdity.

Sample from 2000 (PA): "In a survey of 416 registered voters who said they planned to vote in November, Bush led Gore 44 percent to 40 percent ... Buchanan was supported by 5 percent and 11 percent were undecided"

p.s I am not accusing conscious, conspiratorial bias when I say "GOP leaning" merely that their methodology likely overpolls the GOP side.

[Sydnie]

13. Absolutely not!!!

If Gore had won WV, we wouldn't be in this mess now. They can be salvaged. Kerry needs to talk more about coal and clean coal technology and he can have them I think. If Kerry is busy with the new message that he delivered today, then we need some viable local surrogates to spread the clean coal technology info there and often. He will be losing the anti-abortion crowd there, but coal and the coal economy will win more than we will lose.

Remember, many that live in WV have relatives that live in Ohio and Kentucky and Tennessee and Virgina. Winning them can help a snowball effect in that area.

[dumpster_baby]

7. I sure wish a bunch of CA Dems would move to NV...temporarily....

:-)

[featherman]

12. my daughter just moved from Oregon to NV

she will register and vote Kerry.

[marigold20]

8. In 2000, wasn't one pollster much more accurate at the end?

I keep thinking it was Zogby, but I'm not sure.

[Poiuyt]

14. I think Zogby and Harris were the most accurate

[Ilsa]

11. West Virginia to Bush by 12?

I thought it was closer there.

[WVhill]

40. It would have been except Kerry had to go public supporting the AWB.

He guaranteed that Bush would take WV after that stunt.

[tedoll78]

15. I'm beginning to wonder..

if we should change our focus to Florida, rather than Ohio.

The nation's cameras will already be on Florida. International vote monitoring organizations will be everywhere there. The Democrats will have an army of lawyers there. Heck, Michael Moore will be there as well.

It seems as though the polls in Florida have pretty much consistently leaned to Kerry. The only exception has been the past month, and that was from the worst 6 weeks of Kerry's political life.. and even then Bush only managed a small lead there.

Thus, I wonder.. can we outsmart Jeb for those 27 electoral votes? With those 27 electoral votes, we could even afford to lose a medium-sized Gore state. Or several small ones.

I wonder..

[Teaser]

17. I have some thoughts about the significance of these #s

Which I will post on Burying the Lede today or tomorrow. It's really good news for us.

[mfritz]

18. Colorado is a battleground now

Polls have shown the state as a virtual tie for a while now, and the Democrat is ahead in the senate race. I wish Zogby would add Colorado to his battleground poll.

[abrock]

19. We're on our way to victory guys! This confirms!

[cthrumatrix]

20. making a move....somebody wake up the media weenies from their nap

[Doosh]

22. Kerry up 9 in Minnesota, 12 in Oregon

looking good!

still alive and kicking in tennessee

[Zech Marquis]

24. I think Kerry can get Nevada!

He's still in striking range for Ohio, and Nevada is virtually tied :bounce: a victory in Florida would be sooo sweet!

[blondeatlast]

25. Nevada will go Kerry over Yucca Mountain.

The southern part of Nevada is supremely pissed over the issue.

[johnnyrocket]

27. Ohio blows my mind...WTF is up with them Ohioans...

they're barely southern redneck dolts...and have taken an ass whooping from Bush's domestic policies. They should be all over Kerry's side. Someone needs to give them a clue.

[mconvente]

28. agreed

I don't believe that after Ohioans lose the most jobs under Bush they can still want to vote for him. Ohio is within our reach. We need Edwards and Clinton there (hopefully he'll recover soon) to rally the independents.

[0rganism]

29. The most important results in this batch are PA, WI, MI, and FL

PA, WI, and MI are "must win" for Kerry. If bush can bust a lead in the northern rust belt, we're toast. Lately they've been swinging around quite a bit, and it's nice to see the Zogmeister posting some good news about 'em.

FL is very interesting, because if Kerry wins Florida, he picks up a lot of permutations for the combinations of states needed to get 270 ev.

[mconvente]

31. lol

permutation implies order matters, but that's not so when it comes to EV totals (that would be a combination)! Sorry if I sound elitest, but I'm a math guy so I gotta keep it real!

[0rganism]

44. actually, you (unintentionally) raise a good point -- does order matter?

Florida is on the east coast. If it gets called for one of the candidates early, it could have an impact on the other state turnouts -- at least, that's what the cranky panhandle voters claimed in 2000. So let's say we hear turnout is really heavy in the Democratic counties of south and central Florida, that might encourage a lot of people to vote in the other Democratic states. Let's hope it happens.

re terminology, you're right. I meant combination. Glad you could bust a gut over it. Six years ago, I might have too.

[Yo_Mama_Been_Loggin]

30. Interesting

Zogby has Kerry leading in Iowa while the latest Gallup shows * with a sizable lead there. Then again after Gallup's latest Republican heavy poll I think both them and Ralph Nader have a bit in common. They once stood for something but are now has beens.

[theHandpuppet]

33. Won't load for me...

Anyone care to provide a summary? Thanks!

[johnnyrocket]

34. Then WHATS UP with these 'electoral polls' that shows Bush way up?

They're bandied about DAILY on tv and radio and the papers. Who has seen a major news story on Kerry winning the EV in the polls?

[4dsc]

35. Zogby stated the winner of the election

would be decided on what issue was being talked about in November. During an radio interview on the Ed Schults show, Zogby stated that if national security was the main issue of the country that Bush would win. He stated if the issues of jobs and health care were the main issues being dicussed, then Kerry would be the winner..

Listen to Zogby please..

[Roland99]

37. I KNEW TN was in play!

[in_cog_ni_to]

38. Pa. too! FL.????

wow! I DO like this poll. :bounce:

Thanks for posting it! Too bad we won't see THIS on the state owned cable networks. :grr:

[Fiendish Thingy]

41. And this poll was after only 1 week of "taking the gloves off"

even with all the memo crap. Just think what the next Zogby will show after a week or more of "stubborn incompetence". Go Johnny Go!

[Kal Belgarion]

43. Remember: Zogby also called the election for Kerry months ago

Looks like he could be right again! :)