PEW poll. Kerry 46 Bush 45 RV. Bush 48 Kerry 45 LV

[jezebel]

http://people-press.org/reports/display.php3?ReportID=232
Bullshit LV. this must be the poll Josh Marshall was talking about.

President George W. Bush holds a slight edge over Senator John Kerry in the final days of Campaign 2004. The Pew Research Center's final pre-election poll of 1,925 likely voters, conducted Oct. 27-30, finds Bush with a three-point edge (48% to 45% for Kerry); Ralph Nader draws 1%, and 6% are undecided.

The poll finds indications that turnout will be significantly higher than in the two previous presidential elections, especially among younger people. Yet Bush gets the boost Republican candidates typically receive when the sample is narrowed from the base of 2,408 registered voters to those most likely to vote. (Among all registered voters, Kerry and Bush are in a virtual tie: 46% Kerry, 45% Bush).

Pew's final survey suggests that the remaining undecided vote may break only slightly in Kerry's favor. When both turnout and the probable decisions of undecided voters are taken into account in Pew's final estimate, Bush holds a slight 51%-48% margin. The poll, taken over a four-day period, found the recent video tape from Osama bin Laden had no clear impact on voter preferences. Interviews conducted after the tape was released on Oct. 29 generally resembled the polling conducted on the two previous days.
The potential still exists for changes in voter opinion and, equally important, in the composition of the electorate on Nov. 2. While 6% of likely voters are undecided, another 8% still leave open the possibility of changing their vote.

[Moonbeam_Starlight]

1. Which is more reliable?

RV or LV? I'm thinking RV?

[TexasSissy]

3. No one knows yet. But LV depends on how a pollster defines LV,

while RV is a straight up and down fact. But RV includes people who, of course, probably won't bother to vote.

[Old and In the Way]

7. Neither.....MV

Motivated Voters

Definition: People who may not have voted last election, may not have been registered last election, may not be easily categorized...but all know that this is the most important election of their lifetime.

We be hearing about this new phenomenon for weeks after the election as the polling firms explain why the results would totally outside of their normal margin of error.

[Demfromct]

2. Not worried

Seems like statistical flucuation. And LV models in all these polls are suspect. Bush still at 48. This race is tied baby. GOTV!

[sonicx]

4. partisan breakdown...

"Each candidate garners the support of about 90% of their partisans. Kerry holds a slight 48%-44% margin among independent voters."

Kerry's really ahead.

[Orangepeel]

5. uh, a 3 point lead is as much as a "virtual tie" as a 1 point lead.

even holding aside this likely voter estimation.

A tie in the polls is good news, I think, if turnout is as high as most people expect.

[pabloseb]

6. internals are GREAT

Kerry ahead: 1 point among those who have already voted, 4 points among independents, 3 points in battleground states.

All despite repug oversampling.

[DemocratSinceBirth]

8. they oversampled pukes..

39% Pukes


35% Good Guys


25% The Clueless

[tritsofme]

9. Why are polls consistently oversamling goopers

that don't weight their results?

This hasn't happened consistently in other elections.

The only two options I see is that either there has been a silent shift to the GOP over the past few years, or cell phones are having a much bigger impact than previously thought.

[DaveinMD]

13. remember

the effect of caller id. There is a much higher refuse rate for polling than in the past.

[tritsofme]

14. I don't know about other folks

but I think I've had a caller ID for the better part of 10 years.

Has there been such a drastic change in just 4 years?

[DaveinMD]

15. the mass dispersement of caller id

is fairly new. The services used to cost more. And with all the telemarketing many more people are not answering the phones. It takes longer to do an accurate poll nowadays than it used to when I first started working in politics.

[Philostopher]

16. It's been theorized before ...

that the pollsters do a fixed number of calls -- say 1,500 -- and then report the number of actual responses. That's why you'll see inexplicably odd numbers like 'out of 456 respondents ...' -- because that's who picked up the phone and agreed to be polled.

I won't theorize on it, I'll only say apparently, Republicans are more likely to be home, answer the phone and commit to being polled. Oversampling may not be intentional -- it may just be that they're the ones who answer the phone and agree to sit through a poll.

Any pollster who doesn't report that up front is being dishonest, regardless of the weighting of the sample. Oversampling of Republicans probably isn't intentional, if they report it or note their statistical method for weighting around it. If they don't admit it, whoever's oversampled, or report the breakdown or the statistical method, the poll may well be biased.

[tritsofme]

18. But I don't understand

why this would be a relatively new phenomena, if anything conventional wisdom would suggest to me that Democrats are more fired up about this election and would be willing to spend time talking to pollsters.

[DaveinMD]

19. Dems could be

less likely reachable with traditional polling.

[Philostopher]

20. Here's an article by Jimmy Breslin about the subject.

I know Breslin's partisan (for our side), but it addresses some of the issues you're asking about all in one place (I've seen several different articles addressing different issues, but this is one that collects them):

http://www.newsday.com/news/local/newyork/columnists/ny-nybres163973220sep16,0,5538561.column

[Demfromct]

10. Oversampled?

I feel better. And also, I wonder how many Zell Miller Dems they have included in that survey from the South. If u are weightting, and you get a dem from the south, that skews the poll because. Get it.

[DaveinMD]

11. remember

the registered voters group is likely to be more relevant in this high turnout election.

[West Coast Democrat]

12. They came to nearly the same results in 2000...

but gave Bush* 49% to Gore's 47% in their final 2000 poll (w/ Nader at 4%).

So according to Pew, Bush* is polling 1% lower than he was in 2000,

[NOWMDNOGWB]

17. KERRY + 5% IF WEIGHTED CORRECTLY!!!