More words of wisdom from Ruy on Gallup's screwy poll numbers.

[jefferson_dem]

If you just look at their results for registered voters, there has actually been very little change in the Gallup Poll results since early October. Bush had a 1 point lead then, he has a 3 point lead now. Not that different from what most other polls have been showing. Almost all of the change since their last poll is in the results for likely voters. Unfortunately, the likely voter number is the only one the media will focus on now.

So how do you go from a 3 point lead among registered voters to an 8 point lead among likely voters? By projecting that 89 percent of registered Bush supporters will vote but only 81 percent of registered Kerry supporters will vote. But as we know, this is totally unrealistic.

http://www.emergingdemocraticmajorityweblog.com/donkeyrising/

By the way, did anybody ever get the "official" word on partisan breakdown within their sample? How bad was the repuke oversampling?

[wryter2000]

1. I think that's clearly what's going on

In the recent past, Republicans have been more reliable voters than Democrats. All that has changed this year, though. With our "base" motivated as never before and with all the new folks we've registered, we're going to beat those number like crazy. In fact, I'd predict that in the near future, we're going to see Democrats over-represented in polls to reflect our new resolve to take our country back!

[FloridaPat]

4. Actually when they talked about how many came out in the last 4

elections, the Democrats got around 10% more people out. For some reason (gee, I wonder why) they insist that more repubs will come out this year.

[DemocratSinceBirth]

8. That's Not Entirely Correct

Republicans do better in mid terms than presidential elections because presidential elections motivate the casual voter who unfortunately is disproportionately Democratic...

[Guaranteed]

2. Yup, they ARE weighting the numbers with that poll.

http://www.usatoday.com/news/politicselections/nation/polls/usatodaypolls.htm#method

Of course, they don't give us the party affiliation numbers that they're using, but they're playing games with the turnout.

[tritsofme]

3. They don't weight for party ID

They take the top 55% of their sample who gets the highest score from their LV model and use that as their LV number.

[jefferson_dem]

5. Now that's just wrong - to filter out voters based on a projected overall

election day turnout number, regardless of how "likely" they are to actually vote based on their own indicators. No legit way a 3 point difference should inflate to 8 points.

It's also pretty damn clear that their RV starting point of a 3 point edge for chimpy was based on a disproportionate number of repukes in the first place.

GallupSux.

[unfrigginreal]

6. And that's where they miss the mark

their LV filter accounts for enthusiasm for a candidate but doesn't account for enthusiasm against a candidate. I suspect that I might not make their top 55% because I am not enamored of Kerry. But I damn well can assure you that NO Bush supporter is more likely to vote for him than I am to vote against him.

[mcg]

9. They overpoll republicans because their boss is a big GOP donor

The danger of the Gallup poll

<http://wilsonhellie.typepad.com/for_the_record/2004/09/the_danger_of_t.html>

Why You Should Ignore The Gallup Poll This Morning - And Maybe Other Gallup Polls As Well

<http://www.theleftcoaster.com/archives/002806.html>

"...the Gallup Poll, despite its reputation, assumes that this November 40% of those turning out to vote will be Republicans, and only 33% will be Democrat."

Well, they can go Cheney themselves.

You know, they did have one poll which had more democrats, and in that case, guess what, Kerry was ahead. I think they did that as a feign, to make it seem like they are unbiased and as a setup for this. Pretty slick.

[tritsofme]

10. Gallup doesn't assume anything in terms of party ID

They just run with what they get from their random sample for party ID.

[mcg]

11. Sep 13-15 Gallup poll likely voters 40% GOP 33% Dems

<http://toughenough.org/2004/09/gallup-exposed.html>

and there have others like that, e.g.
Sep 24-26 GOP: 328 (43%) Dem: 236 (31%)

<http://www.theleftcoaster.com/archives/002881.html>

I don't trust them one iota.

[Cha]

7. Well, gallop and his poll can just LOSE

their reputation Even more for reliability!!