USA Today Gallup: Dems advantage on generic ballot narrows

[Sam Spade]

But in all fairness Mike they got they predicted that Bush would win and he did.

No they didn't. They predicted a 49%-49% tie in the popular vote and they predicted Kerry would carry Florida and Ohio while Bush handily carried Pennsylvania. This is why no one takes them that seriously any more.

As for these numbers, there may be some truth to them but they are already dated. This is Bush's mini 9/11 bounce. Rasmussen showed an uptick as well a few days ago (while this poll was taken) but as of today Bush's approval according to him is 40% with 58% disapproval.

Gallup was only wrong there because for some reason they assumed that the undecideds would break 2% for Kerry and 1% for Nader.

That was a serious smoking crack moment for Gallup.  Otherwise, the poll had shown Bush up 2, basically the final margin.

Comparing Gallup's crappy state polls to the national poll is flawed reasoning.  Please stop unless you wish to lose credibility.

Anyway, this poll is merely one data point in the string of points.  Take that for what it is.

[Mike in Maryland]

It should also be pointed out that the 48-48 tie among "likely voters" is not matched by registered voters, where Democrats have a 9 point lead (51-42). 

Gallup's likely voter screen in 2004 usually (not quite always) produced more pro-Republican results than its registered voter polls, as well as polls by other organizations. Deja vu all over again? 

[Amenhotep Bakari-Sellers]

Anyway they predicted a 274 winner no matter what states they got wrong it still was correct they predicted the winner. And they performed accuratly in 2000, polled it the exact numbers.

[ATFFL]

Generic ballot can swing wildly based on wording and position in the poll.   It is difficult to poll well and, even when you do, it is pretty meaningless.

[The Vorlon]

Summer polls just don't work.

The kids are back in school, the campaign has started. 

Wake me up when Mason-Dixon starts polling.

[Moooooo]

There is this common agreement that MD is the best, but Ive yet to see any concrete evidence why.  If I recall correctly, they were pretty far off in the 2005 VA Gubernatorial race.  When I was running down the Rasmussen state-by-state polls for 2004 the other day, they seemed to be pretty damn accurate.  What make MD that much better?

[The Vorlon]

Generic ballot can swing wildly based on wording and position in the poll.   It is difficult to poll well and, even when you do, it is pretty meaningless.

I believe that the Dems have been ahead of the GOP, by varying margins, in the Generic ballot for each of the last 5 election cycles. - shows you how well it works

You pretty much have to go through it race by race because, well, there are only 40 races that count at the House level, so 90% of those polled really don't matter.

[The Vorlon]

There is this common agreement that MD is the best, but Ive yet to see any concrete evidence why.  If I recall correctly, they were pretty far off in the 2005 VA Gubernatorial race.  When I was running down the Rasmussen state-by-state polls for 2004 the other day, they seemed to be pretty damn accurate.  What make MD that much better?

Rasmussen did indeed have a very good 2004 - he had a brutal 2000.

For the record, I think Rasmussen will crash and burn again in 2006 - but hey, that''s just an opinion

Everybody, including MD, has a bad poll now and then, but MD has fewer than most.

In 2000 they were almost dead on in every state (they got one state wrong)

In 2002, they were the most accurate pollster in the nation according to an anyalsis done by the National Council on Public Polling.

In 2004 the were basically dead on in ever state (Except Minnisota where they missed by 4.5%)

Everybody has a good year now and then.  MD seems to have a good year EVERY year....

One poll is, well, one poll.... but year in and out these guys have a record nobody else can really match....

[Moooooo]

All I needed to know.  I always hear they are the best and even repeated it myself a few times, but never really knew why...

[Eraserhead]

But in all fairness Mike they got they predicted that Bush would win and he did.

No they didn't. They predicted a 49%-49% tie in the popular vote and they predicted Kerry would carry Florida and Ohio while Bush handily carried Pennsylvania. This is why no one takes them that seriously any more.

As for these numbers, there may be some truth to them but they are already dated. This is Bush's mini 9/11 bounce. Rasmussen showed an uptick as well a few days ago (while this poll was taken) but as of today Bush's approval according to him is 40% with 58% disapproval.

Comparing Gallup's crappy state polls to the national poll is flawed reasoning.  Please stop unless you wish to lose credibility.

If it has Gallup's name on it I hold them responsible for it and expect better. If you don't want to, thats alright. I do.

[Wiz in Wis]

Generic ballot can swing wildly based on wording and position in the poll.   It is difficult to poll well and, even when you do, it is pretty meaningless.

^^^^^^^^^^^^^^

Props here, Generic Polls in September are as predictive as Tea Leaves and Michael Jacksons facial hair.  Look at Bush approval as a guide (albeit a weak one) before generic vote.

[Silent Hunter]

Welcome back, Vorlon. Haven't seen you for a while.

Let's see if we get some polls like this.

[RBH]

Yeah, that kind of sample was used for this Gallup poll?