Kerry holds a 49-47 (rv) lead in Gallup poll when weighted correctly.

[59millionmorons]

1. The Gallup RV data, while preferable to their LV data, may themselves be flawed. Their RV sample has a 2 point edge for the Republicans in party ID. If that sample is re-weighted to conform to the 2000 exit poll party ID distribution, Kerry leads Bush by 2 points, 49-47.

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

[liberalpragmatist]

1. What're the current Harris Poll Party Reg. Numbers?

In Feb. 2004, the latest I could find, it should be waited 33% Dem, 28% Repub, 24% independent.

I also wouldn't discount the fact that W's probably gonna get a larger number of self-professing "Democrats" than Kerry will get Republicans. Rockefeller Republicans have long since stopped calling themselves Republicans - now they call themselves independents if not Dems. But in the South, esp., plenty of dixiecrats continue to call themselves Democrats but vote Republican in every national race (and most state races too).

Weight the polls with those numbers and see what you get.

[dolstein]

2. Weighed correctly? What a load of crap.

I have a simple question for you -- did the 2000 exit poll party ID distribution conform EXACTLY to the 1996 exit poll party ID distribution? Indeed, has the partisan composition of electorate EVER remained unchanged from one general election to the next? I seriously doubt it. The suggestion that in order to correctly weight the sample you must use the party ID distribution from the last general election is almost certainly wrong.

Does that mean that the Gallup organization's sample isn't skewed? Not necessarily. It is possible that Democrats are underrepresented, either because they are more likely to use cell phones or are less likely to have someone at home full-time. But the suggestion that the Gallup polls are wrong because their partisan composition differs from the partisan composition of the 2000 electorate is, to put it bluntly, moronic. The partisan composition of the electorate IS NOT STATIC. If it were, we would never have seen the New Deal realignment, or the breakdown of the New Deal coalition during the Reagan era.

[KittyWampus]

3. You Obviously Have An Agenda. How Many Fucking Times Do DU'ers

have to rebute your crap?

[dolstein]

7. Try doing it just once

This post suggests that the poll is wrong unless the sample is weighted to reflect the partisan composition of the 2000 electorate. This is JUST PLAIN WRONG. There is NO REASON to assume that the partisan composition of the 2004 electorate will be IDENTICAL to that of the 2000 electorate.

There are all sorts of arguments once could make as to why the Gallup sample is skewed one way or another. But the argument that the sample is skewed because it's partisan composition differs from that of the 2000 electorate is specious. You can't simply ASSUME that the elecorate will remain unchanged. While one poster below has noted that the partisan composition of the electorate changed very little from 1996 to 2000, I don't think you can simply rule out any change from 2000 to 2004, especially with a Republican in the White House.

[jefferson_dem]

4. Survey researchers have two choices: to weight or not to weight

based on partisan affiliation. Given this choice, the greater the load of crap resides with the latter.

If one's goal with a sample is to accurately reflect the voting population, why not make the best effort to do that rather than rely on the convenient position of a "fluid" electorate which bears absolutely no resemblance to the modern political era? Can you point to any election since the New Deal when more self-proclaimed Repukes have turned out than Dems? In fact, recent pre-election season polls have clearly shown that Dem identifiers still hold a 5-10% point edge. By the way, the New Deal realignment was reflected not in survey responses but on election day, 1932. If, heaven forbid, the Repukes overtake all the governing apparatus in such a landslide on Nov 2, then we'll talk realignment.

[DemFromMem]

5. Actually, party ID DID conform EXACTLY from 1996 to 2000

You may have a point that 2004 may not exactly look like 2000, but don't assume that polling organizations can't manipulate the numbers. When you look at the methodology of the various polls out there, you're seeing wildly different numbers in how they're being weighted for Dems and Reps. Party identification evolves. It does not swing wildly from week to week.

You asked a question of whether the 2000 numbers conformed exactly to 2000. You can actually see the numbers for 1996 at http://www.cnn.com/ALLPOLITICS/1996/elections/natl.exit.poll/index1.html and 2000 at http://www.cnn.com/ELECTION/2000/epolls/US/P000.html.

In 2000, Democrats represented 39%, Reps 35% and independents 27%.

In 1996, Democrats represented 39%, Reps 35% and independents 26%

That's amazingly stable.

If there has been a radical shift in party affiliation, I'd be really surprised and the burden should really be on the pollsters to present evidence of such a change.

[dolstein]

8. Oh waiter! One order of crow, please.

Of course, the stability of the partisan composition of the electorate from 1996 to 2000 doesn't mean that the partisan composition has remained equally stable from 2000 to 2004. Luckily, in two weeks this will all be academic.

[Weembo]

12. Trends

As you say, the proof is in the pudding. Even Realclearpolitics, the RW rag I snoop through to try to figure out the zombies, has dropped Bushes "sure things" from 284 EV to 227 EV in the last couple of weeks.

[IIgnoreNobody]

6. They won't be able to cook the books as easily on November 2nd

[okieinpain]

9. you can tell the polls are tight because none of the headlines

show numbers.

[BootinUp]

10. I didn't think they reported registered party numbers

The gallup question is which party do you identify with, not which are you registered with. So while I think some correction is needed its probably doesn't need to be quite as pronounced.

IOW, some of the people who identify with repukes are actually registered dems.

[AP]

11. The assumption is more republicans are voting this year...

...which conforms with polling. Republicans vote more regularly and are more interested in the race this year.

Why? Because Democrats are running against Bush, which isn't a good enough reason to go to the poles when compared with being motivated by the idea that you're defending your guy from unfair attack.

Bottom line: Dems have to get out the vote (precinct walk and phone bank!!!) and have to remind Democrats what they're voting FOR as much if not more than they tell people what they're voting against.

[George_S]

13. We may be surprised at the dem turnout. Bet we will be.

[AP]

14. I'm taking nothing for granted.