JMHO

The question is why an established polling firm like Gallup would allow its credibility to be trashed by leaving the problem unaddressed?

...why should they feel any need to change? At this point, for most average Americans, "Gallup" is synonymous with "polls" in the same way that "Coca-Cola" is synonymous with "soft drinks." And you saw what happened with "New Coke."

I have no doubt that, if a smaller, less-well-known pollster was experiencing such problems with disparity of results compared to the norm over several cycles, they'd have made changes to their methodology long ago. But, as long as the nightly news will quote Gallup as the "gold standard," why change? If they start losing their hold on the conventional wisdom that they remain the "pollster of record" in U.S. politics, then you'll see them making changes -- not before.

...as Nate points out, their RV numbers are also out of the norm in Romney's favor. Also, Gallup has been prone to extreme swings -- larger than in other comparable polls run at the same time -- over a short period of time, something which a week-long tracker is supposed to minimize, which it clearly doesn't seem to be doing.

What is interesting about Gallup is that they run a seven-day race tracker and a three-day Obama job-approval tracker, so the two of them diverge on polling periods. The job-approval number has really been all over the map -- it was +11 for the days of the rebound a few days following the first debate Romney bounce, sank to merely +1 over the weekend, then began a climb back up to +6, where it is now. That suggests that there were some truly bad polling days for Obama late last week, and that part of what we're seeing here might be the earlier, positive Obama days of the rebound dropping off while these bad days remain. It also suggests, based on the climb in the short-term job-approval tracker, that we may well see Obama's numbers in this poll climb dramatically between now and the beginning of next week -- and won't that be fun to tweak the wingnuts about!

Let us look at the 2004 election when Kerry was leading Ohio and then just after 11pm on election night Ohio said their servers got overloaded with queries and they had to transfer the electronic election results to the Republican affiliated company in Tennessee that made the voting machines, SMARTech, then SMARTech would send the data back to Ohio. IT analysts that examined how the SMARTech software works determined it was a "middle man" system where SMARTech would have the ability to alter the information before sending the data back to the Ohio Secretary of State's office. Somehow during that transfer Kerry went from winning to losing. The exit polls had Kerry winning by almost 7%.

Guess what...current Ohio Secretary of State Jon Husted has decided to use the SMARTech machines and software for this election as well. He also hired IT group Triad which has Republican ties as well. So what needs to happen for this to work? Simple, Karl Rove, Husted, and Romney's big money connections need to get the left leaning polls such as Rassmussen, PPP, and Gallup on board to give the appearance of a tight race, hence why the GOP is pushing the term 'horse race.' Notice how the right leaning pollsters have started to release polls on the same days with each other almost lock-step. That would make sense in regards to these crazy outlier polls.

PLEASE email SMARTech, Triad, Jon Husted, and your local Senators to let them know you are onto any possible manipulation and that the word is getting around about SMARTech/Triad/Husted. Most people won't try something illegal if they feel people are onto any possible schemes.

But with Democratic turnout already even higher than it was in '08, it'll be even tougher for them. And even more people are fired up against Republican fraudsters, too. So it makes our job a little less challenging.

I agree...high Dem turnout is a big help. But the goal is to write SMARTech, Triad, and Husted and let them know that many people are aware of this possible scenario. All I ask is for people to write them in a civilized manner putting these parties on notice.

but thats just me.

Here's some stats, btw:

http://elections.gmu.edu/early_vote_2012.html

I do not like to whine about polling per se, but their numbers have been interesting. It will interesting to see what their numbers will be a week from now.