You;re thinking of the Bradley effect, and it’s been analyzed a bit – most statisticians (Nate SIlver being one) don’t seem to think it exists. Or ratehr, if it does it gets drowned out.

And remember, the margins of error are +/- 3-4 percent, usually. Florida was well within that range. (Also remember, a poll says it has a margin of 3 percent and you see a 45-48 number, that means candidate A could 42-48 and candidate B could be 45-51. And those two percentages don’t have to correlate at all, necessarily.

THe point isn’t whether one poll is off; it’s taking them as a whole. For polls to be so skewed Democratic that they were ALL showing a move towards Obama would require some kind of conspiracy among pollsters to what, make Obama look good? But since they don’t know people’s political affiliation before asking — it’s what you are trying to measure, after all — that doesn’t make a lot of sense. Especially as the results were pretty consistent even among Republican-leaning organizations.

All of that BEFORE coffee? Wow, I’m impressed.

The Rasmussen poll did have some impact on my feelings about Ohio, but also two of the other most recent polls (Gravis Marketing, and a newspaper poll) show a race closer than I thought it was.

We might dismiss the newspaper poll because it might not have as good polling techniques, and Rasmussen tends to show a more republican slant than other polls, and one from Gravis is not necessarily a trend…

Individually I would not worry about any one of them, but collectively they suggest a very close race — definitely within a margin of error.

Additionally, it is not clear to me if these are registered or likely voters — or what a likely voter is for this election. In 2008, young voters turned out in droves to vote for Obama. This time, not so much. If the models for likely voters are off by 2%, that could alter the election.

Add to that the issue of under sampling — many groups are simply not represented by polls (such as those with cell phones, and those who have odd work schedules). In a ‘normal’ election their ultimate choices can be modeled well by the choices of other, similar demographic groups. But in this election I am no longer sure those assumptions are correct.

I’m not sure more polls of Ohio would correct either of these problems — there is a distinct possibility that most (or all) of the polls are off by as much as 5%. Exit polls are usually accurate, but that has to wait until election day.

The only hard piece of data we have is that early voting has strongly skewed democratic, as it usually does. But even that does not tell us which way they voted. A disillusioned democrat would likely be as likely to vote early — for Romney — as a loyal supporter of the party (or of Obama).

So… I think it is close. I also think Florida and North Carolina are becoming much closer than I expected. I would have placed them in the Romney camp, but I think it really depends on turnout rather than polling. Obama may have fallen flat in the first debate, and Libya is surely on the minds of the republican base, but the middle voters don’t really remember that far. If Obama shows excellent organization and leadership dealing with Sandy, it might make those middle voters ambivalent again… then again, maybe not. If I were betting, I’d bet like you on those states.

I think it’s funny that you’re from Wisconsin — because I am, too (north central). I’m a transplant of ten years, so I came in after many of the extreme shifts in voting. I don’t agree that it’s a close state — I think instead that Wisconsin voters get very excited by ideology rather than party. Rather than being democratic or republican, I think the state is attracted to shiny new ideas — something akin to a moth to a flame — because the new idea is interesting. Maybe Wisconsonites like political theory too much (if we two are any indication, that much is certain!) — but I think they embrace New more than D or R.

With that in mind, we have an issue. Obama is no longer new. And Romney hasn’t given much of ANY ideology — or, when he does, he later says something else. So… we have a case where the voters aren’t enamored by either candidate. On the other hand, Ryan is new — he appears to be a genuine believer in his ideals, which he was unafraid to share (even when they are unpopular). That sort of passion is far more likely to excite voters in Wisconsin, I think, than party politics.

That said, Ryan is not the top of the ticket, and he has acknowledged that he will govern as Romney’s subordinate rather than the driving force. So maybe we are back to neither candidate driving home a win.

It is interesting that Minnesota and Wisconsin are so different politically. I lived in the Twin Cities for 5 years, and that state was as blue as lake water (at least it was ten years ago). And south and east of Wisconsin are two strongly blue states, too. What makes Wisconsin different?

Is it the cheese curds?

You know, markc, being hostile instead of polite is not always the best way to live your life. I did not make anything up. In fact, the article YOU CITE states that we have had “…more than a month of media narratives about the advantage Democrats are getting in the early vote…”

So, clearly, I had reason to believe it was true. It is pretty sad that you accuse me instead of merely pointing out you believe I was wrong.

If the new Gallup poll shows that to be wrong, then so be it. Gallup is often correct (though I notice you have not mentioned that their unemployment stats have shown steady, even dramatic improvement in the last few weeks… odd that you noticed the one and not the other).

For the past few elections (Since the 2000 debacle) I have heard allegations like this coming from both sides… potential (for) fraud. Not a whole lot has ever come of them. That could be for a few reasons… the ones perpetrating the fraud aren’t doing it well enough (or are doing it too well), the allegations are baseless, or the potential for fraud isn’t being exploited. I have no problem with people pointing out these potential problem areas, but I’m also a bit wary of them unless something really odd happens, like you said: all the exit polls show strong leads for Obama, but Romney manages to eke out a win.

That said, electronic voting still doesn’t seem secure to me. With each new added technology or bit of security, someone points out a flaw. It make me glad that my state still uses paper ballots. Paper doesn’t malfunction/can’t be hacked. The voters might malfunction, the the ballots don’t