Gallup’s Editor-In-Chief tells us to stop asking impertinent questions, the polls are just fine.

Now if a given poll in Ohio in this election shows Obama with a 10-percentage-point lead, one should just ask, “How likely is it that Obama would be ahead by 10 points if he won by five points in 2008?” — forgetting party identification, which we assume is going to be higher for the Democratic Party if Obama is ahead, anyway.

In a column full of weak arguments (exit polls use different questions! party ID changes! we don’t ask 2008 affiliation!), this is a particularly egregious sin against logic. It assumes we can’t know more about the expected party ID breakdown on election day than we know about the expected election results, but of course that’s nonsense, in fact we know quite a bit more: Gallup has recorded a 38-point swing in enthusiasm (from D+26 to R+12) to the GOP since 2008, in swing states the Democrats have lost 800K registered partisans to 80K for the GOP, the net favorable/unfavorable view of the parties has reached an all-time high of R+3  (down from D+25 in 2008), and Rasmussen party affiliation has reached an all-time high of R+3.7 — higher even than in Sept 2004, when the actual election split that year was even.

So this doesn’t begin to explain why so many polls are using D/R/I splits that are so ridiculously pro-Democrat. Some have gone as far as D+16, even though this election appears be headed for a roughly even exit poll split. But have you ever seen an R+16 national poll? Have you ever even seen an R+6 poll? Of course not, that would have Mitt Romney up 10 or 20 points. Doesn’t fit The Narrative.

The problem isn’t that the polls are just inaccurate based on party ID, it’s that they’re strongly biased to Democrats, except for the occasional obvious agenda poll from Plainly Partisan Polling that tries to keep Akin in the Missouri Senate race in preference to a stronger GOP candidate.

And as the always-perspicacious Jay Cost has pointed out, there’s a bimodal distribution here, which means the error is not random.   I won’t claim to know how much of this is blatant partisanship (ahem, PPP) and how much is just honest error, pollster groupthink,  or a “fight the last war” tendency after 2008’s Dem wave, but the 2012 polls are clearly off-kilter in a fairly deliberate way.

So how to cut through the chaff?  A salutary (if overly optimistic) effort has been made to reweight polls to the Rasmussen affiliation numbers, but there’s a very simple rule to interpret national polls: whoever is winning independents is winning the election.  You can safely ignore the rest of the information in a national poll — while it’s just barely possible to very narrowly lose independents and win the election, as Bush did with 49% in 2004, a lead outside the MOE is a pretty strong indicator.