Steve Hayward points to this interesting nugget in Michael Barone’s latest column:

Interviewers can inject bias in the results. The late Warren Mitofsky, who conducted the 2004 NEP exit poll, went back and found that the greatest difference between actual results in exit poll precincts and the reports phoned in to NEP came where the interviewers were female graduate students — and almost all the discrepancies favored the Democrats.

Steve responds: “Not sure whether this says something about polling, or about female graduate students.” This is just a guess, but couldn’t it also say something about people who vote? Seems to me that Republican women and conservative men might tell young women they voted liberal when they didn’t for admittedly different reasons. Some small fraction of, say, married women might not want to seem “un-progressive” with young clipboard-carrying women who effuse progressiveness (we can debate another time whether young liberal female grad students have “tells” about their ideology even when not wearing their “keep your rosaries off my ovaries” buttons). Meanwhile, some men might be inclined to tell young attractive ladies what they want to hear. Not me, of course. But I hear rumors that this sort of thing happens.

Update: A great many readers have located the key flaw to my theory. For example, from a reader:

Jonah,
I don’t recall young female grad students uniformly being attractive—or even presentable (and that’s being kind). My recollection from the mid-80s is that the typical female grad student observed an embargo on makeup, cultivated ginsu-dulling forests of “personal” hair, owned any number of pairs of cable-guy workpants and flannel shirts, and seemingly had a relative in the chuck-a-boot business. Not that there’s anything wrong with that—