Live from Cincinnati: It's Tuesday afternoon! Is that quick and bloggy enough for the TPM Café? Well, here's my not so bloggy beginning, with a thought to loosening up later in the week once I get my first, coherent two cents on the cyber-table.

Yes, the standard vote-choice question developed by Gallup and imitated by countless other pollsters certainly does minimize the "undecideds", especially during the earlier time periods preceding the election, such as the summer. So David's post and critique of the standard practice in the field (elaborated in his book, which I just finished reading last week) is largely on target, but I don't think David goes quite far enough in digging into a even bigger issue: the meaning of saying you're "undecided" or that you "haven't made up your mind" or "could change your mind before Election Day".

Mark's right: "A preference is often not a final decision." preference is often not a final decision." And yes, Nancy, question wording may just be the tip of the iceberg for all the reasons you have enumerated, particularly the likely voter screens and turnout models, a lot of which appears to be proprietary and therefore unavailable for us un-insiders to scrutinize.

But I'm thinking of a deeper iceberg: What exactly does it mean when someone says they're "undecided", that they "might change their mind", or that they "lean" this way or that way? And do these response categories mean the same thing in late October or early November, when voters' preferences are presumably firmer, as they do in June, July, August or September? I would wager that an in-depth cognitive interview probing the meaning of these response categories would turn up a lot of apples, oranges, and pears sitting underneath the exact same response category chosen by different respondents, and that this would vary over the campaign period as well. This issue strikes me as reminiscent of the problem with middle response and don't know/not sure categories in survey questions. If a respondent says he or she is "undecided", for example, it might mean they're genuinely ambivalent and torn between the candidates; or they want to avoid committing himself or herself publicly to an interviewer asking about their vote preferences; so the safe option is "undecided" or "might change my mind". Or perhaps they're consciously concealing their vote preference and just don't want to tell us who they're voting for (a place for the "Bradley Effect" or "Reverse Bradley Effect" to hide?). Or maybe they just haven't thought aboutthe election that much yet, so in this case, saying you're "undecided" represents another form of a "don't know" or indifferent response, especially during the summer or spring months preceding the election.

So David's suggestion for a new vote question, along with Nancy's and Mark's ideas, will certainly help us get a better handle on how the uncertainty of voters' preferences fluctuate across the course of the campaign and how we can go about getting better info on what the pollsters actually do in the dark, proprietary recesses of their operations--fuggettabout getting at the "truth"! But I don't think we have a very good grasp, if any, short of some well-designed cognitive interviews, at what it actually means to be uncertain about who to vote for, now, and how that differs from being uncertain last summer, winter or spring? There's lots of ambiguity sitting in the standard question and in David's new question too. So we have lots and lots to think about before we sleep...more later.

George Bishop is a professor of Political Science at the University of Cincinnati.