Reply #14: oh, hey, how about that

[OnTheOtherHand]

foo once got you to say something substantive, sort of:

Have you given any thought to what's wrong with using the Panel Study OTOH offered as an example of "false recall"?

First, let's be clear: it doesn't matter whether we construe it as "false recall" or not. The question is whether people report their past votes accurately, and if not, whether there is a somewhat predictable bias to their misreporting.

The evidence -- from several different sources including that panel study -- is strong that misreporting occurs, and that people are more likely to misreport having voted for rather than against the previous winner. So I remain curious why you have nothing to say about it. Or maybe that's the reason. Let's see.

ETA: Well, heck, while we're at it... you asked, "Do you really find plausible a position that 1 out of 7 Gore-2000 voters defected to Bush in 2004 (possibly OTOH's most absurd reconciliation of the 2004 recorded vote count and Bush's 3M mandate)?" Well, it's not my "position" -- we don't have enough evidence to know the actual proportion. However, we do know that in the 2000-2004 NES panel study, in fact 14% of people who said in 2000 that they voted for Gore, said in 2004 that they voted for Bush. So I'm curious about the basis of your confidence that that could not possibly be true for the electorate as a whole.

Bottom line: I'm curious whether you believe yourself, and if so, why.