> In my messy room, I'll see if I can find the 5 pg. or so memo
>from the Field Poll, that compares the projection they made in
>each Governors race and Presidential election since 1948. Right
>every time, and within a tenth or two of a percent in all the
>races, even '92 when Perot got 19%.And from asking about a thousand
>folks that reflected the demographic totality.
Sample sizes in serious polls are large enough to be representative, so that critique is wrong. But the answers you get aren't always clear. Asking someone which candidate they're going to vote for is a bit different from asking whether they're "for" or "against" a war. A vote is unambiguous - you vote for X or Y or neither or don't vote at all. With a war, people may hold contradictory opinions while saying they support it - because that's the "right" patriotic thing to say. But if you probe, or offer alternatives, you may get a more complex answer. Surely you've had this experience doing polls, Michael?
Doug