We know how the 2000 Presidential race came out: it was close. Nothing has happened in the meantime to radically redraw the map. So the obvious way to handicap this year’s election is to assume that each party starts out with what it had last year, and ask about changes. For some reason, none of the polls I’ve seen have been reported that way, until the latest Zogby effort:
Mr. Kerry is showing a 2-to-1 lead (50% to 25%) amongst voters who didn’t vote in 2000, while winning three-quarters (75%) of Ralph Nader’s voters and stealing twice as many (8% to 4%) of Mr. Bush voters in 2000 than Bush is stealing of Gore voters in 2000.
That, plus some sort of effort to estimate the sizes of the various groups (and of the Bush/Gore vote among those who have died in the interim) ought to provide a reasonable guess about the outcome.
Of course, since this is an immediately post-Convention poll, you can’t take the current encouraging numbers too seriously. But my point is about method; I’d like to see all the pollsters report results in this format from here on out.