Wampum

The more I look at the electoral college map, the more I am convinced of the centrality of Wisconsin. Gore barely eeked out a win in 2000. Gore received 1,240,431 votes while George Bush received 1,235,035. Nader pulled in over 93,000 votes. Nader will again appear on the Wisconsin Presidential ballot.

Wisconsin's eleven electoral votes seem likely to be decisive. A new University of Minnesota poll has Bush +1 at 48%-47% in a three way race.

A CNN/USAT/Gallup poll has Bush ahead by 6 among likely voters and by 8 among registered voters. If that poll is accurate, then Kerry is in big trouble in Wisconsin.

As a result of those polls, Slate has moved Wisconsin into the Bush column saying that "he has taken Wisconsin outright."

What are the implicit turnout assumption of those polls? For the Gallup poll, which I suspect is the one driving Slate to move Wisconsin, I can not say. USA Today does not report the cross tabs or the party ID breakout of their sample.

If anyone has that information, please post a comment. Are you listening Steve Soto?

We can, however look at the turnout assumptions of the U of M poll because it does report the cross tabs. Its sample has Mr. Bush drawing 94% of Republicans, 5% of Democrats, and 35% of Independents in a three way race.

Mr. Kerry gets 92% of Democrats, 5% of Republicans, and 46% of indenpendents.

If we apply those cross tabs to the party identification of the 2000 exit polling from Wisconsin (37% Democrats, 32% Republicans, 31% Independents) we get the following:

Bush 42.8%

Kerry 49.9%

Wow, a one point Kerry deficit turns into a seven point Kerry lead if the 2004 party ID turnout mirrors 2000 party ID turnout.

It is apparent that the U of M poll has an implicit assumption that Republican turnout will be much better than it was in 2000 and that Democratic turnout will not keep pace.

Is that likely? I really could not say. Neither can anybody else as Wisconsin as same day registraton. The Seattle Times does report:

In Wisconsin ... voter registration is up by at least 200,000 this year, estimates Kevin Kennedy, the state elections chief. He says the rise is being driven by the high stakes and close race as much as by activist organizations.

"The municipal clerks are hearing from people who have been gone from the country for 20 years and they want to vote," Kennedy said.

Still, appraising the registration drives is a dicey business, cautions Curtis Gans, director of the Committee for the Study of the American Electorate.

Gans, noting that "registration doesn't necessarily speak to turnout," said voter rolls grew nationally in 1996 and 1998, largely because of motor voter laws, but turnout was down in both years. In 2000 and 2002, by contrast, registration was down but turnout was up.

Milwaukee County Executive Scott Walker, who many political observers suggest is quietly campaigning to become the Republican gubernatorial nominee in 2006, is not sending good signals about his respect for the voters of Wisconsin or this state's electoral system.

Aware of massive increases in voter registration in Milwaukee, and the intense interest in the presidential race among people of color, young people and others who might not be inclined to vote for George W. Bush, Walker pulled one of the most bizarre stunts ever seen in Wisconsin electoral history. When city of Milwaukee officials requested that more ballots be printed in order to accommodate the expected rush of new voters on Nov. 2, Walker refused.

So, registration is up in Milwaukee and interest is high among those unlikely to vote for George Bush but the reported sample has an implicit assumption that turnout will be much less favorable to Kerry than it was to Gore.

I have no idea what Wisconsin turnout will be but if the cross tabs from the U of M poll are remotely accurate, Kerry will win if turnout mirrors the 2000 election.

I will do a similar anaysis if someone can get a hold of the cross tabs from the Gallup survey.

Edited slightly.