...this story isn't good news for John Kerry:

Despite weeks of bad war news — including a rising casualty toll and increasing criticism of the administration’s management of the war from Democrats in Congress — the presidential race in Michigan remains a toss-up.

Bush is in a statistical dead heat with presumptive Democratic presidential nominee John Kerry, with Bush favored, 44 percent to 40 percent, over the Massachusetts senator, the survey found.

“If I were John Kerry, I would be concerned,” pollster Steve Mitchell said. “The race is tough for Kerry. Voters may disagree with Bush on some of his tactics, but he is perceived as being strong.”

If Bush is even in Michigan (which he lost to Gore by about four points in 2000), the Dems are in big trouble.

The article doesn't suggest one of the possible reasons why. Southeast Michigan has the highest population of Iraqi-Americans (one of whom is my Baghdad-born sister-in-law) in the country.

In years past, undecideds have tended to break against the incumbent, so Kerry might gain some solace in that fact. However, the economy continues to improve. There was a very big jump in consumer confidence this month. These things work against Kerry.

I've been hearing that the "undecideds tend to vote against the incumbent", but I really don't know what to make of it, and I'm increasingly feeling like the statement has no practical value, except perhaps for propaganda value.

After all, at the time voters pull the curtain on the voting booth, how many of them are "undecided"? Zero, for all practical purposes. So, what if a majority of zero vote for the challenger on election day? How much does it help the challenger?

(I'm willing to entertain a claim that 3 out of 4 "undecideds" voted for Mondale in 1984. Yet, Reagan won 49 states. What was Mondale's consolation? "Sure, I only won one state, but at least I took a majority of the 'undecideds'!")

The whole thing depends on how you measure "undecideds", and whether you are even able to track their changing opinions longitudinally, I guess.

Sure, three months prior to the election, a majority of the undecideds may be considering the challenger. As the voters drive to the polls on election day, however, how many of them are still "undecided"? Is there any reliable data on that?

The whole claim seems to lack a foundation based on controlled observation.

---Nally

The timing of being "undecided" is certainly a question. I agree. But the idea of breaking against the incumbent rests on the reasonable proposition that the incumbent has had 4 years to convince the undecided electorate to be in his column and has so far failed to do so. This makes sense to me.

The failure of Binary Thinking.

There is a third option for undecideds, stay home

Hell, there is a fourth, vote nader or other third party as well.