ARG reports Kerry lead over Bush in Michigan as a tie.

[Feanorcurufinwe]

To be fair, they are right, the lead is statistically insignificant. Keeps this in mind when Bill Schneider tells you a move of 2 pts means people are deserting Kerry.

John Kerry and George W. Bush are tied among likely voters in Michigan according to a survey by the American Research Group. A total of 48% of likely voters say they would vote for Kerry if the presidential election were being held today and 45% say they would vote for Bush. A total of 1% of likely voters say they would vote for Ralph Nader and 6% of likely voters are undecided.

In a race between just Bush and Kerry, Kerry is at 49% and Bush is at 46%, with 5% undecided.
http://www.americanresearchgroup.com/mi

[Bill McBlueState]

1. big fan of the

ballot lead calculator

thanks for the link!

[Doosh]

2. damn

[0007]

3. Michigan will bring it home for John Kerry, not a doubt in my mind

[nodictators]

4. Actually, with these 3% leads, Kerry is more likely to win than Bush

ARG is correct to say that Kerry's 3 percentage-point lead is NOT statistically significant at the 95% confidence level. But it might be statistically significant, for instance, at the 75% confidence level.

In other words, at this time the probability of Kerry winning Michigan is considerably higher than the chance of Bush winning there.

[TWiley]

5. That will SOON change as

I will be manning a phone line at the local democratic headquarters. I plan to work my ass off down there.

[fujiyama]

6. While ARG

has shown Kerry to have slipped since he chose Edwards as the nominee (he was up 7 then), other polls have shown him with the same lead of around 6 or 7 points, including Strategic Vision, a republican firm.

Kerry will win MI and PA, barring a Bush landslide victory. The only thing that can get Bush a landslide victory is if he captures Osama. SwiftLiars aren't enough to win him these two states.

[VirginiaDem]

7. "Statistically insignificant" is "significantly misleading"...

and it's sloppy shorthand for "Bush might be leading". Assuming the margin of error is 4 (just grabbing a number for the sake of argument since it's not listed here) and the confidence level is 95 percent then there's a 95 percent chance that Kerry is leading by negative one to seven points. Notice that seven out of the nine possibilities (using only whole numbers for simplicity sake) are of a Kerry lead. Under the same above assumptions, unless I'm mistaken, there's a 68 percent chance that Kerry leads by between one and five. This is because going out two standard deviations gets you a 95 percent confidence level while going out to just one standard deviation gets you 68. One standard deviation is half of two so you just divide the moe by 2. Of course, nobody does this in reality because 95 is so much more solid than 68 but it gives you a common sense look at what's being estimated.

[yellowcanine]

8. That would be true if it were only one poll. But the fact that several

polls are saying the same thing - that Kerry leads Bush - gives more confidence it is an accurate assessment of opinion at this moment, imo.

[VirginiaDem]

9. Agreed

We can be pretty sure that we're up by a few points at least in MI. Unless most of the polls that have us up are somehow using bad methodology, which is unlikely but there's really no way of knowing for sure until the votes are cast. Here's hoping...