…and it’s a dead heat.
P.S. If the “incumbent rule” is right, though, Bush is in trouble, in terms of the popular vote at least. He’s at 48% or below in all but two of the nine final national polls, and at 50% or higher in none.
As for the electoral vote, it’s all about turnout in the key states, but Kerry appears to have at least a halfway-decent shot in Florida, which is better than I expected a few days ago. Ohio is totally up in the air. I get the sense that Wisconsin and New Mexico are leaning Bush, while New Hampshire, Michigan and Minnesota are leaning Kerry, and Iowa is a total toss-up. I could be wrong, though.
Most likely tie-election scenario at the moment: Bush holds Florida, loses Ohio and New Hampshire, gains Wisconsin and New Mexico. Everything else stays the same as 2000. Electoral College deadlocks, 269-269.
Kerry Spot says the Republicans are rather upbeat.
It’s just too, too close to call.
P.P.S. In far more important news, the Mystery Pollster had a baby — on my birthday!!
P.P.P.S. The BrendanLoy.com community is currently predicting a 270-258 Bush win, with Wisconsin (10 EVs) tied.
|
Categories: Election 2004
|
November 1st, 2004 at 3:51:29 am
What happens to the aggregate totals if you throw out Charles prediction? Statistically it does appear to be a bit of an outlier and skews the results.
November 1st, 2004 at 3:56:58 am
The aggregate is based on the state-by-state results, not an average of the totals, so Charles’s predictions don’t have much of an effect. For example, he picks Bush to win almost every solid blue state, including places like Massachusetts, California, etc. But these predictions have no effect — as you can see, MA, CA, and all the other solid blue states are in the Kerry column in the aggregate map, indeed they are in the “more than 80% pick Kerry” column.
So the only effect Charles really has is, he is just another person picking Bush in virtually all the red and battleground states. His outlandish predictions don’t really figure in, however.
To answer your question more directly… the only affect that Charles’s removal from the sample would have, at the moment, would be to move Wisconsin from “tied” into Kerry column (10 out of 19 instead of 10 out of 20), so Bush would win 270-268.
November 1st, 2004 at 4:02:17 am
If you were to throw out both Charles’s and Toni’s maps — the two most pro-Bush in the contest — it would become 268-266 Kerry, because New Hampshire would move from the Bush column into the “tied” column.
But again, I don’t see that as justified; there is nothing statistically “outlier”-ish or invalid about their picks in those battleground states, and their crazy picks (no offense Toni & Charles :) in the blue states have absolutely zero effect.
This is precisely why I chose to use an aggregation method rather than an averaging method, though. If you were to simply average the 20 vote-total predictions, it’d be 287.5 to 250.5 Bush. That clearly would be skewed by the fact that we have two, ahem, exuberantly pro-Bush maps, and no equivalent pro-Kerry maps. By contrast, the state-by-state aggregation method (I don’t even know if “aggregation” is the right term, but I’m guessing it’s at least close since Mike hasn’t corrected me thus far :) doesn’t have this problem.
November 1st, 2004 at 12:26:12 pm
Hey, how can I see other people’s electoral college maps? That sounds cool. And uh, do we get a prize if we pick the best map?
November 1st, 2004 at 12:51:22 pm
I believe the winner, if her name is Becky, gets rewarded for having an accurate Electoral Map with lots of sex.
November 1st, 2004 at 1:07:12 pm
All the maps are in the original post here… just scroll down. :)
November 1st, 2004 at 1:08:24 pm
As for the prizes, I discussed that here.
November 1st, 2004 at 2:24:12 pm
Andrew - I had heard that the first prize was dinner with Brendan … and the second prize was a weekend with Brendan …
Was that just another Rather silly rumour ?
(innocent grin)
November 1st, 2004 at 6:29:25 pm
LOL! You Republicans are Hilarious today! :)