I think the margin of common error (primarily due to cell-phone voters and the unkown efficacy of the field campaigns, espeically Obama's) is significanltly greater than McCain's margins in national polls.
Further, one of the reasons there have been so few elections where the popular vote went one way and the EC the other is that there have been relatively few elections that were very close in popular votes. A cross-outcome like that isn't likely, because chances are someobeody will win the popular vote by at least 3%. But within 3%, the chance becomes significant. The following graph is based on simulations run from June polling data, so it wouldn't look quote the same now. But the general spread is interesting.
I dunno, out of 4 extremely close, undisputed elections (1880, 1884, 1960, 1968) the EC and PV winners were different in only 1 of them. I'm not saying it can't happen this time, but I just wouldn't put must weight on state polls this far in advance when more national surveys are conducted than in individual states, and trend lines can be more easily detected. I'll look at electoral vote projections the week before Election Day, and worry about them then.
I understand the sources of error in these surveys, but I prefer not to assume their biased one way or another (baring clear evidence to the contrary). Doing so provides too many opportunities to discount unpleasant news.
I'm not discounting anything. I would say that if there's a large error, it's probably undercounting Obama's likely vote. But I'm not at all convinced there is such a large error. My point is that even relatively small errors are larger than McCain's reported margin, so we can't have justified confidence in suppositions of who would win, if the election were held now. At even money, I'd bet on McCain. But I don't think I'd give 3-2. And since we're still in the period that was expected to have a small, temporary bounce for McCain, David's Chicken Little routine is just silly.
no subject
Date: 2008-09-12 05:53 pm (UTC)Further, one of the reasons there have been so few elections where the popular vote went one way and the EC the other is that there have been relatively few elections that were very close in popular votes. A cross-outcome like that isn't likely, because chances are someobeody will win the popular vote by at least 3%. But within 3%, the chance becomes significant. The following graph is based on simulations run from June polling data, so it wouldn't look quote the same now. But the general spread is interesting.
no subject
Date: 2008-09-12 07:24 pm (UTC)I understand the sources of error in these surveys, but I prefer not to assume their biased one way or another (baring clear evidence to the contrary). Doing so provides too many opportunities to discount unpleasant news.
no subject
Date: 2008-09-12 08:26 pm (UTC)