Replies

Almost all of them were close or in the ballpark... I'm disappointed that I was given a false sense of assurance here... Next time, count me as a firm believer in the polls. LS, sorry bud, but the LS rule just can't ever be injected into any threads in '08... actually you can, but you'll be reminded of '06 time and again.

I do think most of the polls are really biased until the last week before the election. SurveyUSA is famous for doing that.
I did not get worried until the polls got worse in the end, instead of better.

You're absolutely right, and I said so on my "Losers and Winners." The pollsters were big, big winners. So until they show me they are off, I, like you, have to ditch the LS rule.

But we today have some final turnout numbers, and the numbers showed that unlike 2002 and 2004, the Dems turned out in higher percentages, and that in some states (including OH) our turnout was a little down.

That suggests that Rs, rather than voting for Ds, stayed home a little, and extra Ds came out over 2002. Given that, my model for predicting turnout, hence, poll error, was correct as far as it went. No one---certainly not pollsters---can anticipate "extra" turnout. Anyway, we were tracking turnout at 1:00 and it was really good, leading us to think that it would be a re-play of 2004. But all our people who were going to vote, voted early. So after 1:00, we got just 20% more of our voters, whereas in 2004, we got about 40% more.