this was originally written as a comment on the thread about the Braley campaign's response to the DMR poll. Since there is another diary that mirrors my own thinking, I thought I would just throw it up here.
1. Remember polls are samples, and what you are usually given is a margin of error (eg +/- 3.5%) with a 95% confidence interval. That means that 19 times out of 20 the results will be within the margin of error of the poll, but one time in twenty it will not, even with a great pollster. IT is simply a reality of sampling, and a course in basic statistics should make that clear.
2. What the Braley response does not include is that among those in IA campaigning for him this weekend is Secretary of Agriculture Tom Vilsack, former two-term Governor, who still remains personally popular and can give Braley some cover on farm issues - assuming the race is close, which is what other polls seem to indicate, that could be enough of a boost to carry Braley to the win.
I have not yet written off IA in my modeling. So let me offer a more complete review.
1. There were 11 Democratic seats considered in jeopardy. I consider two lost -- WV & MT - and four off the table for us - NH, MI, NM, NC. In that modeling there are 5 seats left, of which the Republicans have to allnk five while losing NONE of the three of theirs (KY, GA, KS) that are in some jeopardy.
2. I tend to think we keep CO. Matt Barreto of Latino Decisions has an email/post out today where he talks about how most of the polling in CO is not accurately representing the Hispanic vote. For example, one poll had 50% of the Hispanics included as having college degrees, when the national average of adults is 30% and for Hispanics more like 15%. Others had the Hispanic vote evenly split or even had Gardner ahead. That is simply not going to happen. These are similar kinds of problems to what was seen in NV between Engle and Reid. BTW, only two pollsters got that race right - Latino Decisions and Mark Mellman, and the latter still had CO as a tossup.
3. Even if we consider CO a tossup, if I am write about the first six states, the two we lose and the four we keep, we need only ONE of the following to guarantee control:
CO, AK, GA, KY, KS, AR, LA, SD, IA.
4. I tend to think SD is lost. Pressler has faded badly and is not taking enough votes from Rounds to keep it competitive. We will not know about LA until the guaranteed runoff, and it is possible that Landrieu could win. I make it a one in three chance that Nunn will hit 50%, which would pretty much cinch things. I think AK may be 50-50, particularly if one considers the synergy with the Governor's race where people are going to split from the current Republican Governor. KY and AR are hard sells, and I expect to lose them. Assuming Orman wins KS, he is far more likely to caucus with Dems than with Rs, I think even if he holds the balance of power. And if in that election Repubs lose the Governor's race and the Secretary of State race, I am sure he will caucus with Dems unless Repubs have 51 other seats, and I do not see Republicans likely to get to 51 without him - they would have to hold GA and KY while winning four of SD (likely), AR (likely), CO (tossup), AK (tossup), LA (your guess is as good as mine about how a runoff if control is in the wind will play out) and CO (which my gut is we are going to win.
5. In my own thinking , I am still inclined to think we win CO and still think we may win IA, in which case we are guaranteed 50-50 regardless of what happens in the three Republican seats. If we win either, or win AK, or win GA, then we do not need KS to get to 50-50. As long as we have 49 without runoffs (guaranteed in LA and possible in GA) I think Orman commits to Dems to give us 50, if he wins.
So I remain optimistic.
How about you?
