|KY||Survey USA||3/15/08||McCain +36||McCain +10|
|MO||Survey USA||3/15/08||McCain +14||McCain +2|
|NH||Rasmussen||3/16/08||McCain +3||McCain +6|
|OH||Survey USA||3/15/08||McCain +7||Clinton +6|
|OH||PPP||3/16/08||McCain +8||Clinton +1|
The polls in Ohio, where both Survey USA and PPP now show McCain losing to Clinton but beating Obama, and to a lesser extent Missouri, would tend to be the biggest concerns for Obama. It's clear that the Jeremiah Wright controversy has had at least some near-term effect on Obama's trial heat numbers. Essentially all of these polls were taken over the weekend, including on Friday, which according to Rasmussen was the worst single day of polling for Obama (in his head-to-heads against Clinton) since the primary campaign began.
I have already blogged about the Wright "anti-bounce". On the one hand, the fivethirtyeight.com polling averages are designed to be forward-looking. If you're going to make electoral projections this far out, you'll find you'll do better to look at a wider cross-section of data, rather than just the polls that came out in the last week; history tells us you'll make a better prediction that way. So, Obama's new, "bad" polling numbers are a part of his reality -- but only a part of his reality. On the other hand, any exercise based on polling is inherently backward-looking to a certain extent, as we are at the mercy of awaiting new polling data. There may be more bad polling data coming down the pike for Obama -- or there may not be as the Wright controversy fades.
For the time being, I think we can say there has been a measurable effect, especially in white, working class states like Ohio. These are to McCain's ultimate benefit, rather than really to Clinton's., who has little to brag about in these numbers.
How long will the effect last? That we cannot say yet.
Obama's overall win probability against McCain is now 51.4%; Clinton's is 39.0%.