Saturdays are typically slow for new statewide poll releases, but with Election Day just two weeks away, new surveys yesterday from PPP and Daily Kos/Research 2000 are worthy of an update. In particular, a new West Virginia survey from PPP helps move West Virginia and its five electoral votes back into the lean McCain column.
The new West Virginia poll from PPP is only the third there this month and one of only seven since the conventions. It shows the McCain-Palin ticket leading Obama and Biden by eight points (50% to 42%) and shifts our trend estimate nearly four points in McCain's direction, to a 49.5% to 45.3% lead, just enough to qualify for lean McCain status on our map.
The new Kos/Research 2000 poll in Maine showing the Democrats leading by 17 points (55% to 38%) also has a big impact on our trend estimate, bumping the Obama lead up nearly four points -- enough to shift Maine into "strong Obama" dark blue. Maine is also a state with very few recent surveys, just two in October.
Normally, our trend estimates are "small-c" conservative, tending to minimize the impact of any individual poll, especially if it deviates significantly from the overall trend. However, as I noted when a new poll shifted West Virginia into the toss-up category a month ago, the scenario in states like West Virginia and Maine -- very few polls, and a new poll that represents a big difference from the previous polls -- makes our trend estimate more sensitive. In the the West Virginia example, the trend line is now largely ignoring an outlier value of 42% for McCain from ARG in early October, and basing the recent McCain trend on McCain results of 50%, 49% and 50% on the most recent Rasmussen, InsiderAdvantage and PPP polls respectively.
Obama's lead on the national popular vote trend has narrowed since earlier in the week. It was down to 5.2% (49.1% to 43.9%) when I generated the table above, (although two new daily updates from Rasmussen and DailyKos, entered as I was writing, widened it slightly back to 5.4% -- 49.4% to 43.9%). The first table above, which compares individual poll results for the national trackers to their last non-overlapping samples from earlier in the week, shows a tightening margin on four of the six this week. As I did yesterday , I will try to post an update later today when we have the results from all six tracking polls that release on the weekend.
Follow Mark Blumenthal on Twitter: www.twitter.com/MysteryPollster