A few states moved in the past two weeks. Movement was to be expected, of course, since we’re narrowing the zones for “Tossup” and “Leans”. But that’s exactly why this edition is such a surprise…we had a couple states move into “Tossup”.
Here is the current map:
Now for the details.
As always, “Continuing” refers to the seats in Senate Classes 2 and 3, which are not up for election this cycle.
Here the highlights of the past two weeks, walking from the Pacific to the Atlantic:
- Hawaii: Why do I list Hawaii? Because there’s an ever-growing drumbeat suggesting that former Governor Linda Lingle has a shot at beating Representative Mazie Hirono (D-Honolulu) in the general election to fill the seat being vacated by Senator Daniel Akaka. Polling hasn’t lent much support to the claims of a close race; Hirono’s lead has ranged from five to 22 points over the past 18 (!) months, with no readily discernable trend. The last poll, conducted in July by the Honolulu Star-Advertiser, gave Hirono a 19-point lead in a “likely voter” model. Real Clear Politics calls Hawaii a “Leans Democrat” state, but I feel pretty confident from the polling data that it’s “Likely Democrat”.
- Nevada: Recent polls from the Las Vegas Review-Journal/SurveyUSA and Public Policy Polling make the Silver State look ever more certain to keep incumbent Senator Dean Heller. His lead over Representative Shelley Berkley (D-Las Vegas) is five points in the former, and an unadjusted two (adjusted five) in the latter. Five points is a lot to ask this close to the election, particularly since Heller has led the race since its inception and there has yet to be a trend of narrowing. For those reasons, I’m moving Nevada to “Likely Republican”.
- Arizona: We now know that Representative Jeff Flake (R-Mesa) will be facing off against former Surgeon General Richard Carmona in November. Flake has polled fairly well against Carmona, but the polling has been light and infrequent. I’m torn between keeping Arizona at “Leans Republican” and moving it to “Likely Republican”. There have been hints of a narrowing trend, but it’s hard to distinguish between signal and noise here. For now I’m leaving it where it is, since we have had no polls since July, but my gut is telling me that it’s probably time to move it.
- Montana: One new poll from Rasmussen gives Representative Denny Rehberg (R) an unadjusted four-point lead over incumbent Senator Jon Tester. Rasmussen has been the only firm to poll Montana since April, and the amount of noise in his polls in other states means that we have large error bars here. I just don’t have enough confidence to move Montana out of the “Tossup” column.
- New Mexico: Rasmussen is here again, this time showing Representative Martin Heinrich (D-Albuquerque) ahead of his predecessor, former Representative Heather Wilson (R-Albuquerque) by seven points. The number is consistent with every poll since April. Wilson has yet to even tie a single poll, let alone show a lead. The only thing keeping New Mexico out of the “Likely Democrat” column is the low frequency of polling, coupled with the high noise of the firms conducting the New Mexico polls. It’s still a “Leans Democrat” state, but if we don’t see a poll with a Wilson lead by October 1, I will move it to “Likely Democrat”.
- Missouri: You know you’re in trouble when your party abandons you and begs you to drop out of the race. This is the fate of Representative Todd Akin (R-Wildwood), whose intemperate comments regarding rape and pregnancy have left him deeply unpopular with all but the hardest of hard-line Republicans. While Akin had a lead over Democratic incumbent Claire McCaskill in every post-February poll before he made his fateful gaffe, he has not done well in the polls since then. Rasmussen saw a ten-point unadjusted lead (11 adjusted) for McCaskill, while The Post-Dispatch/Mason-Dixon came away with a nine-point lead and Public Policy Polling a one-point unadjusted lead (though that adjusts to two points in favor of Akin). PPP could be an outlier, but it was conducted much later than the others — on Day 3 of the Republican National Convention — so we might also be seeing a convention bump here as well. Intrade, in one of the few individual Senate markets with enough activity to mention, has McCaskill with a 62 percent chance of reëlection. Because of the PPP poll, I’m not ready to go so far as to call Missouri “Leans Democrat” (though Real Clear Politics is). I am, however, confident enough to move it to “Tossup”, pending another round of polls.
- Wisconsin: Former Governor Tommy Thompson is polling well against Representative Tammy Baldwin (D-Madison). With the primary behind us, CBS/New York Times/Quinnipiac’s six-point unadjusted (four adjusted) lead, Marquette University’s nine-point lead, Public Policy Polling’s five-point unadjusted (eight adjusted) lead, and Rasmussen’s 11-point unadjusted (ten adjusted) lead, all in Thompson’s favor, things look mighty good for this to be a flipped seat. I’m leaving Wisconsin in the “Leans Republican” column for one more round, but only because I’m not yet sure if this is a bump or a genuine base.
- Michigan: Three polls in the past two weeks, all of which have been a surprise to me. Both Mitchell Research and Baydoun/Foster show former Representative Pete Hoekstra with a lead over incumbent Senator Debbie Stabenow, by one and two points, respectively. The Detroit News, on the other hand, gave Stabenow an eight-point lead. The Detroit News poll is in line with the historical trend, but it’s unusual (though hardly unprecedented) to have two outliers be that close to each other. When I get data that are that much in conflict, I sit and wait for more before reacting. I’m leaving Michigan as “Likely Democratic”, but frankly I’m baffled by this latest set of numbers.
- Ohio: Three new polls this time, from The Columbus Dispatch, Ohio Poll/University of Cincinnati, and CBS/New York Times/Quinnipiac. The Dispatch poll indicated a tie, and the Ohio Poll gave incumbent Senator Sherrod Brown a one-point lead over state Treasurer Josh Mandel. The Quinnipiac poll, on the other hand, gave Brown a seven-point unadjusted (nine-point adjusted) lead. It’s hard to discern whether we’re seeing a tightening race, since Quinnipiac is right at the historical trend, while the other two point to erosion of Brown’s lead. I was expecting to move Ohio to “Likely Democrat” this time, but the latest data provide enough uncertainty that I’m leaving Ohio in the “Leans Democrat” column.
- Florida: Only one poll this time, from CBS/New York Times/Quinnipiac, who saw a nine-point unadjusted lead for incumbent Senator Bill Nelson over his challenger, Connie Mack, IV (R-Fort Myers). It adjusts to an 11-point lead. As I said last time, Florida’s polling has been relatively noisy, and Rasmussen’s particularly so, so his latest poll from two weeks ago didn’t give me much confidence. With the addition of the Quinnipiac poll, the picture is even clearer than it was last time, when I moved Florida into “Leans Democrat”. No change here.
- Virginia: Two polls, and two ties: Rasmussen and Public Policy Polling both saw no lead, before adjusting for house bias. After adjusting, they indicate a one-point lead for former Governor Tim Kaine over former Governor George Allen in Rasmussen’s case, and a three point lead for Allen in PPP’s case. At this point, I think Virginia will be a “Tossup” right up to election night.
- Connecticut: Three new polls, from Quinnipiac, Public Policy Polling, and Rasmussen. Quinnipiac and Rasmussen saw a three-point unadjusted (one point and two points adjusted, respectively) lead for former WWE executive Linda McMahon over Representative Chris Murphy (D-Cheshire). PPP had the opposite view, with four unadjusted (one adjusted) points in Murphy’s favor. This latest round makes it all but impossible to put Connecticut in any column other than “Tossup”, a move from “Leans Democrat”.
- Massachusetts: One new poll from Public Policy Polling gives incumbent Senator Scott Brown a five-point unadjusted (eight adjusted) lead over Elizabeth Warren. As a single poll, it’s impossible to tell whether we’re seeing an outlier or signal; it’s pretty far from the trend that has been in place since April. In the only other Intrade individual Senate market with enough trading activity to discuss, Brown is trading at a 59 percent win rate, while Warren is trading at 45 percent. It appears from the historical charts that the market is reacting to the PPP poll. Absent confirmation that PPP wasn’t an outlier, though, Massachusetts stays a “Tossup”.
We’ve had a few moves: Nevada, Missouri, and Connecticut.
And so, our former four tossups become six: Indiana, Massachusetts, Montana, and Virginia are joined by Connecticut and Missouri. It seems credible for Republicans to win at least three of the six, but winning three would not hand them the majority unless Representative Paul Ryan (R-Janesville, WI) becomes Vice President. Moreover, North Dakota, which has been lightly polled for some time, may not still be “Leans Republican”; the absence of data leaves us effectively blinded there. The Intrade markets moved a decent amount: they now give Republicans about a 54 percent chance of taking the Senate (up four points from two weeks ago), with a 26 percent chance of Democrats holding at least 51 seats (down five points from last time), and now a 25 percent chance of exactly 50 seats (up five from last time). The sum is 105 percent, so there’s clearly some market error. Most individual Senate race markets are still too lightly traded to produce any meaningful data.
How credible do you think those market numbers are? Do you agree or disagree with my state analyses above?
- Senate Watch: August 18 (logarchism.com)
- Todd Akin Within One Point Of Claire McCaskill In New Poll (outsidethebeltway.com)
- Missouri Sen. Claire McCaskill polls 10 points ahead of Rep. Todd Akin in latest Rasmussen poll — @TheAtlanticWire (theatlanticwire.com)