Fewer key states were polled for Sen­ate can­di­dates over the past two weeks than we had last time. But we did get some move­ment in a few races.

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:

  • Ari­zona: One new poll, from Pub­lic Pol­icy Polling, shows Rep­re­sen­ta­tive Jeff Flake (R-​​Mesa) ahead of Demo­c­ra­tic for­mer Sur­geon Gen­eral Richard Car­mona by one unad­justed point (which adjusts to four). Polling has been light and infre­quent, but Car­mona has yet to lead in a sin­gle poll. Last time I said I was torn between keep­ing Ari­zona at “Leans Repub­li­can” and mov­ing it to “Likely Repub­li­can”. I’m not torn any­more; Ari­zona is now “Likely Repub­li­can”. Intraders seem to agree; Flake is trad­ing at 87 per­cent, and Car­mona at 22 percent.
  • Mon­tana: One new poll from Pub­lic Pol­icy Polling gives Demo­c­ra­tic incum­bent Sen­a­tor Jon Tester a two-​​point unad­justed lead over Rep­re­sen­ta­tive Denny Rehberg (R). This adjusts to a tie. Cou­pled with Rasmussen’s poll from last time, I still don’t have suf­fi­cient evi­dence to con­sider Mon­tana any­thing but a “Tossup”. Intraders are more inclined to believe Rehberg will win here; he is given a 65 per­cent like­li­hood, with Tester at 44 per­cent. It’s a lightly-​​traded mar­ket, though, so I don’t give much cre­dence to it.
  • New Mex­ico: Another pair of new polls. One was from the Albu­querque Jour­nal, who found Rep­re­sen­ta­tive Mar­tin Hein­rich (D-​​Albuquerque) ahead of his pre­de­ces­sor, for­mer Rep­re­sen­ta­tive Heather Wil­son (R-​​Albuquerque) by seven points. The other, from Pub­lic Pol­icy Polling, indi­cates a nine-​​point unad­justed Hein­rich lead (which adjusts to seven). The num­bers are con­sis­tent with every poll since April, and lines up per­fectly with Rasmussen’s poll from last time. Wil­son has yet to even tie a sin­gle poll, let alone show a lead. More­over, when look­ing at the same polling firms over time, the trend is a con­sis­tently increas­ing gap between the two can­di­dates. For this rea­son, I’m mov­ing New Mex­ico from “Leans Demo­c­rat” to “Likely Demo­c­rat” a cou­ple of weeks ear­lier than I had orig­i­nally expected to.
  • Mis­souri: Since last time, we’ve had one poll in Mis­souri, from Ras­mussen. He saw Demo­c­ra­tic Incum­bent Claire McCaskill ahead of Rep­re­sen­ta­tive Todd Akin (R-​​Wildwood) by six unad­justed points (which adjusts to seven). Ordi­nar­ily, this would be cause for mov­ing Mis­souri from “Tossup” to “Leans Demo­c­rat”, since no polls after Akin’s rape com­ment have shown the chal­lenger with a lead. How­ever, Ras­mussen saw McCaskill ahead by four more points two weeks ago, which could sug­gest a nar­row­ing lead. Until the polls show me more, Mis­souri stays a “Tossup”. Intraders dis­agree, giv­ing McCaskill a 63 per­cent chance to Akin’s 40 percent.
  • Michi­gan: One poll in the past two weeks, from EPIC-​​MRA, gives Demo­c­ra­tic incum­bent Sen­a­tor Deb­bie Stabenow an 11-​​point lead over for­mer Rep­re­sen­ta­tive Pete Hoek­stra. Last time, I noted that both Mitchell Research and Baydoun/​Foster looked like out­liers in show­ing Hoek­stra ahead, given the his­tor­i­cal trend, though it’s uncom­mon to have two out­liers in the same time period. It now appears my cau­tion was jus­ti­fied, so I’m leav­ing Michi­gan as “Likely Demo­c­ra­tic”. Intraders agree, giv­ing Stabenow an 85 per­cent chance to Hoekstra’s 23 percent.
  • Ohio: Three new polls this time, from NBC/​Wall Street Jour­nal/​Marist, Pub­lic Pol­icy Polling, and Gravis Mar­ket­ing. Marist gives incum­bent Sen­a­tor Sher­rod Brown a seven-​​point unad­justed lead (which adjusts to five points) over Trea­surer Josh Man­del. PPP came away with Brown up by eight unad­justed points (five adjusted). Gravis sees Brown up by six unad­justed (eight adjusted). Last time, I thought we might be see­ing evi­dence of a tight­en­ing race. This time, it looks like Brown has a com­fort­able lead. Given how noisy polls have been in Ohio this year, I’m leav­ing Ohio in the “Leans Demo­c­rat” col­umn until we have a longer period of time with­out the close num­bers of two weeks ago. If we see the same thing in the next install­ment, I’ll move Ohio to “Likely Demo­c­rat”. Intraders are as bull­ish for Brown as they are for McCaskill, giv­ing him a 65 per­cent chance to Mandel’s 25 per­cent. It’s a rather lightly-​​traded mar­ket, though.
  • Florida: Four polls this time, from Pub­lic Pol­icy Polling, Gravis Mar­ket­ing, Sur­veyUSA, and NBC/​Wall Street Jour­nal/​Marist. If you look over the two-​​week period, incum­bent Sen­a­tor Bill Nel­son has gained a lit­tle, but his chal­lenger, Con­nie Mack, IV (R-​​Fort Myers), has rapidly dropped about four points over that time. Nelson’s lead is approach­ing dou­ble dig­its. I moved Florida into “Leans Demo­c­rat” a month ago, but I’m near­ing the point where I’ll push the Sun­shine State into “Likely Demo­c­rat”. At this point, I’m wait­ing to see if this is a tem­po­rary bump or a new plateau. This mar­ket is too lightly traded on Intrade to be useful.
  • West Vir­ginia: This is a rarely-​​polled state, and only Pub­lic Pol­icy Polling has both­ered to look any­way. Regard­less, with mar­gins well above 20 points each time, there’s no doubt that Demo­c­ra­tic incum­bent Sen­a­tor Joe Manchin will hand­ily defeat ultra­con­ser­v­a­tive Repub­li­can John Raese. West Vir­ginia is “Likely Democrat”.
  • Vir­ginia: Two new polls this time. One from Gravis Mar­ket­ing gives for­mer Gov­er­nor George Allen a five-​​point unad­justed lead (which adjusts to two) over for­mer Gov­er­nor Tim Kaine. NBC/​Wall Street Jour­nal/​Marist sees an unad­justed tie, which also adjusts to two points in Allen’s favor. Two points is within the mar­gin of error of the his­tor­i­cal trend of tossup. Until I see data to the con­trary, I will con­tinue to believe Vir­ginia will be a “Tossup” right up to elec­tion night. Intraders are more con­fi­dent in a Kaine vic­tory, giv­ing him a 59 per­cent chance of win­ning, to Allen’s 30 percent.

Cal­i­for­nia, Min­nesota, New Jer­sey, and New York were also polled, but their out­comes have never really been in question.

The num­ber of seats left up for grabs is dimin­ish­ing, with the only remain­ing lean­ers of Ohio and Florida on the left (both rapidly mov­ing to “likely” sta­tus), and pos­si­ble flip­pers North Dakota and Wis­con­sin on the right. Our six tossups remain so: Con­necti­cut, Indi­ana, Mass­a­chu­setts, Mis­souri, Mon­tana, and Vir­ginia. Repub­li­cans’ best chance is Indi­ana, while Democ­rats’ best is prob­a­bly Con­necti­cut. The other four are cur­rently coin-​​flips. That makes it 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, a decreas­ingly likely event. 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 45 per­cent chance of tak­ing the Sen­ate (down nine points from two weeks ago), with a 33 per­cent chance of Democ­rats hold­ing at least 51 seats (up seven points from last time), and now a 21 per­cent chance of exactly 50 seats (down four from last time). The sum is 99 per­cent, which sug­gests that these mar­kets are being traded heav­ily enough to carry some weight.

How cred­i­ble do you think those mar­ket num­bers are? Do you agree or dis­agree with my state analy­ses above?

Related arti­cles