U.S. Senate 2010: The Final Predix

You knew it was coming.

Over the past year, I’ve posted more voter models and projections than I can remember, and now, with less than one week to election night, I think I’m ready for some final analysis. Not once in my monthly projections have I projected a GOP take-over of the U.S. Senate, and that remains the case today. Some races have trended right (Wisconsin, Florida, Ohio), but the races which most matter to a take-over (Connecticut, Washington, Delaware) don’t appear poised to trade hands. In fact, as you’ll see in my forthcoming calculations, I think there are a number of under-the-radar races where the Democrat is poised to overperform. In some cases, big time. (Be prepared to catch your jaw when Iowa comes up.) Oh, and speaking of these calculations…I’ve opted to only include races where I foresee a winning margin of 20% or less. (Sorry, Christine.) Also, I’ve included third-party candidates when I foresee that vote garnering 3% or more.

So, with succinct exposition on the table, I say it’s time for some numbers…

D -> R FLIP:

North Dakota

Arkansas

Indiana

Colorado

Wisconsin

Pennsylvania

West Virginia

Post-midterm U.S. Senate composition: 52D/48R

Alaska

R – 44%

I – 37%

D – 19%

McAdams – 4/24/64 = 23%

Miller – 54/29/6 = 36%

Murkowski – 42/47/30 = 41%

Arkansas

D – 40%

R – 38%

I – 22%

Boozman – 18/98/65 = 58%

Lincoln – 82/2/35 = 42%

California

D – 43%

R – 34%

I – 23%

Boxer – 86/3/42 = 49%

Fiorina – 12/95/47 = 47%

Other – 2/2/11 = 4%

Colorado

R – 36%

D – 32%

I – 32%

Bennet – 8/89/48 = 46%

Buck – 92/11/52 = 54%

Connecticut

D – 44%

R – 32%

I – 24%

Blumenthal – 92/11/55 = 57%

McMahon – 8/89/45 = 43%

Florida

D – 38%

R – 38%

I – 24%

Crist – 33/14/49 = 30%

Meek – 59/2/14 = 26%

Rubio – 8/84/37 = 44%

Georgia

R – 41%

D – 40%

I – 19%

Isakson – 98/18/64 = 59%

Thurmond – 2/82/36 = 41%

Illinois

D – 47%

R – 34%

I – 19%

Giannoulias – 86/2/42 = 49%

Kirk – 12/96/49 = 48%

Other – 2/2/9 = 3%

Indiana

R – 46%

D – 38%

I – 16%

Coats – 92/14/55 = 56%

Ellsworth – 8/86/45 = 44%

Iowa

R – 38%

D – 37%

I – 25%

Conlin – 4/86/42 = 45%

Grassley – 96/14/58 = 55%

Kentucky

D – 47%

R – 43%

I – 10%

Conway – 86/8/41 = 47%

Paul – 14/92/59 = 53%

Louisiana

R – 45%

D – 43%

I – 12%

Melancon – 8/87/46 = 47%

Vitter – 92/13/54 = 53%

Missouri

D – 40%

R – 38%

I – 22%

Blunt – 8/96/57 = 52%

Carnahan – 92/4/43 = 48%

Nevada

D – 43%

R – 34%

I – 23%

Angle – 8/86/44 = 42%

Reid – 87/7/42 = 49%

Other – 2/4/5 = 3%

None – 3/3/9 = 6%

New Hampshire

I – 37%

R – 32%

D – 31%

Ayotte – 57/95/13 = 55%

Hodes – 43/5/87 = 45%

New York (B)

D – 50%

R – 30%

I – 20%

DioGuardi – 6/89/48 = 40%

Gillibrand – 94/11/52 = 60%

North Carolina

D – 44%

R – 37%

I – 19%

Burr – 14/94/56 = 52%

Marshall – 86/6/44 = 48%

Ohio

D – 42%

R – 36%

I – 22%

Fisher – 84/3/36 = 44%

Portman – 16/97/64 = 56%

Oregon

D – 38%

R – 33%

I – 29%

Huffman – 6/86/45 = 43%

Wyden – 94/14/55 = 57%

Pennsylvania

D – 46%

R – 42%

I – 12%

Sestak – 88/8/47 = 49%

Toomey – 12/92/53 = 51%

Washington

D – 36%

R – 32%

I – 32%

Murray – 92/6/49 = 51%

Rossi – 8/94/51 = 49%

West Virginia

D – 48%

R – 39%

I – 13%

Manchin – 84/8/48 = 49%

Raese – 16/92/52 = 51%

Wisconsin

D – 39%

R – 38%

I – 23%

Feingold – 89/4/44 = 47%

Johnson – 11/96/56 = 53%

***Voter models based on analysis of ’04-’08 CNN exit polling data + actual voter registration numbers + recent public polling data***

21 thoughts on “U.S. Senate 2010: The Final Predix”

  1. iowa: i don’t think conlin will do as well as culver does

    colorado: bennet’s gaining, and Colorado has a lot of environmentalists who will be more energized by recent statements by Buck

    louisiana: too red of a year for that close of a margin

    nevada: reid’s in more trouble than this

    north carolina: see louisiana

    west virginia: manchin’s strong, raese is weak

  2. Hope you’re right about IL- Giannoulias is struggling, but I sure hope he pulls it out. I think btwn the Green and the Libertarian, they’ll probably take more than 3%. I do think/ hope/ pray that Alexi pulls it out, but I think it’ll be closer than 1%, and he’ll win with no more than 47%, Kirk- 46.5%.

    Also, Boxer’s race will be close, but I don’t think 2% (I did think that a few days ago), I think a 4-5% cushion is more likely.

    Would be curious to hear your thoughts on the why behind the #’s you project.

    Well done, otherwise.

  3. Thanks for sharing it.

    Re: NV, I’m wondering how you came up with your partisan breakdown. Based on the first week of early voting, it seems that there’s close to parity between Ds and Rs. Do you think that Ds will turn out in greater numbers on Election Day?

    http://www.nvsos.gov/index.asp

  4. As others have noted, California is way too close (5% is probably more appropriate), Colorado not close enough (I still think Buck wins, but it’ll be within 5. Your turnout numbers assume a slightly more professional R candidate.)

    Connecticut is also not close enough. I’d put Blumenthal maybe 5 or 6 points up over McMahon, tops. No way he wins by less than 4-5, but it’ll still be closer than that.

    Kentucky will be closer, probably a mirror image of the 2004 KY-Sen result. Seriously, if a guy as crazy as Bunning needs Bush-level turnout to carry him over the finish line against Dan freaking Mongiardo, a guy as crazy as Paul vs. an actual Democrat is going to be similarly close, despite the tea-flavored nature of much of the state.

    Nevada will be much closer and will probably go to a recount, but Reid will win.

    I’d give a few more points to Gillibrand in New York – more like 63-37.

    Your results for Washington are WAY too close. It’ll be more like 54-44 Murray. Turnout in King County is HUGE this year, and the “cell phone effect” really complicates polling in western WA. Also, Dino Rossi is a joke of a candidate who will never win a statewide race because King County utterly loathes him and wants his political career to be over for good. (Trust me, I used to live in Seattle. Even in 1994, Rossi probably wouldn’t win statewide.)

    The two races I disagree completely with you on outcome are Pennsylvania and West Virginia.

    Sestak’s “late surge” strategy makes it really difficult to poll PA, but I think Toomey got caught napping and is collapsing, if the polls tightening in the last few weeks is any indication. We’ll see on election day if I’m right, but I see a final result of more like 52-48 Sestak. Don’t underestimate Sestak’s appeal in the Philly and Pittsburgh burbs, which is way more than that of Generic D.

    In WV, Raese has had too many scandals and is also collapsing. The state obviously wants to elect a Republican to the seat, but they’re not so keen on a carpetbagging rich guy from Florida, and Manchin’s personal popularity also needs to be taken into account. It’ll be more like 53-47 Manchin.

    There are a lot of other races I agree with you completely: Alaska, Arkansas, Florida, Georgia, Indiana, New Hampshire, Ohio, Oregon, and Wisconsin (+/- 2, but Johnson still wins).

    I also agree on Louisiana. People are underestimating Melancon’s strength downstate and in New Orleans (especially after Deepwater Horizon), and overestimating Vitter’s ability to not be an arrogant ass. Melancon has been campaigning relentlessly, Vitter not so much. It’s only the Shreveport/upstate vote that will return Vitter to office at all, but it’ll be within a couple of points, as you think. At least that’s my take.  

    I think Illinois and maybe Missouri are “wait and see”.  

  5. AK-Sen:  I increasingly think Murkowski not only might win the write-in, but Miller might be nosediving all the way down to 3rd, into the 20s.  I’ll go out on a limb and say he does finish 3rd, and McAdams beats him.  I really think the latest revelations this week will cost Miller almost all votes outside the hard-core GOP base.  And if Murkowski doesn’t win it, McAdams will, although that scenario would be a much closer 3-way with Miller holding more support to hold down Murkowski.

    CO-Sen:  I disagree, I think Bennet has momentum and eeks it out.  He’s peaking at the right time, and Buck still can’t stop putting his foot in his mouth.  I think indy women break toward Bennet.

    WV-Sen:  Disagree based on all recent tea leaves.  Only Rasmussen (including its Fox Pulse alter ego) has Raese leading, others say Manchin leads, and Charlie Cook confirmed in a column a few days ago that Manchin has a small but consistent lead.

    NV-Sen:  I agree Reid wins but by much less, because I think your turnout model is WAAAAAY off given what we know about early voting and the fact this is a midterm when Dem vote share is worse than in a Presidential.  I also think you make two offsetting mistakes in that you understate both Reid’s share of Republicans and Angle’s margin with indies; both will be considerably higher than you predict, based on my close reading of all polling over these many months.

    IN-Sen, IA-Sen, KY-Sen, NC-Sen, OH-Sen:  I think in these 3 you actually have the Democrats OVERperforming, I expect them to do worse than these numbers.  Grassley has run his best campaign ever, I think better than when he beat John Culver to win the seat in 1980, and that more than any mistakes by Conlin will push him over 60%.  Conway is tanking, I think.  And Ellsworth has been a perpetual “meh” in a conservative state in a Republican year.  And Marshall has tanked, she never had the money to compete against Burr’s huge warchest.  NC will be a blowout, perhaps not quite 20 points but definitely in double digits.  For Ohio, same as NC but much worse because Fisher has run a worse campaign than Marshall, in addition to getting swamped financially.  I actually think Fisher fails to reach 40%.

    IL-Sen:  this one I don’t disagree, but neither do I agree.  To me, this is the hardest-to-predict race of all simply because neither candidate can consistently poll at or above 45%, and that makes it impossible to guess how it ends up.  The 3rd-party support will bleed, but how much?  Indies will break which way?  Do we know for sure that more, or even most, undecideds are Dems and Dem-leaners?  Even 6 days out, too many unanswered and perhaps unanswereable questions.  I tend to think in my gut, sadly, that Kirk eeks this out, but objectively your prediction is equally likely as my gut.

    All in all, I say we end up with 53-54 seats, rather than the 52 you’re predicting.

  6. but disagree on 4 of the individual races.  In 3 of those races, Independent voters are the key.

    NV: Reid’s GOTV will help maximize his vote from Dems, but he will lose Independent voters by much more than 2 points.  Independents in Nevada tend to be more on the conservative side.  The 43D-34R-23I projection is also too optimistic.  Unlike most here, I think Reid will not quite make it.

    IL: Despite his resume problems, Kirk has a voting profile that has worked well for Republicans in the past in Illinois.  He will do better than +7 among Independents.  Also, Giannoulias is not keeping it as close as he needs to downstate in order to make up the difference in Cook County.

    CO: Social issues are playing a greater role here than in any other state and Buck is on the wrong side of those issues.  Independents in Colorado, particularly women, will side with Bennet by 10+ points.  My guess is Bennet will squeak out a narrow victory in the closest Senate race on election night.

    WV: Manchin is very popular among all voter groups, from liberal to conservative.  This will enable him to withstand any negative sentiment towards Democrats nationally.  Manchin should win this race fairly easily.

    Among the other races, Alaska is the most intriguing.  It would be great to see a poll there to quantify how far Miller has imploded.

  7. Latest Quinnipiac has Dems moving to Crist – he was down 50-38, now up 51-36. Now, it is probably an outlier and Rubio still takes it comfortably but just for fun what do you think Crist needs to sneak by? He has 51% of Dems, 19% of Repubs and 43% of indies according to Q while you currently give him 33/14/49.    

Comments are closed.