Princeton Election Consortium

A first draft of electoral history. Since 2004

Senate update, October 22: 53 D/I, 47 R

October 22nd, 2012, 10:30am by Sam Wang


Since my last update, the terrain has shifted by about one seat away from Democrats. This may be a coattail effect from the Presidential race. The median outcome is 53 D/I, 47 R (1-sigma confidence interval, 53-54 D/I seats). Retained control by Democrats/Independents is still extremely probable: 98%.

The details:

The following seats are relatively uncertain, either because of a small margin or because polls are too few to determine conditions of the race accurately. Margins are median +/- estimated SEM. For the calculation I used all polls since 10/4 or the last 3 polls from Pollster.com, whichever is greater. Statistical leaders are indicated in bold.

  • AZ: Carmona (D) vs. Flake (R). R+6.0 +/-4.1% (n=5).
  • IN: Donnelly (D) vs. Mourdock (R). R+4.0 +/- 2.1% (n=3).
  • MT: Tester (D) vs. Rehberg (R). D+2.0 +/- 0.9% (n=3).
  • ND: Heitkamp (D) vs. Berg (R).  Tie +/- 6.0% (n=3).
  • NV: Berkley (D) vs. Heller (R). R+3.0 +/- 3.4% (n=7).
  • VA: Kaine (D) vs. Allen (R). D+1.0 +/- 1.9% (n=11).

The following seats have a clear leader:

  • CT: Murphy (D) vs. McMahon (R). D+4.0 +/-1.1% (n=5).
  • MA: Warren (D) vs. Brown (R). D+4.5 +/- 3.1% (n=6).
  • WI: Baldwin (D) vs. Thompson (R). D+3.5 +/- 0.7% (n=10).
  • MO: McCaskill (D) vs. Akin (R). D+6.0 +/- 4.1% (n=5).

At an individual level, knife-edge Senate races are where donations and volunteer efforts would be most effective. Key races are listed at ActBlue (D)Crossroads GPS (R) is likewise strategic in their approach. Show your support!

Finally, let’s try an experiment to control comment overgrowth. If you have something to say, please focus on Senate issues. No Presidential freakouts please!

Tags: 2012 Election · Senate

28 Comments so far ↓

  • Rick in Miami

    It’s been interesting to see that the % chance of D/I control has barely dipped, in contrast to the presidential meta-margin. It fell to a post-debate 1 minimum of 83% on 14 October, then rapidly shot up with a few polls released in the next day (back into the mid-90s) and then crept up to 98%.

    Like Sam says, there are 3-6 seats *really* hanging in the balance – tipping most of them one way would make a big difference down the road. The difference between a 52 seat majority and a 54 seat majority could mean a lot in 2014.

  • HSG

    I’m wondering if this is indirectly good news for Obama in states like WI, OH, VA, etc., where the D Senate candidate is leading and thus the weaker R candidate is a drag for Romney. I seem to remember reading this somewhere — 538?

  • divF

    First-time commenter … I get more out of this site than any other, thanks.

    Regarding your Senate commentary. The statistics are what they are, but I would view Kaine’s chances as better than Tester. His lead, while small, has been remarkably stable over the last three weeks, if not longer.

    • Jay Bryant

      Second time posting for me, and I just want to say that this is the best of the 6 sites I check every day for my political information. That includes 538. Nate’s methodology includes too much noise. Sam’s methodology makes more sense to me.

      Good work, Dr. Wang!

  • Ms. Jay Sheckley

    Because I need to make every penny count, I never give politically, I just volunteer in my own way. I used to make rapid calls for candidates, now I creatively agitate online with gentle humor and info-mongering, and on election day I do radio call-in of politically-aligned swing-state based radio to get out the vote.

    But you’ve changed me. Don’t tell my husband I’m ready to give a third time from your link. [As if he won't read this the minute he wakes up.] I just need to know:

    In the link, are the recipients now listed in the order you think will help most? I’m willing to pay my PEC dues again, but I’m a bit uncertain what to do. Does the order of listings mean that giving to the party is currently less useful than to the top candidate?

    • Ms. Jay Sheckley

      I also need to know because I want to share your donation link, and this’d help me write a header. I thought the party was likelier to give to a tied race, so maybe the more challenged races are needier?
      BTW, are you saying the Senate is more influenceable now than the house?

  • Andy

    I feel the difference is that people tend not to think nationally when they’re voting quote-unquote locally. I don’t know the stats on how many people choose the “straight ticket” option…you hear often about ticket splitters who do so strategically. Personally, I wish more people voted out their incumbent Republican Senators and House members solely because of their obstructionism…I can imagine a lot of W.Va. Dems vote for because of a local issue like coal, without thinking how their vote affects the balance of power in D.C.

    • ted

      On my ticket, I have a choice of senators, a republican jack*** who preaches the self serving libertarian lines. (i.e. remove the irs!) or a libertarian whom I’ve never heard of.

      Personally, I’m going with the devil I don’t know over the devil I know said “I can barely feed my family on 250,000 a year!”

      Given the “libertarian” bent the unknown guy may actually suck up a significant number of votes from the repub. Sadly, no democrat on the ticket…not even a pretend one..

  • Olav Grinde

    Dr Wang, to what degree do you believe the recent swings can be attributed to financial disparity between the campaigns, and/or SuperPAC involvement?

    I seem to recall that Donnelly, Carmona and Berkley were doing somewhat better not long ago.

    In Massachusetts, I understand that incumbent Scott Brown and challenger Elizabeth Warren signed and agreement to not use outside money — and that Warren has raised much more money than Brown.

    One more question: Mayor Michael Bloomberg of New York recently used the S-word (socialist) to attack Warren and endorse Scott Brown. Any thoughts on what impact this may have?

    • arguethefacts

      Olav, Karl Rove and the Koch Brothers found a loophole in the agreement Brown and Warren made. Warren is still sticking to the agreement. In the last couple of weeks Rove and the Koch Brothers have pumped in millions of extra dollars into Brown’s campaign. Although it seems to have had the opposite affect since most voters are looking on Rove and the Koch Brothers as outside interference. Once they began exploiting that loophole Warren’s numbers started going up. I think ultimately their interference has so angered the voters that it will give Warren the win.

  • Joel

    Sam, any insight as to why some races (Baldwin-Thompson, Murphy-McMahon, and Tester-Rehberg) have much smaller deviations? Larger poll samples? It looks like the N of polls is basically the same, with the exception of Baldwin-Thompson (and the Kaine-Allen race has plenty of polls but still a higher error).

    • Sam Wang

      My first guess would be less ethnic diversity in those states. All polls have to be re-weighted according to the likely composition of actual November voters. if a sample is really heterogeneous, there are more ways to do the weighting.

      Virginia poses two challenges: considerable diversity (the areas near Washington, plus rural areas), as well as the presence of multiple partisan polls. My gut says Kaine is ahead and race is stable. I’m reporting the medians as opposed to my gut.

  • Joe Sarachek

    The interview on NPR was fascinating. Can you point by point illustrate how critical Ohio is to the outcome again.

  • Justin S

    Very maddening! This whole election has been driving me nuts. I’m so nervous

  • CH

    Dr. Wang, do you think you could provide a House update in the next few days? Your take a few weeks back was rather bullish, more so than ‘conventional wisdom’ has been. I don’t put very much stock in the latter; I’m hungry for some serious analysis of the data out there.

  • Justin S

    Its a little frustrating that Rep pollsters seem to be flooding the news with polls while more reputable ones are putting less info out

    • Obama 2012

      Yup. I just posted about this happening in Virginia but I think it’s happening all over the country.

      If this model ends up being biased towards GOP I think we will know why (the flood of obviously biased GOP pollsters)… there may be some sort of minimum standard that a poll needs to reach before being included?

    • Howie Weiner

      As I said some polling companies are being used in the same was as political ads. They are essentially advertisements for their candidate. By flooding the market with more polls the Republicans hope to tip the public perception that their candidate is doing better which has the twin benefit of mobilizing their base and demoralizing the opposition. It is all part of psychological warfare.

    • Olav Grinde

      Howie, the interesting thing is that if the Republican and Republican-leaning pollsters succeed in driving the narrative, influencing public perception, motivating a high turnout for their base and depressing turnout for the opposition, they will have been right!

      …which would mean that in the next round there can be a debate about why the other polls were so skewed against the Republicans.

    • William

      I don’t think you’ll see that effect, Olav. Obama’s GOTV is too good for that level of demoralization to come into play.

  • Obama 2012

    I’me very surprised to see you list the Kaine/Allen race as a toss-up; I was under the impression Kaine was running away with it.

    I think this is a situation where right wing polls are flooding the state and throwing off the median poll …

    Check this out: Of the last 10 VA Senate polls 6 of them are GOP pollsters (Rasmussen, Wenzel Strategies, Pulse Opinion X 2, We Ask America, McLaughlin)

    The other four polls (which also includes the questionable YouGov tie…) gives Kaine an adverage lead of 3% … it includes 7 point lead for Kaine from CBS/Quinnipiac…

    • Sam Wang

      It is usually the case that partisan polls are, in part, a messaging strategy. I believe this is why RCP does not list commissioned polls. We follow Pollster.com’s approach, which is to take everything that appears to be legitimate. I agree that there is an issue here. But one thing I won’t do is pick some partisan polls and not others.

      I do agree with you that Kaine is probably ahead now, which is why I am not listing him under ActBlue.

  • LarryinLA

    I’m interested in the effect of quantization noise on both the Senate and Presidential models here. Because of the use of median statistics and the integer percentage reporting of all polls, there’s added noise introduced into the model. It seems to be pretty significant, as moves in the positions of individual states and senate races happen in 1 point steps (or occassionally 0.5, but this depends on whether n is even or odd, not any intrinsic element of the race). I think without this effect, moves in the MM would be much less frequent.

  • Mark F.

    So, Romney’s most likely path to victory is FL, VA, NC, OH plus either CO or IA. NH also feels like it will go Romney.

    • arguethefacts

      I’m in New Hampshire and the Gubernatorial race is dragging Romney down. The Tea Party won massively in 2010 and took over the state government and proceeded to try to enact every ALEC bill they could. Fortunately the Governor vetoed most, but some of his vetoes were overridden. And fortunately the legislature only meets two weeks a year. He is a very popular governor but not running for reelection.

      The Republican nominee for Governor is somewhere to the right of Attila the Hun. He has also run for Governor before and been soundly defeated. He’s a typical rape blame it on the woman governor who believes there should be no exceptions for rape, no contraception, and felony convictions for doctors who perform abortions. And he has hinted he would also put the mother in jail too, for having an abortion. He is also for pulling the state out of medicare and reducing/eliminating early childhood education, as well as privatizing everything in the state. Hopefully, his extreme views seem to be having an affect not only on his race, but also on Romney’s poll numbers. Romney has endorsed this wingnut.

  • Zach C

    My question about VA has already been answered but I do have to ask why is PEC more certain about Chris Murphy than 538? Is it just because the 538 model is giving more weight to older polls where McMahon was polling better? Those two states seem to be the only races where 538 and PEC are giving different answers. Thanks.

  • jharp

    Hi Sam,

    First visit and awesome place. Daily read for me from now on.

    Found out about you from a poster on Taegan Goddard’s Political Wire.

Leave a Comment