Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

Senate: following Obama back up

October 31st, 2012, 10:25pm by Sam Wang

(original version published on temporary server)

Rick in Miami has helpfully calculated some time series. As has been the case for several months, Democrats/Independents will control 53-55 seats (middle 50% of outcomes) – a change of zero to +2 seats for the Democrats.

Rick has added Pennsylvania and Nebraska, whose races have gotten close. Come to think of it, maybe that’s why Romney is hitting Pennsylvania – boosting downticket races. I don’t think it’s because he has much chance of winning there.

There’s something interesting What’s interesting about this graph is the decrease in Democratic/Independent expected seats in the first two weeks of October. This parallels President Obama’s precipitous drop after Debate #1. Unlike the Obama-Romney EV margin, Senate prospects have fully recovered to pre-debate levels – and then some. What individual race(s) could have done this? The implosion of the Mourdock (R-IN) campaign comes to mind.

(Here is the original comment thread at DailyKos from the temporary site.)

Tags: 2012 Election · Senate

55 Comments so far ↓

  • Brian

    Are we pretty sure that King from Maine will caucus with the Dems?

    • Charles Zuckerman

      I think so. King was a moderate governor who at the outset stated he would watch how he was treated, respectfully or otherwise, if and when he needed to caucues. The dems got out of the way, barely running a campaign and investing no national party money. The republicans went on the attack. If both his political attitude and his initial statement have any weight, he will caucus with the democrats.

    • Jennifer de Poyen

      he hasn’t actually committed to caucusing with democrats, presumably to leave himself the option of joining the republicans, should they somehow capture the senate, or of being in the catbird seat of being a senator both sides need to woo for a majority. but everything about him suggests he will caucus with dems. this would be why dems have deserted their own candidate to help independents with king.

    • Philip Diehl

      The odds that King caucuses with Dems = the odds the Dems control the Senate. Why? Because the Dems will organize the Senate and determine which committees King sits on. Plus, he’ll be in the majority, rather than the minority, always an advantage. King’s committee assignments have a big effect on his chances of reelection. Finally, he’s more aligned with Dems on most policy matters.

    • James Kocot

      King has said he would caucus with the Democrats in a debate held in Maine.

  • George

    At one point when O was riding high you were running probabilities on a house switch. Seems highly unlikely now – but thoughts on gains or losses?

  • William

    Any thoughts on the House?

  • Tapen Sinha

    At this point, I would say forget the House. However, it will be in play in 2014 when the economy turns around – big time.


    • Ray Umashankar

      Excellent point. 2014 will be time to thow all the tea party candidates overboard!

    • Philip Diehl

      I wish we could be optimistic about the House in 2014, but the historical record shows the incumbent’s party takes a hit in the by-election of his second term. Even if the Dems were able to escape this fate in 2014, it’s highly unlikely they could capture enough seats to take the House, especially in the era of Citizens United and banana-republican voter suppression.

  • Sean Ernst

    Uh oh. Dr Wang you (and Nate Silver) have been called out again, sort of… This arguement is at least a bit more rational. Any thoughts?

    • JohnJacobs

      I wouldn’t take this guy seriously – he’s a partisan hack.

      “(4) A certain number of intangible factors favor Obama. Hurricane Sandy changed news coverage dramatically in the last week of the election: a disaster brings Americans together, lets a president seem presidential, … Moreover, every minute spent on dramatic Sandy stories is a minute the news isn’t spending on Obama administration failures in the economy, in energy policy, in deficit reduction, in Libya. It’s hard to think that the double-standard of “fact-checking” that more harshly (and often incorrectly) scrutinizes Romney claims than Obama claims doesn’t have some effect on undecided voters.”

      This paragraph makes it clear he is not an objective player. He cites a largely unquantifiable (mythical?) pro-Obama Sandy factor, and then launches an attack on Obama and a defense of Mitt’s mendacity.

      People find it hard to accept objectivity when it favors one party, but the objective truth is that Obama has been a good president and Romney has campaigned with a near total disregard for the truth.

  • StratMan

    I agree. GOP-friendly redistricting after 2010 put the house out of reach this time around.

    • Craig

      The average Cook PVI of a House seat was previously D+.5, with a median of R+2. It’s now D+.2, with a median of R+3. That’s about exactly where it was in 2004, and it’s mostly due to Democratic votes being naturally concentrated in urban districts. The GOP doesn’t have a single district of R+30 or above, while the Democrats have fifteen, including two in NYC which are D+41 (Rangel and Serrano).

  • Chris

    All the senate polls are systematically favored toward democrats, or didn’t we all get the memo?

  • Amitabh Lath

    I wrote in jest a while ago that Obama rope-a-doped debate#1 because he knew he would eventually pull up (and he has) and that it would get Republican money to flow away from downticket races towards the Presidential.

    Sam called it the “fake broken wing” theory.

    But look! Pre-conventions the senate was stuck at D 51, now it is (almost) D54. I know some if it due to idiotic comments about abortion. But still.

    • Ray Umashankar

      I felt the same way. At the time there was a lot of talk about Rove & Koch brothers diverting money to the senate and house races.

  • P.B.

    Slightly OT but thought I’d post on the freshest thread. Nate Silver and Josh Marshall are both pointing to this bit of devil’s advocacy. Would love to hear Sam’s thoughts. (PEC is mentioned several times.)

    • Sam Wang

      That is a very long essay.

      I am not sure what to think. In my reading, state polls do quite well – when aggregated using the median. If that’s true, then I have to figure out what on earth people like that writer are going on about.

  • Steve in Colorado

    Gave some money and campaigned for Joe Miklosi yesterday (Co-6). Hopefully Coffman’s Tea Party antics and birther statements will be remembered by more than just me.

  • Olav Grinde

    Kudos to Andrew and Sam for getting PEC back up and running! And thank you so much for the temporary fix…that allowed me to get my daily fix of Meta-Margin and Obama re-elect probabilities.

    It’s calming to see the Meta-Margin back at ±3% and the EV prediction hovering around 320..

    And it’s very soothing to see the strong likelihood of continued Democrat control of the Senate.

    Hope the Republican control of the House is whittled down to a majority of less than a handful of votes.

  • whirlaway

    I don’t know how this matters. The Republicans are not afraid to filibuster. So nothing will useful will pass the Senate, just like the last 4 years..

  • xian

    what as the offset calculated from incumbency and gerrymandering? wasn’t it about 2.5%? If Obama’s margin is more than that, then the House is in play, yes?

  • Rob

    Sorry if this is a stupid question, but the EV prediction is at 323 right now, with Romney at 215. The only combination I can see that ends up with this number is giving Colorado and NC to Romney. However the map seems to show Colorado is lean Obama. Am I missing something?

  • piktor

    Dan emocrats need 35 gross seats to control the House. 25 net gains without any seats lost. Historically, since WW II the party in the White house has never gained more than 15 seats in an election. All this from Charlie Cook interview at The Daily Beast.

    • xian

      Honestly, I don’t care bout “since [arbitrary moment], [entity described in some crafty way] has never [done some specific thing with an arbitrary cut off point]” models.

      I care about the math.

  • BrianTH

    I don’t think lack of historical precedents should be given much weight. I’d like to see more actual polling on the House.

  • TAW

    I’m in Illinois 10th.

    Republican candidate has more money than God. This should be a Dem seat.

    Incumbent Republican Bob Dold was a tea party 2010 victor. They have him tacking heavily to the center to reflect the district.

    The amount of telephone, direct mail, signs, and TV (although I don’t watch it) is unbelievable.

    • Olav Grinde

      The situation reminds me of a bumper sticker I saw a couple of decades ago.

      Support Democracy. Buy a politician.

  • DaveM

    The direct mail stuff is becoming increasingly offensive. The grainy black-and-white photos emblazoned with litanies of monstrous misbehavior are annoying, but it’s the profligate waste of paperboard that’s really insane. We’ve gotten enough of those miniature billboards to wallpaper our house. They go straight into the recycling bin. There are certainly no environmentally responsible state or local candidates competing for my vote.

  • Pat

    The restored site seems to be missing posts between 10/31 and 11/3. Will they be restored? Very glad you are back up.

  • InDecided Voter

    Have dropped about £2000 ($3200) on Obama at Betfair, will put another £1000 for the EV based on your prediction Mon eve/Tues morning. Can still get silly on Obama 310-329 and 330-349, it’s just free money.

  • Jeff Smith

    OK, just because I worry about most everything, anybody else worried about the recent RAND data trending toward Romney?

    • Sam Wang

      No. It’s just another survey, though one that tracks the same people repeatedly rather than resampling at random. It would be interesting to test whether the nature of the survey alters the result. As I have written recently, I think its most interesting feature is its ability to track the switching of individuals between the two candidates.

      I have a post on undecided voters (November 2, 4:30pm) that I am about to port over here that addresses the RAND survey a bit.

  • Billy

    Hi Sam, the NYT has a slick infographic of possible outcomes for each candidate:

  • Matt McIrvin

    It looks like it follows the higher-frequency wiggles in the presidential race remarkably closely, but there’s a general Democrat-ward trend superimposed relative to the presidential race. I wonder why that is? Maybe people are more disgusted by Generic Republican than by Mitt Romney at this point.

  • Matt McIrvin

    The latest Romney Advantage of Doom is Democrats not returning their mail-in ballots:

    If they try to vote in person but don’t bring the mail ballot with them, they’ll have to vote provisional.

  • Craig

    There is some…funny polling going on in PA-SEN. All of the R firms have at 0-2 Casey, while all of the D and nonpartisan firms have it at about 8-12 Casey.

    One of those R firms is Susquehanna, who said outright that they not only weight by partisan ID (doh) but they do so according to the 2010 midterms.

    Their demographics are also badly skewed: 88% white (82-81% in 2004-2008), 48-49% male (46-47% in 2004-2008), 10% young (21-18% 2004-2008).

    Wenzel Strategies had similarly skewed demographics (Wenzel thinks that 40% of the voters will be 60 or over, which doesn’t even come close to matching the 2010 results, let alone presidential years.) McLaughlin, Rasmussen/Pulse, and some marketing firm called GS Strategy Group did not release internals.

    Pro tip: Don’t hire marketers to do political work. You shouldn’t hire me to do your marketing either.

    • Sam Wang

      Weighting by turnout from an off-year election like 2010 is a particularly bad move. No wonder their results are always out there.

      See, this is the kind of hijinks that happen in polling. Luckily we have great crowds of these organizations to cancel one another out.

    • Craig

      Right – it would have been just as improper to use the 2006 midterms as the demographic baseline for 2008. You may have ended up being right in the end, but it would have been for the wrong reason. Midterms and presidential electorates are just very different universes.

    • Matt McIrvin

      Wenzel’s the one that has Romney ahead in Ohio, right?

      Just heard a guy over on Robert Wright’s Atlantic blog saying that Long Island Newsday’s endorsement of Romney indicates that Obama will lose Illinois.

  • Matt McIrvin

    I’m assuming that you won’t be able to rely on the “great crowds of organizations” for long: state poll aggregators have gotten so much attention this year that sockpuppet spamming of polls is coming, if it’s not here already.

    • MAT

      Yea, the great polling arms race of 2016 should be interesting. The fun will be seeing if both sides play along equally. If so, the PEC should continue to work. If not, then something like Pollster’s rolling house effect adjustment may be needed.

      Also, if PEC, 538 & company do hit the target this year, will the media focus shift to them for the horse race commentary in the future (thus encouraging what Matt McIrvin suggests above), or keep breathlessly reporting every poll. I think I know the answer, but…..

    • Philip Diehl

      Partisan parity in the polling arms race is not likely. The GOP has three big advantages: thanks to CU, they have more money than the Pope to out-poll the Dems; their polls are cheaper and faster to conduct because they use robo calling and ignore cell-only voters; and robo calling and excluding cell-only voters automatically tilts their results GOP.

      The GOP will inform their polling strategies with an analysis of the aggregators’ modeling methodologies, taking advantage of the aggregators’ transparency, and craft their polling strategies to manipulate MSM and public assessments of the state of the race.

      So, something along the lines of a house-effects adjustment will be necessary by 2014 (manipulation is easier in off-year elections) and no later than 2016.

  • E L

    Now the Main Stream Media is ” its own polls to quiet the right wing and keep the suspense going:

  • John

    The PEC’s best guess on the probability of an Obama win is currently 99.8%.

    Can you please give a high level explanation of why PEC’s model is so much more sure of an Obama win than the other quantitative models out there (e.g. – Nate Silver) and the prediction markets (e.g. – Intrade)?

    In particular, Intrade is currently estimating an Obama win at about 2 out of 3 probability and is offering 1-2 payoff on an Obama win.

    If your model makes objective sense, then why aren’t objective people driving Intrade’s odds much closer to an almost sure Obama win? I mean, it’s free money on the table for the taking if your model reflects anything close to reality.

    • Matt Davis

      Dr. Wang already has given that explanation in an earlier article discussing the different aggregators’ approaches.

  • ArtsyJane

    How do you figure that it’s +2 to 0 change for the Democrats in the Senate?

    From where I stand I see

    Elizabeth Warren
    Tim Kaine
    Claire McCaskill
    Tammy Baldwin
    and a few others

    all pretty likely to win. And they are all new seats for the Democrats.

    Please explain.

    • Philip Diehl


      The picture is complicated, but it boils down to:

      * Baldwin is not in the bag and Kaine is close, though he’ll probably win.
      *Retaining Montana will be tough but it’s possible
      *We’ll lose North Dakota and Nebraska.
      * Indiana will be a pickup as will Maine assuming King caucuses with the Dems, which has a likelihood of about 90-95%.

Leave a Comment