Princeton Election Consortium

A first draft of electoral history. Since 2004

The 1-point Sandy bounce?

November 5th, 2012, 10:02pm by Sam Wang


(Today’s our all-time high of traffic – over 300,000 350,000 views. That’s three times our 2008 traffic. Welcome, readers of Andrew Sullivan, Paul Krugman, Wired, Deadspin,Kevin DrumPeter Norvig, The BlazeThe Blaze? ZOMG!!! I thought you were PECer-heads! Seriously…everyone go read that piece. Alternate reality.)

Some of you expressed concern at an “unnerving” drop in the EV estimator today. I’ll make a confession: part of that was a recent data glitch. Pollster.com was feeding us the rolling daily Ipsos/Reuters  averages, which were redundant, especially in North Carolina. It’s fixed now. After the dust clears we’ll scour for similar earlier mistakes.
>>>
Where I live, Halloween was postponed until tonight because of Sandy. The last of the trick-or-treaters have trickled by.

Judging from the return toward semi-normal life here in hard-hit New Jersey, voting mechanics should only be affected in some regions, and less so in other (less blue) states. So the direct physical effect of Sandy is probably small. But what about indirect political effects?

Pundits on the right have claimed that Sandy “stopped Romney’s momentum.” Hmmm. First, let’s point out yet agaain that in terms of state polls, Ro-mentum stopped on October 13th or earlier, the week after Debate #1. The pattern was one we’ve seen before: a force drives opinion away from a Meta-Margin “set point” of Obama +3.0%, then relaxes back – like what happens after you pull on a spring. We’ve been seeing a version of this in the form of steady movement toward President Obama.

This recovery has not been visible in national polls, which until this week have been flat. I am puzzled about why they would differ from state polls in this way. Maybe nonswing state voters were less attentive to political events after Debate #1.

However, Sandy certainly caught everyone’s attention. Maybe that’s why national polls (which, although of questionable absolute accuracy, are still good for detecting relative change) have done this:

That appears to be a 1.0-point swing toward President Obama. This will require further analysis. But it suggests that as I prepare final predictions, November 1st is a date to look for late-breaking trends. This is a challenge for prediction, since I had hoped to integrate data over a longer period using a secret sauce (did you click on that? secret no more).

>>>

Tonight I’m preparing final predictions, as well as collecting useful links for Election Night – a Geek’s Guide. Regular readers, I seek your help. In comments, could you please cite your favorite places for various bits of information:

  • election returns as they come in;
  • early voting;
  • liveblogging;
  • what you’ll be watching for; and
  • other useful information?

Tags: 2012 Election · President

167 Comments so far ↓

  • Keenan Walsh

    Hi Sam,

    I just want to thank you for your work on this site. I only discovered it at the end of September to be honest, but it took the edge off my days–not just because you give Obama good odds, but also (perhaps more so) because the intellectual stimulation curbed a lot of my visceral anxiety. Because of this site, I was able to approach the election with more curiosity than dread, regardless of the outcome.

    That is a remarkable thing–to ease someone’s worry. And I doubt that I’m the only one you helped in that regard.

    So, thank you. Let’s cross our fingers for tomorrow.

    -Keenan

    • Khan

      Ahhhh, no way Jose. Even the utterance of that R word gives me the shivers. It’s like Beetlejuice–say it too many times and it happens.

  • Amitabh Lath

    Other than the top ticket races, I will be looking for Bill Foster to win his congressional race. I know the House is probably not going to change hands, but I would like a little Sandy-inspired coattail effect downticket.

    PS: We’ve been without power since last Monday due to the hurricane. But if Sandy can push a couple of house races +D, then it will have been worth it.

    As for returns, I think the CNN website is a little quicker than most.

  • Martin

    Chances of voting irregularities in FL and OH tomorrow night thanks to republican sec’s of state shenanigans: 100%!

    Chances of the PEC site crashing tomorrow night: 99.9999999999999%!

  • tarylcabot

    In answer to your question about favorite places for information, I still prefer the cable news shows: by design, the cable operators place them next to each other, so CNN, Fox, FNC, and MSNBC can be scanned quite easily. By contrast web sites get so much traffic, the updates are often slower. I only care about results as they come in & projected winners of president & the handful of ‘tossup’ senate seats.

    Live blogging is good for debates, but not as enjoyable for election results.

    The one exception for me is CA propositions – I usually go to LATimes.com. Local news isn’t an easy button push away.

    • Ralph Reinhold

      I prefer the foreign press. They usually don’t
      select the stories based upon a preconceived notion of what’s important to their biases. In particular, I like the Guardian, Economist and stuff.co.nz. The latter usually has one or two US stories that are important to me that I don’t find from the normal US sources unless the newsfeeds from UPI, AP or Reuters.

  • Michael

    I was interested in your description of the dynamics of this race since mid-summer, with occasional impulses (e.g., the first debate) followed by a spring-like reversion to the Meta-margin “set point” of Obama +3.0. So I went to the post concerning your first prediction of this cycle, in which you pointed to the 2004 and 2008 race dynamics as a guide to the 2012 prediction. I can see how the data has informed this year’s forecasting model, along with a certain intuition about how “locked in” most voters appear to be regarding their candidate preference. I wonder, however, whether this sort of approach will yield a useful predictive model in future elections. Do you think at all about the possibility of a random process characterized by regime shifts or other sorts of dynamics that would not be captured by this year’s model?

  • Pat Brown

    Thanks Dr Wang and Andrew for keeping me sane during the last few months. I am a lurker and hope to remain that way but needed to voice my appreciation. I have stayed away from Facebook but had to join to keep up during your blackout. I have become obsessed with this election and the end can’t come to0 soon. I have sent all my friends to your site.

    I follow live blogging on TPM and Andrew Sullivan’s Dish. I listen to NPR or other radio to get election returns and seldom watch TV. The Election Night Helper looks as if it will be useful.

    On the radio just now a caller was complaining about the “liberal media.” The Host said it was because they were listening to all the college professors. The caller-in tally is Obama 7 Romney 5 on a Republican Host show.

    Thanks again for all your numbers that a lay person can follow and I look forward to you final
    prediction.
    Pat L

  • Dean

    Thank you so much, Dr. Wang and Mr. Ferguson. I have been visiting this site since around the time of the NPR interview. I have been talking about this site to my friends and family. I’m no mathematician or statistician, by any means, but I tell people that this site strives to offer very precise predictions, with as little extraneous “clutter” as possible.

    I’m sorry I have no advice for following the election that’s better than what people have already offered in this thread. I’m a Chicagoan who will be watching the election on TV, only a few miles from where the president will be, literally just south on Lake Shore Dr.

    I first saw President Obama in Springfield, IL when he was a state senator supporting my union, AFSCME, around eight years ago. He took his time to speak to us in the statehouse rotunda, when other politicians ignored us. I feel that so much is at stake in this election, that so much of what we fought for (union rights, government safety net programs, women’s rights, etc.) is threatened by Republicans and Romney. I also fear that what belongs to We the People, government services, will be privatized.

    • wheelers cat

      Dean
      That is what is so great about this site…Dr. Wang cuts through all the clutter and pedantry so you can see the power and the beauty of the maths.
      My advanced calc prof (Insane Husain we called him) used to splash a great big proof up on the board, and then step back, spread his arms wide, and declaim– see the beauty!
      I remember the beautiful celebration in Chicago in 2008, and people laughing and celebrating in the streets all over the country.
      Thanks to the Nerdocracy of math-heads, like Nate and Sam and Drew, the pundits and republican claquers havent been able to steal the will of the people.
      We have overcome.

  • Greg P

    Dr Wang, thanks again for a great site. Last time, you provided very precise estimates of the electoral, but I am used to a confidence interval. Just personally, what electoral vote count would you consider to to be so low that you would reconsider the validity of the assumptions of the model (the median poll is the most accurate, etc). Or more more succinctly, what number of Obama electoral votes would be your “eat your shorts” number? Would it be 290, or simply anything less than 272?

  • Suja

    I usually end up watching CNN, even if it makes me stabby. Then I confirm their calls with others.

    Specifically watching for NH and VA to see how it all matches up to polling numbers. Locally, I will be watching for the results of the Kaine-Allen matchup in VA, the fate of Prop 6 in Maryland, and further afield, what happens with the McCaskill-Akin race and Mourdock-Donnelly race in Indiana.

  • David

    Hart’s Location, Obama 23, Romney 9, Johnson 1; Dixville Notch, 5 to 5. That’s your results fix until tonight.

  • Jeffrey Sax

    For those who want to dig deep into precinct level details (vote counts and demographics) of the 2008 election, there’s the Stanford Election Atlas (http://atlas.esri.com/Atlas/VoterAtlas.html).

  • pechmerle

    And Dixville Notice has reported:

    Obama 5. Romney 5.

    Exactly as the “it’s all a toss-up” horse race folks have said. ;-)

    • RDT

      As reported above, Hart’s Location was more helpful: Obama 23, Romney 9, Johnson 2.

    • pechmerle

      RDT: Thanks, I hadn’t noticed David’s post a few minutes before mine. (Maybe it wasn’t visible yet, I don’t know.)

      Obviously now we don’t need to see any other results. ;-) Or as we sports fans like to say, “Stop the game right here!”

  • Seager Mason

    My mind is now at rest regarding the potential for computerised voting fraud.

    http://www.theawl.com/2012/11/the-truth-about-voting-machines

  • Elio

    Thanks for the great work and analysis. I’m sure you follow it already but I agree with the other poster that Talking Points Memo does a great job putting out live results on election day. Other than that, Sullivan will provide some emotional color commentary and the cable new stations (which I generally don’t enjoy) are useful for a sense of the developing conventional wisdom and comedy.

    I want to put in a request for some detailed post-game analysis. Beyond just the raw score-keeping we’ll get other places, it would be nice to see if there was evidence for Nate Silver’s high uncertainty compared to what you and Drew Linzer are predicting. Since the states may have correlated shifts from the polls, it isn’t clear to me what the best way to compare across the models would be.

    • wheelers cat

      I think Nate is converging on Drew and Sam as we speak.
      All Praise the Holy Central Limit Theorem.

  • Rusty

    If you’re looking for important information, see if it’s raining or snowing in Northeastern or Midwestern cities, especially (obviously) in swing states. Bad weather depresses Democrat votes. It’s not a trivial effect, either.

  • 538 Refugee

    However it goes later today, I’d like to thank EVERYONE that made this an enjoyable site to visit. Sam and Andrew may have provided the meat and potatoes but the posts and links were the spice and just as important to me. Sam’s analysis is a great anchor but doesn’t exist in a vacuum. Again, thank you all for your HONEST efforts in helping to bring a complete picture of the data into sharper focus.

  • Steve Schlichtenmyer

    I enjoy PEC. For early polling data George Mason Univ. has the most comprehensive data:
    http://elections.gmu.edu/early_vote_2012.html

    Keep up the great work. I’d love to see you provide your analytic expertise to major social issues, e.g. the effectiveness of incarceration in preventing crime (three strikes comes to mind), various issues in education, etc. Thanks again.

  • Karen

    I know everyone says OHIO, OHIO, OHIO but I will be watching VA because it’s in the first wave of poll closings and if Obama wins I’m pretty sure it’s over for Mittens.

    I will be checking out Andrew Sullivan’s live blog, watching MSNBC (can’t help it love Chris Matthews) and DVRing the live DailyShow/Colbert

    • Ralph Reinhold

      If Obama wins VA, he has won because of the large margins in three of the swing states.

  • Karen

    Oh, and at some point I will raise a glass to Tim Russert. Still miss that man like crazy.

  • SomeNorCalGuy

    Election Returns: I’ll be checking the returns habitually at nytimes.com. Their paywall goes down at 6:00 EST so I’ll be there all night. I’ll also be watching CNN for the most part (unless they break out the damn holograms again. Then I’ll be watching the Avengers on mute and listening to NPR.) I’ll also occasionally be checking in on MSNBC and [shudder] Fox News. If Bill O’Reilley’s head looks like it’s going to explode and Chris Matthews is as giddy as a schoolgirl, you know it’s going to be a good night.

    Liveblogging: Andrew Sullivan, all the way. He’ll also be on Colbert, so, bonus.

    What you’ll be watching for: First things first – Virginia is one of the first states to have the polls closed at 7:00 EST. If that gets called quickly for Romney, that’s not a good sign. But if it’s too close to call, I’ll remain optimistic. If it gets called quickly for Obama, then the game’s over for Romney.

    Then at 7:30 Ohio and North Carolina have their polls close. Nobody will be calling Ohio for a good long time. Exit polls could have Obama 60-40 in Ohio and they will not call it. It’s just too crucial. North Carolina less so. If it gets called quickly for Romney, that’s okay. If it’s too close to call that’s good for Obama. If it gets called for Obama, then that’s another sign that it’s game over for Romney.

    But the Super Tuesday of the night is 8:00. Some semi-swing states at that hour: New Hampshire, Michigan, Pennsylvania plus Florida – another state that nobody is going to call anytime soon. But there is another bellweather in New Hampshire. If it gets called for Romney, then keep a wary eye on Iowa, Colorado and, possibly, Wisconsin.

    Other useful information:
    A map of poll closing times: http://www.politico.com/news/stories/1012/82595.html

    512 ways to the white house: http://www.nytimes.com/interactive/2012/11/02/us/politics/paths-to-the-white-house.html?ref=politics
    Cool interactive graphic to explore electoral possibilities as states get called

    Another cool tool for playing around with state results: http://www.270towin.com/ (Which I’m sure everyone is familiar with) Not only do they let you play with the electoral college, you also get to play with both Senate and House maps. (Warning – the House Map crashes my Flash, but that could just be me)

    Oh, also I live in California, so there’s more than a few propositions I’ll be looking for, in my state and others. Gay marriage, the death penalty, campaign finance reform, marijuana legalization – even GMOs are on the ballot in mine and other states. I’m very interested where we stand as a nation on these issues.

    • BrianTH

      Just a WAG, but I think if the normal criteria allow an early-ish call in Ohio, at least one network will do it, on the theory that the networks are still trying to be “first” with breaking news. And in fact I would specifically guess that network would be Fox.

    • SomeNorCalGuy

      BrianTH I disagree for two reasons:

      1.) Generally speaking, I believe that repubable news organizations (and Fox) have a general sense of journalistic integrity. That means they believe that being right is more important than being first.

      2.) The only thing more important than getting viewers the most accurate up-to-date information is getting viewers. Since it is generally agreed that whomever wins Ohio wins the Presidency (all things being equal) then I think most viewers will shut off whatever they’re watching shortly after Ohio is called. The news organizations need to create the illusion of drama on election night (or, even better, actual election night drama) and they’ll keep stringing viewers along until even a blind dead dog can call Ohio (and Virginia and Colorado, possibly Florida too, plus New Hampshire and Iowa).

      Bottom line, it’ll be midnight Eastern until reputable news organizations call Ohio and the winner of the election is clear. (Of course if Ohio really, really is too close to call, as well as Virginia and/or Colorado, then the results will still be in question Wednesday morning and for several weeks to come. Oof. Just the thought is making me sick.)

  • pechmerle

    It seem obvious that the most intelligent commentary (other than in posts/comments here of course!) will be dispensed on the Daily Show.

    I can’t think of a journalist who has done a better job during this cycle than Jon Stewart. All that, and plenty of LOL’s to boot. Can’t lose proposition.

    (Special commendation: the comments here at PEC by wheelers cat, which were the wittiest and most literate by far.)

  • Some Body

    A data question: does the national polls chart take into account those trackers that stopped tracking because of the storm (most notably, the greatest R-outlier, Gallup)? Knowing Sam’s preferences, I’d guess it doesn’t. But then, this could explain part of the jump (Gallup came back yesterday, with an R+1, which is a change from the R+5/6/7 territory they were in before, but they’re still dragging the median one poll down).

  • Chicky

    Sam, Andrew, and all commenters:

    I’ve been a first-time lurker since early October. I’m a US citizen/voter living in London. I’m a news junkie and certainly an election junkie, as well as a member of Democrats Abroad. It goes without saying, probably, I’m an Obama supporter.

    I found this site to be the only place for truth, not punditry, on this here inter-webs. Having found it, I don’t think I will ever, EVER, listen to a pundit again because time and time again has shown that my guess, my punditry, is no worse than theirs. Sam and Andrew, though, have the science/numbers/facts, not the pointless wishful thinking. Plus, the occasional not-un-called-for snark.

    Then, there are the commenters: wheelers cat, 538 refugee, khan, Ms Sheckley, to name a few, that have truly brightened my day every time I see their commentary here.

    Thank you for a truly lovely experience.

  • Olov (not Grinde)

    Thanks PEC!

    I don’t know what they’re doing for tonight, but the Economist’s liveblogging of the debates at democracy in America is usually wicked funny, and suits my biases to a tee. so I’ll be hanging out there. Since I’m in Sweden, it’s probably going to be hard to find the right channels on my TV. And however much of a democrat I am, I don’t think I can stand an evening of MSNBC punditry, even if that will probably be streamed online. Nonetheless, Chuck Todd: I wish I could quit you!

  • Nancy

    Why suffer through CNN when you can watch on the eminently rational, thoughtful BBC? I watched their coverage in 2008 with a UK friend and was so glad I did; the instant I turned to CNN and its ilk I wanted to hurl myself off a cliff. The Beeb is an oasis of calm and sanity.

    http://www.bbc.co.uk/news/world-us-canada-20103937

  • jhm

    I tried making my own forecaster last cycle, but I couldn’t get the lighting bolt to animate it.

  • Marty

    I want to add my heartfelt thanks to Sam, Andrew and the regular contributors.

    I live in Canada, and the coverage I will be listening to will be Micheal Enright, one of our best ‘talking heads’, who is in Washington DC but will be broadcasting on CBC.

    • Rose

      I second that.
      Thank you! I am glad I stumbled into this site the week or so before elections. It kept me optimistic.

  • wooderson

    Did anyone read the comment section of that blaze article? absolutely astonishing…

Leave a Comment