Princeton Election Consortium

A first draft of electoral history. Since 2004

Blown away

October 29th, 2012, 11:11pm by Sam Wang


Not by Sandy, though the winds are pretty high (check out this excellent moving representation).
I’m more blown away by the web traffic over the last few weeks. Take a look.
The above screenshot is what I see backstage. Traffic has doubled every 1-3 weeks since July; lately the doubling time has been 6 days. We’re running at about three times the traffic we got in 2008. Today, the Princeton Election Consortium received nearly 200,000 views. Note the number of referrals from Reddit, Facebook, and Twitter. That’s a new development. The list used to be heavier on politically-oriented sites such as Andrew Sullivan, Kevin Drum, and Balloon Juice. Liftoff?

Thank you all for reading and visiting. Andrew Ferguson and I will try to keep it interesting in the remaining days of the campaign.

Tags: 2012 Election · Site News

243 Comments so far ↓

  • Joel

    Sam, you’re in the bigtime now! Pretty soon, you’ll get invited to all the exclusive clubs, so that you can rub elbows with our innumerate elites:

    http://www.youtube.com/watch?feature=player_detailpage&v=171FURqSIQc#t=18s

  • don in fl

    looked at the 12:10 report, romney had no blue in the ev distrubition graph.good. hope it stays that way.

  • SurfaceThought

    I can only begin to imagine the absurd backtracking that unskewed polls, Rush Limbaugh, the Fox newsters, and other conservatives who declare that Romney is going to win in a landslide are going to have to do on November 7th.

    Actually, come to think of it, its sort of scary that sort of narratives that might emerge.

    • Matt McIrvin

      They won’t do any. All that stuff will go down the memory hole in an instant.

    • SurfaceThought

      Really? Even the unskewed polls guy? After managing a website for half a year whose sole purpose is to discredit the vast majority of polls validity, you don’t think he is going to try to attribute a Obama victory so some sort of conspiracy rather than coming out and saying “Despite the presence of house affects in individual polls, the polls as an aggregate had no systematic skew. Basically, everything I have said for the past six months is wrong, and this was a failed enterprise.”

  • Mary

    What is up with this alarming post by Weigel on Slate about Gallup’s early voting numbers for Pres. Obama looking “terrible”?

    He writes: “Gallup asked voters whether they’d cast ballots or intended to before election day. The early voters broke 52-46 for Mitt Romney. The dawdling voters who would vote before election day were tied, 49-49. The voters waiting for November 6 broke for Romney, again, by a 6-point margin.”

  • Amy

    As an epidemiologist, the geek factor of your statistics discussions is beyond enjoyable. That you assuage my angst over the election is even better. So glad to see a dissection of how a little bit of knowledge can be dangerous in swaying the uninformed. Keep informing and educating! The more i see people talk about your site and how this kind of analysis doesn’t take into account all the “parameters” the more i wonder about our education… but then someone posts a response explaining how using the mean (most likely) accounts for potential biases or unmeasured cofounders. Yay!

  • L. Murray

    Well, I know that I have been telling all my fellow Democrats that this site will help calm them down.

    I had to evacuate because I live on the Quinnipiac River in New Haven. The friends I was staying with lost power but we could still text to friends who had The Internet and could give us Princeton Consortium updates.

    Now I am back with power restored and can check every few hours myself.

  • Erin O'Daly

    Ah, up to 2.32% at the 3:01 update! Sorry, this site is like polling crack to me. It’s just nice to get the real, unbiased, unvarnished information. We all have the tendency to cherry-pick the polls we like, and discount the ones we don’t, and to be obsessive about a single poll. So it’s great to come here and just see all that mess of information condensed into a single top line number.

    • Debbie

      “polling crack”….Touché lol.

    • S_Andersen

      Problem? What problem? I don’t have a problem! I can stop reading Wang whenever I want… I’m just not ready yet.

  • Ohio Voter

    I can’t help but notice the pageviews graph looks incredibly similar to Meta-Margin graph.

    • skmind

      See, it is that kind of a joke that gets taken as gospel by rightwingnuts. I can imagine a wingnut blog that actually will cite that someday

      “Dr. Wang’s stats are based on how many hits he gets . Ha ha ha!”

  • Rayongamma

    Here is a question for good statisticans like Sam. Instead of calling the winner of the election, one could ask what are the chances that the presidential election result might not be known by midnight EST (given delays by hurricane Sandy, absentee ballot counting etc.)? I am wondering if I should stay up til 6AM watching CNN from France… Is it negligible (say less than 5%, possible (5-50%), probable (50-95%), or nearly certain (>95%)? Given the opinion polls, one can try to factor in with some Bayesian priors the delays in reporting from absentee ballots, from precincts operating under candlelight, etc.

    • Richard Vance

      Yes, I’m wondering if we have to wait for Hawaii to close to find the final result:-) Not really, But it would be an interesting bet on what state called by what network puts OB over the top.

  • wheelers cat

    so here is a tasty bit of empirical data.
    im a technodroid and a culture maven, so I thought I would compare a couple of celebrity endorsements.
    Here is Clint Eastwood for Romney
    https://www.youtube.com/watch?feature=player_embedded&v=on3rmLq0e3g#!
    31k views since Oct 24
    Here is Joss Whedon for Obama
    http://www.youtube.com/watch?v=6TiXUF9xbTo
    1,814k views since Oct 28

    Does pop culture count? Can a party that is wholly disenfranchised from American contemporary culture actually win?

  • Karen Carroll

    So the prediction of 98% is really nice, and I believe the model, in that if people actually vote their intentions, the model will be vindicated. However, what if people can’t vote because of the hurricane damage? In 2000, not only was Gore the popular vote winner, but he was also the choice of the majority of Floridian voters, had it not been for the butterfly ballot, the purging of the voter rolls, and other shenanigans. I’m very concerned about a repeat. Anyone else?

    • Dean

      I was told by some that they fear voting shenanigans more than polls and Ro-mentum. I commend anyone who fights against voter suppression.

      Since we’re discussing this site’s growing popularity, I have a thought I would like to share that is amusing me. What if, in some solitary place, away from everyone, or at night when everyone is asleep, Romney stares glumly at the MM and EV estimator?

    • wheelers cat

      @Dean
      probably not. Romney is a bot, he likely dreams of electric sheep.
      but i guarantee Bibi Netayahnu does.
      ;)

    • xian

      well, everyone says he’s a “data guy”

    • Alex Simmons

      Fortunately, most of the the states affected by Sandy are either safely Democratic or Republican –and even turnout won’t change their dynamics.

      However, there is one exception that I worry a bit about and it is Virginia. I’ve sort of conceded Florida to Romney so it really doesn’t figure into the important and lynchpin electoral math.

      Virginia is, by all scholarly accounts, at this point leaning Obama. Turnout volume here could definitely affect the outcome since it is probably within a few hundred thousand votes either way. But Virginia wasn’t devastated by Sandy the way that more northward states were.

      All of this is not obsessively worrying to me though, since Ohio is pretty clearly Obama’s and there is no way for Romney to win — even if he takes Virginia. Also, one third of the voters have voted in Ohio early and the anectodal issuance is that Obama leads by 2-1 margins. There’s really not too much that Ohio could do to put their hand in the political till because they are being highly scrutinized and Obama’s GOTV effort is emblamatic of consumate efficiency.

      At least that’s the way I see it.

    • Richard Vance

      My evangelical cousins in Ohio are being heavily affected by the Ralph Reed pew voter guides. Many of them were wavering but are now pro-RR and telling their friends. I have been telling them how much evil dirt is involved in the Christian Coalition and how can they vote for a Mormon Bishop who lies? They are attracted to Ryan and his abortion stance, the Ryan choice was really a good shore up the base move that’s paying off now when the ground game is what counts. The pews will be active and I’m worried. Yes I know the odds of 98% but that is NOT 100%. I work with software safety critical systems and 98% won’t get me to sign you off as ready to fly, no way. When Romney’s chances get down to less than one in a million I will relax a bit. I suggest that all of you sign up to do calling to swing states on behalf of the OB campaign. I called Colorado in 2008, the red red Colorado Springs area and was pleasantly surprised to find people willing to talk and wondering why they had not already been called!!! We do need intelligent well spoken representatives to do the calling ant that you folks. All that’s required for evil to prevail is for good men to do nothing – attributed to various people, usually Edmond Burke.

  • Brian

    Certainly is nice to see the “watch zone” above the 270 line. Now if only it would stay there for one more week…

  • jd351

    Well does anyone have any insights into this article. I find it amusing at best..

    http://news.yahoo.com/blogs/signal/national-polls-meaningless-stage-election-135415586.html

  • Froggy

    Nate Silver has posted a series of messages on his Twitter feed (https://twitter.com/fivethirtyeight) raising questions about Pharos Research, which has released a number of state polls in the last few days, notably an OH O+3 poll today. He cautions, “Really not sure that aggregators should be including ‘polls’ from Pharos Research Group in their averages.”

    Looks like the only thing that would change if we removed Pharos Research polls from the model here would be a shift of OH from O+3 back to O+2.5, where it was before the 3pm update.

  • Olav Grinde

    Officials hand-copying ballots in Palm Beach County

    @Karen Carroll, I noticed this news from Palm Beach County, which I wouldn’t exactly call reassuring:

    “Earlier this month, Supervisor of Elections Susan Bucher announced that more than half of some 60,000 absentee ballots contained a printing error. The mistake wasn’t in the proofs, she said, so her office mailed them out without noticing headings were missing over the names of Florida Supreme Court justices up for retention.

    Richard, who now represents three justices on the ballot, said the lack of a heading could confuse voters who might lump the justices in with candidates for other races.

    But the mistake has national implications, too. Without the headings, tabulation machines couldn’t read the ballots. So Bucher ordered teams of workers — supervised by herself, a county judge and a county commissioner — to hand copy votes onto good ballots. With more than 27,000 ballots streaming into the office, Bucher anticipates the copy frenzy to last through Election Day.

    On a recent afternoon, hundreds of workers buzzed over their tedious labor deep inside an industrial park north of West Palm Beach.”

    (The quote is from the Sun Sentinel, the emphasis is mine.)

    http://articles.sun-sentinel.com/2012-10-27/news/fl-palm-elections-supervisor-problems-20121024_1_absentee-ballots-elections-susan-bucher-election-day

  • Reason

    I lost power so I have been going through withdrawl. Wow! What a nice thing to come back to! Awesome. And Did anyone see Gov Christie of NJ praise Obama and his leadership, invited him to come tour NJ, and told Romney not to come for a photo op?

  • don in fl

    did christie tell romney not to show for photo opt.priceless if true.

    • Tim in Brooklyn

      In effect. Said he had no interest or time for one. Busy dealing with Obama, who has some power.

  • Reason

    @don, he did. He said he does not give a damned about presidential politics, Romney needs to stay out of NJ, and he praised Obama up and down. On all major networks, even Fox. He said his job is to take care of the people of NJ. So for being a Gov, not a Romroid, I give him credit.

    • Michael

      Eh, Christie is a red governor of a blue state experiencing a massive non-political crisis. In this case, his being apolitical is political.

    • David Mann

      If Romney wins, Christie can’t run in 2016. Let’s give him credit for serving his state, but let’s not give him too much credit either.

    • E L

      Christie is now DOA with the Republican base for a 2016 presidential run because he supped with the devil, but he greatly increased his reelection chances as governor in blue NJ.

  • Reason

    @Steve, damn, should have waited 4 minutes. Down to 2.3 now. NC back in R column due to drop off I guess. But still great. I think NC will be like 2008. Very close and winner not decided there until a few days after election.

  • Olav Grinde

    President Obama has a fair shot at winning Virginia and/or North Carolina and/or Florida. I would not be surprised if he takes all three.

    One thing that does surprise me, however, is the Democrats lagging almost 3 % in early voting in Colorado.

    • mediaglyphic

      Olav, i noticed that also. and it worries me. I wonder if there are any granular precinct by precinct comparisons for Colorado. It would also be nice to see what the early voting tally in Colorado was at the same time in the last election (2008).

    • Olav Grinde

      MediaGlyphic, a big question, of course, is which way the so-called Independents will break. Also unsure about third-party candidates in Colorado.

      But so far, with 965,510 votes cast, equivalent to 39.9% of the total 2008 vote, we have:

      Democrat 35.6%
      Republican 38.4%
      None/Other 26.0%

      Hopefully those numbers will change in a Blue direction in the days ahead. In 2008, just under 80% of Coloradans voted early.

      http://elections.gmu.edu/early_vote_2012.html

    • mediaglyphic

      True enough about independents, that is why it would be nice to be able to compare by precinct or same date last cycle. I don’t know if this data is available.

    • mediaglyphic

      Olav, i also read the following
      ” 220,000 fewer Democrats have voted early in Ohio compared with 2008. And 30,000 more Republicans have cast their ballots compared with four years ago. That is a 250,000-vote net increase for a state Obama won by 260,000 votes in 2008. ”

      a red leaning guy Adrian Gray (can you believe the name!) wrote this, but i can’t find a source. is he just blowing smoke?

    • Olav Grinde

      I would love to see where Mr Gray gets his overly-red numbers. To me it sounds like he’s bluffing or quoting an unreliable source. The problem, as I understand it, is that Ohio does not have party-specific registration.

    • Steve in Colorado

      I wouldn’t put it past SOS Scott Gessler to be somehow cherry-picking the reporting here in Colorado. But, maybe not. Hopefully this state will not be needed.
      But for OH, I bet that Adrian Gray is referring to people who voted in the GOP primary as Republicans. Well, there wasn’t an interesting Dem primary, so of course the numbers will look weaker than in 2008. Fox New and Time had polls showing Obama up nearly 2-1 in early voting in OH. (I believe it was 63% to 36%)

    • mediaglyphic

      Steve in CO, thank you , i hadn’t thought of that as even a possibility, if the diary of adrian gray is based on this then he must be thinking he is in real trouble

    • wheelers cat

      ummm…guyz.
      Marijuana is on the ballot. No way Romney is carrying Colorado.
      Let me remind you all, Ras whiffed on Colorado in 2010 by 5 points.

    • Olav Grinde

      Steve, yes, I remember those figures. The PPP poll released on Sunday stated:

      “Obama’s built up a big lead among early voters in the state. 36 % say they’ve already cast their ballots and they report having voted for Obama by a 63/36 margin. Romney’s up 53/45 with those yet to vote.”

      Given that Obama has a 27 % lead amongst the third who have voted, Romney’s 8 % edge among those yet-to-vote seems insufficient, to put it mildly. Unless there’s a huge surprise (such as vote counting or arithmetic), it seems more than probable that Obama wins Ohio.

      http://www.publicpolicypolling.com/pdf/2011/PPP_Release_OH_1028.pdf

    • Martin

      @Steve in Colorado

      Thanks for the link but am wondering if you have any info on the early voting stats from this point in 08???

    • Ohio Voter

      I can’t speak for the whole state, but in Columbus early voting is up x2 from 2008. And Franklin County is solidly democrat.

      Also correct, Ohio does not have party registration. Your ‘party’ is whatever primary ballot you requested last. But you can change that every election if you want.

    • ML

      Ohio doesn’t have party reg, they classify Dems/Repubs by whether or not they voted in the primaries. So the shift in party ID is meaningless given the lack of competitive primary in OH this year as compared to the hotly contested one in 2008. Better to look at which counties have big turnout.

  • zenger

    2.3 to 2.0. At this rate, Romney wins by Christmas!

    (No, I didn’t do the math).

  • zenger

    Oops. Make that 2.32 to 2.3 (same punchline)

  • James P

    The media has done a horrible job with the polls and do not understand that sites such as yours and Nate’s are the best ones to follow. Wisconsin is not a swing state and people think its close there. If a 5 point lead is close, then so are a bunch of states. Romney has to win FL and OH. There is no way he wins since NV is entering the Obama column.

  • Angochay

    Not sure if it has been suggested, but I think a really nice graphic (especially if you end up predicting close to the final result) would be an animated gif of the median estimator line moving across the graph according to the changes in time with that beam of light showing the prediction projected on the right side of the graph. This will nicely illustrate your target prediction as it evolved over time and how it always remained within a certain spectrum. Were Obama to win, pundits will be suggesting the election broke towards him late. A nice graphic will put that assumption in perspective.

  • Steve

    Can someone explain to me how Huffpo Pollster works. They have moved Ohio from Lean Obama to tossup back to Lean Obama and as I write this back to Tossup within the last 24 hours. And yet they have added 4 polls showing Obama leads of 3 to 5 points in that timespan. What gives?

    • InmanRoshi

      My guess is the following series of events….

      a) Obama was a Lean Ohio to start yesterday.
      b) Rasmussan comes out with a poll showing Romney +2 in Ohio (*cough* *cough*), which is enough to push the average back to tossup.
      c) A few small polls come out showing Obama Ohio leads today, pushing the state back to Lean Obama.
      d) Nate Silver voices on twitter concern about the credibility of one of today’s pollsters (Pharos Research) claiming he can’t verify the pollster or their business model, and upon reaching them they can’t answer simple questions (like where their calling center is located). Silver then says on twitter than the polling aggregators should probably exclude them for the moment until they can be properly vetted.
      e) Huffpo/Pollster follows suit and removes the Pharoh poll, moving the state back into tossup territory (barely).

      https://twitter.com/fivethirtyeight

    • P G Vaidya

      As I recall, they take only the recent most post from each pollster. Therefore, if a pollster had Obama up by 5 in a previous posting and a new one comes out showing him up by 3, their overall rating goes down.

  • mediaglyphic

    Joe Klein says he doesn’t know
    http://swampland.time.com/2012/10/29/i-dont-know/?iid=sl-main-arenapage

    and that anyone who claims to know is blowing smoke.

    well Joe, what is knowing? modus ponens, “if p implies q then if p is true so is q” its more a question of how do we know than if we know (epistemology). Joe Klein i thought you were better than this.

    • Obama 2012

      lots of crazy comments there (the article itself was bad enough… but the comments are nuts.)

    • E L

      “But anyone who claims to know who is going to win is blowing smoke.” —Klein. Joe is a veteran reporter who knows that a comment like that puts him beyond later criticism no matter the result. It’s the perfect veteran reporter stance.

  • Eric

    Dr. Wang (or anyone else with expertise in this topic), how frequently has someone won an election without winning his home state? How significant will it be that Romney/Ryan are losing in all three of their home states? (MA, MI, and WI – plug losing in NH, Romney’s vacation home state). If R/R manage to win surely this state of affairs – no pun intended – is unprecedented, no? If they lose, as most expect, will this be the first time that both candidates on the ticket have lost their home states? In 1984 when Reagan won a landslide Mondale still managed to win his home state.

  • Ms. Jay Sheckley

    Sam & Andrew, you naughty fellahs! Are you out helping wet people or not adding a 5:01 update to see how many times we’ll refresh our browsers ? XD
    Those pesky neurologists!

    BTW- Im trying to explain the probabilities up top. I presume it’s not 100% til after the election?

  • zenger

    Gore lost Tenn. while winning the popular vote.

  • Jefflz

    The Italian Fascists had an expression ( the origin of which is disputed) that seems very approriate for the occasion: Many enemies, much honor!

    Long live the Nerds!!

  • Olav Grinde

    The recent movement of the Meta-Margin confirms that Mitt Romney has momentum. Downward momentum.

    • Eric

      Indeed. I wonder what all of those who are screaming “Romney will win by a landslide!” will say one week from now? Oh, what am I saying; they will claim it’s a conspiracy because everyone they had been listening to told them Romney was definitely going to win. What they fail to realize is that they surrounded themselves with those who only thought like they did and who were feeding them disinformation. Instead of doing the proper thing and turning on those who fed them these lies they will instead do the wrong thing and blame the victor for lying, cheating and stealing. It will never occur to them that they were the ones in the wrong all along. I feel sorry for such people but know that there is nothing that can be done to help them because they must first admit they have a fault before that fault can be fixed.

    • skmind

      ‘I wonder what all of those who are screaming “Romney will win by a landslide!” will say one week from now?’

      If Obama wins, they’ll be headed to Wikipedia.

      Trying to change the entry for ‘landslide’ to 230+ electoral votes and winning the most states.

    • Mikey Armstrong

      Eric, they know full well Obama will most likely get a second term.

  • Ohio Voter

    That’s a beautiful MM to close the day

  • Frank

    I admit to visiting your page at least 10 times per day for the past two weeks. Found your link on 538. Your factual analysis brings grounded logic to a sea of media-induced election confusion. I truly appreciate your column.

  • John NYC

    Folks finally figured out that you’re the real deal and Nate basically wishes he was as smart as you.

  • cevangelista413

    I love your site. I check it pretty much every few hours. I was wondering if you could address something I encounter a lot on other sites, where people scream and shout at me about how Romney is ahead by “double digits!!!” and so on. I check a lot of polls, and I don’t see that anywhere. Are there really polls this incorrect, or do you think people are delusional to the point that they make up their own numbers? I’ll take my answer off the air.

  • Michael Worley

    I have a serious question:

    Assume the median poll is accurate. Then the popular vote is between tied and 1 point Romney ahead.

    Then this represents a 7 point shift from 2008.

    In 2008, Obama basically tied in North Carolina.

    The MM has him losing NC by less than a point.

    I find it improbable that Obama maintained NC so well. Is there not a systematic problem in most of the swing states? Ohio is one thing, but North Carolina is another.

    In short, either national polls are wrong (and we’re assuming the median), state polls are wrong, or Obama has accomplished the astonishing feat of maintaining his standing in NC by 7 points above last time, FL by 3 points, OH by 5 points, etc.

    The best national polls for Obama show him up by 1. If this is true, NC has a 6 point swing towards Obama relative to the nation. That’s crazy, even if you think the margin will be huge in red states.

Leave a Comment