Princeton Election Consortium

A first draft of electoral history. Since 2004

A Local Experiment In Crowd Wisdom

July 29th, 2016, 4:00pm by Sam Wang


(updated 10:10pm)
I have finally converted my statistical politics hobby into material gain.

Here at Princeton, the Center for the Study of Democratic Politics has announced the winners of its quadrennial election-prediction contest, a game that is open to CSDP and the Princeton community more broadly. Late in 2015, they put out an open call. Entrants were asked to predict the nominees for the two major parties. The deadline was January 2, 2016. Vice-presidential nominations were used as tie-breakers. I won! The prize: a hooded sweatshirt with the CSDP logo.

Near the end of December, I submitted my entry to the organizer, Michele Epstein along with a note: “If I’m wrong, and there is a good chance I will be, consider this merely a snapshot of what the data pointed toward as of January 1, 2016.” As it turns out, 2016 has been an excellent test of what a polling data-centered approach can accomplish.

Taken as a whole, the contest entries describe the collective opinion of knowledgeable experts – and reveals what a strange year this has been.

There were 69 entries in all. Eight of these entries correctly predicted Hillary Clinton (D) vs. Donald Trump (R). Of the eight correct Clinton v. Trump entries, I won the tiebreaker by saying Tim Kaine would be the Democratic vice-presidential pick. Four picked Julián Castro, and one each picked Joaquín Castro, Martin Heinrich, and Martin O’Malley. Julián Castro was also the betting markets’ favorite at the time.

On the Republican side, the most common incorrect guess was Marco Rubio (40 entries), most often with John Kasich as a running mate (eight entries listing a Rubio/Kasich ticket). The single most common VP choice was Carly Fiorina.

Here are histograms of the answers on the Republican side. Names that were submitted only once are omitted (scroll down to see those):
Recall that these entries were made in late December, before any primaries had occurred. Many of them reflect solid guesses based on the assumption that party mechanisms on both side would work normally, bringing together party officials, donors, and activists – a model that has been informally called “The Party Decides.”

Obviously, that didn’t happen for the Republicans. This fact illustrates that the national Republican Party is, in a deep sense, broken this year. The breakage was hard to recognize, except for the fact that polls and conventional wisdom pointed in such different directions. In the survey, most knowledgeable observers thought Party mechanisms would win out. I chose the polls.

This raises the question of where the Republican Party is headed in coming years. It is by no means certain that they have dark days ahead. For example, if Hillary Clinton and the Democrats come into power this November, it is quite possible that they will lose seats in 2018, since midterm elections are often bad for the party in power. That dynamic could serve many Republicans well, even if their party still suffers from internal divisions.

I did make one wrong call: I incorrectly predicted that Donald Trump’s running mate would be Ted Cruz. As you can see in the VP answers above, this was a common error. I will defend myself by pointing out that Mike Pence covers the movement-conservative bases, making him a mini-Cruz. In some sense, this represents an echo of normal party mechanisms at work.

Below are the summary statistics for the Democratic VP nomination; “Castro” means Julian Castro. Again, single entries are omitted.

Thanks to Michele Epstein and CSDP.

All the entries:

WHO WILL BE THE DEMOCRATIC NOMINEE FOR PRESIDENT?
WHO WILL BE THE DEMOCRATIC NOMINEE FOR VICE-PRESIDENT?
WHO WILL BE THE REPUBLICAN NOMINEE FOR PRESIDENT?
WHO WILL BE THE REPUBLICAN NOMINEE FOR VICE-PRESIDENT?
Bernie Sanders
Cory Booker
Chris Christie
Marco Rubio
Bernie Sanders
Martin O’Malley
Jeb Bush
Marco Rubio
Bernie Sanders
Martin O’Malley
Marco Rubio
Ted Cruz
Hillary Clinton
Martin Heinrich
Chris Christie
Marco Rubio
Hillary Clinton
Robert Gates
Chris Christie
Ted Cruz
Hillary Clinton
Joaquin Castro
Donald Trump
Scott Walker
Hillary Clinton
Julian Castro
Donald Trump
Marco Rubio
Hillary Clinton
Julian Castro
Donald Trump
Marco Rubio
Hillary Clinton
Julian Castro
Donald Trump
Nikki Haley
Hillary Clinton
Julian Castro
Donald Trump
Ted Cruz
Hillary Clinton
Martin Heinrich
Donald Trump
Ben Carson
Hillary Clinton
Martin O’Malley
Donald Trump
Ted Cruz
Hillary Clinton
Tim Kaine
Donald Trump
Ted Cruz
Hillary Clinton
Martin Heinrich
Jeb Bush
Marco Rubio
Hillary Clinton
Tim Kaine
Jeb Bush
John Kasich
Hillary Clinton
Anne-Marie Slaughter
Marco Rubio
Chris Christie
Hillary Clinton
Bernie Sanders
Marco Rubio
Ted Cruz
Hillary Clinton
Deval Patrick
Marco Rubio
Nikki Haley
Hillary Clinton
Mark Warner
Marco Rubio
Kelly Ayotte
Hillary Clinton
Julian Castro
Marco Rubio
Carly Fiorina
Hillary Clinton
Julian Castro
Marco Rubio
Carly Fiorina
Hillary Clinton
Julian Castro
Marco Rubio
Jeb Bush
Hillary Clinton
Julian Castro
Marco Rubio
John Kasich
Hillary Clinton
Julian Castro
Marco Rubio
John Kasich
Hillary Clinton
Julian Castro
Marco Rubio
Joni Ernst
Hillary Clinton
Julian Castro
Marco Rubio
Kasich
Hillary Clinton
Julian Castro
Marco Rubio
Kelly Ayotte
Hillary Clinton
Julian Castro
Marco Rubio
Mia Love
Hillary Clinton
Julian Castro
Marco Rubio
Nikki Haley
Hillary Clinton
Julian Castro
Marco Rubio
Nikki Haley
Hillary Clinton
Julian Castro
Marco Rubio
Nikki Haley
Hillary Clinton
Julian Castro
Marco Rubio
Nikki Haley
Hillary Clinton
Julian Castro
Marco Rubio
Scott Walker
Hillary Clinton
Julian Castro
Marco Rubio
Scott Walker
Hillary Clinton
Julian Castro
Marco Rubio
Ted Cruz
Hillary Clinton
Julián Castro
Marco Rubio
Nikki Haley
Hillary Clinton
Mark Warner
Marco Rubio
Carly Fiorina
Hillary Clinton
Mark Warner
Marco Rubio
Chris Christie
Hillary Clinton
Mark Warner
Marco Rubio
John Kasich
Hillary Clinton
Martin Heinrich
Marco Rubio
Mike Pence
Hillary Clinton
Martin O’Malley
Marco Rubio
Carly Fiorina
Hillary Clinton
Martin O’Malley
Marco Rubio
Carly Fiorina
Hillary Clinton
Martin O’Malley
Marco Rubio
Carly Fiorina
Hillary Clinton
Martin O’Malley
Marco Rubio
John Kasich
Hillary Clinton
Martin O’Malley
Marco Rubio
John Kasich
Hillary Clinton
Martin O’Malley
Marco Rubio
Kelly Ayotte
Hillary Clinton
Sherrod Brown
Marco Rubio
John Kasich
Hillary Clinton
Tim Kaine
Marco Rubio
Carly Fiorina
Hillary Clinton
Tim Kaine
Marco Rubio
Carly Fiorina
Hillary Clinton
Tim Kaine
Marco Rubio
Carly Fiorina
Hillary Clinton
Tim Kaine
Marco Rubio
John Kasich
Hillary Clinton
Tim Kaine
Marco Rubio
John Kasich
Hillary Clinton
Tim Kaine
Marco Rubio
Lisa Murkowski
Hillary Clinton
Wesley Clark
Marco Rubio
David Petraeus
Hillary Clinton
John Hickenlooper
Ted Cruz
Condeleeza Rice
Hillary Clinton
Julian Castro
Ted Cruz
Ben Carson
Hillary Clinton
Julian Castro
Ted Cruz
Carly Fiorina
Hillary Clinton
Julian Castro
Ted Cruz
Carly Fiorina
Hillary Clinton
Julian Castro
Ted Cruz
Carly Fiorina
Hillary Clinton
Julian Castro
Ted Cruz
Jeb Bush
Hillary Clinton
Julian Castro
Ted Cruz
John Kasich
Hillary Clinton
Julian Castro
Ted Cruz
John Kasich
Hillary Clinton
Julian Castro
Ted Cruz
Nikki Haley
Hillary Clinton
Julian Castro
Ted Cruz
Nikki Haley
Hillary Clinton
Julian Castro
Ted Cruz
Susana Martinez
Hillary Clinton
Julian Castro
Ted Cruz
Susana Martinez
Hillary Clinton
Julián Castro
Ted Cruz
Nikki Haley
Hillary Clinton
Martin O’Malley
Ted Cruz
Marco Rubio
Hillary Clinton
Wes Clark
Ted Cruz
Carly Fiorina

 

Tags: 2016 Election · President

39 Comments so far ↓

  • Nick Carnes

    This isn’t really your first CSDP win, though.

    In 2008 — after hearing you give a talk in the Politics Department — I submitted an entry in that year’s CSDP contest in which I answered every question I could using the latest estimates from your website. “My” entry won the CSDP contest in 2008 (Michele can verify).

    When I emailed about the win, you very kindly declined my offer to hand over the 2008 hoodie. I’m so glad to hear that you got a hoodie of your own — and the recognition you deserve — this time around.

    Congratulations on your win, and thanks maintaining such a terrific website.

  • Matt McIrvin

    Kaine was actually a far more popular guess for Democratic VP nominee than I would have expected: as many people picked him as Trump for R nominee, though only Sam named both. He wasn’t even on my radar until the rumors started to go around in the last days before the convention, though as soon as I heard the name I thought he wasn’t a bad choice.

    Nobody guessed Labor Secretary Tom Perez, who emerged as one of the progressive favorites in the final days.

  • Ali

    Love your site! Is there any way you can add a graph that shows the progression of the Bayesian probability? Or does it so closely track the meta-margin that it’s not worth doing that?

    • Matt McIrvin

      So far it’s hardly moved at all. It would be a very boring graph.

    • Kevin

      I believe Sam intentionally dampens the noise on the Bayesian probability estimate by rounding it to the nearest 5%, on the grounds that this noise is not instructive in light of the uncertainty. I do worry about false impression of sudden movement created if the prediction suddenly ticks up or down 5%.

  • Paul

    Sam,
    Yesterday in Bloomberg News, there was an article about polling results in which the question was: ‘who do you think will win?’ as opposed to ‘who will you vote for? With the point being that asking the first question is more informative re the actual outcome than the latter. Logically, this makes sense b/c the first question is asking each person questioned to do a mini poll in their head based upon everything and everyone they know rather than the personal voting question which is simply asking for 1 data point without reflection. On a personal level, I know that I do this, in particular, I focus my attention on people who are not stridently on either left or right. Do you have any thoughs/insights? Spoiler – Clinton is way ahead.
    Paul

  • Olav Grinde

    We demand a photo of the winner in his new CSDP hoodie!

  • Mark J

    Congrats on your correct pick.

    I have noticed in reading your web site and the comments a strong Democratic tilt. I suspect this same tilt in most punditry and media is creating a false signal in this election. There is a very strong anti-Trump bias out there.

    I realize every election in my 60-year lifetime has experienced this same lean and that pollsters and analysts have tools for correcting for bias, but Donald Trump seems different than anyone in the past. I suspect he is actually in better shape than his numbers show. It’s the Bradley effect and I saw it in the Brexit vote as well.

    I’d be quite interested to know if you have concerns about this.

    • Sam Wang

      No concerns, especially since there was no Bradley effect in the GOP primaries. Brexit is inadmissible, since it was a one-off election.

    • Commentor

      I seriously doubt the suggestion that there is a strong Democratic tilt in “most punditry and media.” Its certainly a GOP talking point, but several studies have been released suggesting that Clinton has received much more negative coverage than Trump has. And where do conservative media outlets such as Fox News, the Wall Street Journal, and the NY Post fit into this narrative?

      Frankly, this post seems to echo the rationale behind the “de-skewing” movement from the last election.

  • MarkS

    Congrats! But what was the material gain?? At least lunch at Prospect House, I hope!

  • Amitabh Lath

    Anne-Marie Slaughter! I have to admit she was not even on my way-outside list of possible VPs. Who sent in that pick? Was it someone from the New America Foundation?

    As for the top of the fight card, Clinton and Trump were pretty much consensus picks by late Dec/early Jan, at least here on the PEC. It would be interesting to poll these other experts and find out why they thought Trump would falter? Did they not trust the polling? Or suspect a Herman Cain style collapse in the spring? Or expect the rest of the republican field to collapse around a consensus anti-Trump choice sooner than they did? (well, actually they never did)

    • 538 Refugee

      My guess is a health dose of “The Party Decides”. Let’s face it. It’s really hard for a lot of us to face up to the fact such overt hatred was just waiting for a champion to surface.

      Rosalynn Carter says about Ronald Reagan, “I think this president makes us comfortable with our prejudices.”

    • Amitabh Lath

      A large number of clear-eyed professional political experts did not consider Trump a serious choice. That means his nomination may not be a robust phenomenon. In alternate universes there may be more cases of Trump imploding or being taken down by a coordinated establishment attack than the reality we currently inhabit. A lot of low-probability events had to all line up. For instance note how Jeb was considered so strong early that Romney stayed out, but so weak later than he dropped out before FL.

      Or of course, the experts use the wrong metrics (or worse, their guts) and Trump rules the GOP across the multiverse. Either way, it would be good to diagnose this.

  • pechmerle

    The most optimistic of the forecasting sites currently: Election Analytics, a project of the Univ. of Illinois Computer Science Department:

    http://electionanalytics.cs.illinois.edu/

    Their forecast as of today: Clinton 100% to win [!]

    They got the result right in 2012, but were off on the EV totals. And that year one of their c0-leaders of the project, in his final before-the-election call, said Romney. Why: “basing that in large part on Rasmussen Reports polling data, which he said he considered ‘the most reliable.’” [!!] WSJ 11/7/12 follow-up on the forecasters.

    • Tony

      I hope it’s true, but I’m not sure I’d believe a group that says Clinton’s likely to win Kentucky and North Carolina is safer for Clinton than Washington.

    • 538 Refugee

      “Each forecast provides a snapshot of the current state of the election and represents the outcome if the election was held on that day.” Well, there is 0% chance of them being proven wrong since the election isn’t held “on that day”. Combine that with little current state polling and roll your eyes.

      Expected Electoral Votes 328.78 209.22
      High precision for an impossible number. I’d suggest a drinking game based on misuse of significant digits but alcohol poisoning would ensue.

    • UI Supporter

      New feature added: How will undecided voters impact the Presidential election? Visit “Customize your forecast” at http://electionanalytics.cs.illinois.edu/
      You may not like what they say, but you cannot dispute their record in 2008, 2012, and 2014 (senate).

  • BillSct

    9:13 pm Friday July 29th PEC just got mentioned by Rachel Madow as a “high tech” election tracking organization along with a bunch of other more mainstream election tracking sites.

    • Sam Wang

      That’s nice! Though I have a somewhat contrarian view: I have noticed a tendency for that program to mention PEC whenever we make an extreme prediction favoring the left: Democrats retaking the House in 2012 (a miss), or retaining the Senate in 2014 (that’s what I said in September, though not in October; this is now remembered as a miss). I think I understand the reasons for those misses now (2012, population structure and gerrymandering; 2014, midterms aren’t random in their drift). I think the November prediction for 2016 is built on more solid ground…the coming weeks will tell.

  • anonymous

    Eight people got both presidential nominees right in January, which suggests hefty per entrant wisdom. But nobody got all four correct, which is probably due to the scarcity of political junkies (small crowd = less total wisdom).

    • Jinchi

      To be fair, Trump wasn’t even sure of his VP pick on the day before he announced him, so it’s asking a lot for anyone else to guess it correctly.

    • Matt McIrvin

      All four nominees were individually guessed by somebody, though. The one person who guessed Pence thought he would be Marco Rubio’s running mate.

  • Truthy

    Careful Dr. Wang, today a hoodie, tomorrow a complete site redesign and articles about “How Trump’s momentum makes the race just too close to call!”

  • pechmerle

    Is there a list of the non-Princeton entrants in the competition somewhere? Interested in seeing who the persons/teams were.

  • Canadian Visitor

    Dr. Wang, you’re the best. Congratulations. Internauts from the True North see PEC as a quick and reliable way of making sense of US politics. More wind to your sails.

  • David Fry

    So no one picked Elizabeth Warren as Hillary’s running mate? I would have thought that would be a common guess, although in December she was tacitly aligned with Bernie.

    • Matt McIrvin

      Only one guess for a Clinton-Sanders unity ticket, too–fewer people than guessed Sanders would get the nomination. I suspect these choices were always considered less plausible by political junkies than by the masses.

  • 538 Refugee

    Maybe ‘someone’ can link to this in their mea culpa on missing the Trump nomination early and often. Yes, my handle is my hint. ;)

  • Chip

    Congratulations. Well deserved. I’d have guessed Clinton/Biden and flipped a coin on the (R).

  • Joseph

    Sam, are you sure you didn’t call up Hillary and say you’d split the winnings if she went with Tim? /s

    Congrats!

  • W.Dow Rieder

    Cool, and congrats. But someone should talk to them about using something better than pseudo-3D bar graphs to present results. Tufte would be shaking his head.

  • Alan Cobo-Lewis

    Impressive–but your Kaine prediction couldn’t have been based on data, could it?

  • Matt McIrvin

    So how did you pick Tim Kaine? Data, or gut?

  • A New Jersey Farmer

    Mazel tov to you Sam. Proud to be a PEC fanatic.