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

Any requests?

November 2nd, 2008, 10:30pm by Sam Wang

I’m taking requests for analysis to do between now and Tuesday morning.

Possibilities include the following:

A cheat sheet to get you through Election Night. Presidential, Senate, and House races to watch, and early warning signs of what the evening will yield.

Analysis of inefficiencies and errors in electronic markets. I would also name contracts that are under- or over-valued. (Even if I post, if you send me proof that you have my book or have donated to one of my favorite races, I’ll send you recommendations by late Monday.)

Final Presidential, Senate, and House predictions. Actually, you’ll get this one no matter what. I’ll post them late Monday night, when I enter them into various contests.

A chart for applying biases and “corrections” to polls. This is in case you believe in the Bradley effect or want to assign undecided voters, and want to see how it affects the prediction.

More snark about other aggregators. I’m sure many of you want this one, right? (crickets)

Is there anything else?

Tags: 2008 Election

62 Comments so far ↓

  • Hans

    Dr. Wang: Comment #3, above, asks for a break down of size of shift in meta-margin by frequence of that size shift. I think the file EV-estimate_history.csv contains the data for this, and was about to try it for my own curiosity, except…there are several days that list a meta-margin of -999. What does that mean?

  • Nicholas J. Alcock

    Dear Sam,
    You have stated you will using a t-distribution not a n distribution in 2012.
    Could you please recalculate your 2004 and 2008 calculations in the light of the proposed change in methodology? My guess is that your EV estimations would be unchanged?

  • MN Gumbo Girl

    Please calm me down. RCP dropped Ohio and Virginia into the toss-up category. So now the RCP tally is at 278, when recently (yesterday, I think) it was at 311. That’s a big, sudden drop and it’s freaking me out a bit. Tell me why that shouldn’t worry me. (I know you’ve got an explanation/perspective.)

  • Bob

    An analysis of what early voters have looked like. There has been a lot of discussion that more minorities, young etc would turn out this election. Based on early voting turnout so far, compared to early voting in previous elections, are we seeing a significant change in the proportion of minority and young voters? Has turnout actually increased compared to previous elections?

  • Josie

    wish list:
    Cheat sheet (POTUS and Senate) with recommendations on who has best TV coverage. Should I watch FOX so I can see exactly where and when the boundless GOP optimism ends, or PBS so my coverage is uninterrupted?

    Cage match with Nate Silver, but perhaps Jesse Ventura would be a better ref, considering his political background.

  • matt

    I want #1 too! Cheat sheet — what to watch for??

  • Sam Wang

    Everyone – thanks for your suggestions. The winner by a landslide is a cheat sheet. This will include a bias analysis, which covers the effects of late swings, unexpected shifts, and undecideds. It will take the form of a worksheet where you can find out the answer in real time. It will also have many other items you have requested, such as Congressional predictions.

    In addition, specific replies:

    JeffH (#2), Paul (#11), and others – An analysis of overall, systematic bias in outcomes will be in the cheat sheet as a graph you can read.

    DFS (#7) – In 2004, poll averages called the winner correctly in every state except Wisconsin (Kerry +1%). Except for this one case, any fraud that might have occurred did not create any outcome that contradicted polls. I view poll aggregation as a key line of defense against fraud.

    GAW (#8) – Will come later. First I will email those who sent me receipts.

    George (#12) – I don’t know. There’s a site someplace that tracks poll closing times. I’ll be watching New Hampshire to see if its results match polls. It’s currently at Obama +11%. Virginia is another one to watch. Voting in both states ends at 7:00PM Eastern.

    Arnold Evans (#13) – The power of your vote is given in the right sidebar. It’s exactly what you want.

    Eddie (#19) – My subjective estimate of the probability of a McCain victory is the probability of a massive extraneous event. Given the frequency in our nation’s history of things like Pearl Harbor and the 9/11 attack, I put it about 0.01%

    Behnam (#38) – Today at, Andy Tanenbaum pointed out a great report from Ruy Texeira on the demographics of those states: % whites, % college graduates, that kind of thing.

    David U (#44) – Contingency analyses are easy. Unfortunately we haven’t automated that, or else you could do it yourself. Maybe in 2012. The cheat sheet will contain some “canary in the coal mine” scenarios to tell you which states to watch most closely.

    Bill N (#46) – The 95% CI is calculated from the probability distribution of all 2.3 quadrillion outcomes. In that sense it takes advantage of all interpoll error, state by state, assuming that sampling contributes some minimum error. It’s just what a guy who knows how to use the word “bootstrap” would want to see.

    Hans (#51) – A Meta-Margin of -999 means we had a problem retroactively doing the calculation because more polls are available now than on the day we calculated the Median EV Estimator. We’re going to fix this after the election.

    MN Gumbo Girl (#53) – I have no idea what RCP’s definitions are, but they must not be limited to math. Obama leads in 12 our of 13 Ohio polls, and in 9 out of 9 Virginia polls.

    Thanks for all your suggestions. If you have more to say, post it here – but also email me at sswang a.t princeton dot edu.

  • Max

    This may not be as much an election-night item, but I would love to hear about the impact that an early Obama victory when the east coast polls close could have on downticket races and ballot initiatives in the West due to lower turnout.

    When this has happened in the past, who stays home? Has it favored the winning presidential party’s other candidates, or his opponent’s? I assume the effect would be mirrored in initiatives that break down more or less along party lines (e.g. Prop 8 in CA) but if there’s any difference there I’d be interested to hear about it.

    If this already exists elsewhere on the web, no need to reinvent the wheel — just point to it.

    Thanks again for a great site!

    P.S. I’m all for the cage match too…

  • Observer

    Sam, can the cheat sheet include presence of third-party votes and probable impact of that on the two major-party candidates in that state?

    Thanks for all that you are doing — and have done.

  • Walter

    It’s just what a guy who knows how to use the word “bootstrap” would want to see.

    This made me chuckle, as a grad student who uses such methods all the time. :)

  • Arnold Evans

    I was unclear – I was asking for a comparison of different ways to volunteer or contribute to a campaign.

    How many calls to Pennsylvania have the same impact as one door knock in Pennsylvania. How many dollars contributed have the same impact as one vote.

    I’m not asking for a comparison between Pennsylvania and another state, but for various activities inside the same state, for example Penn.

    Then I asked if the effects on this election of factors such as the economic crisis or unemployment can be separated from other factors such as Obama’s greater access to money than Kerry, for example.

    I doubt the data necessary to measure these things is available, but thought I’d ask.

  • Howard

    Is it too late to request a polling breakdown along generational lines? Baby Boomers in particular, but also Gens X and Y?

Leave a Comment