Princeton Election Consortium

A first draft of electoral history. Since 2004

Make your own personal prediction!

November 4th, 2008, 12:30am by Sam Wang


(A prediction from Bob Vanderbei. Enter yours in comments!)

Based on the simplest possible analysis in which I average the most recent month’s polls assembled by RealClearPolitics and use election data from four years ago for those states that have had no polls in the last month, I predict Obama will win 364 electoral votes and McCain will get only 174. A graphic showing which states will go to whom according to me can be seen at http://www.princeton.edu/~rvdb/JAVA/election2008/.

–Bob

Tags: 2008 Election

41 Comments so far ↓

  • Observer

    I’m going out on a limb – and up late this – :-) – and going with Obama 379 EV’s:

    He squeaks in in Georgia and Missouri.
    But 353 will do, really it will.

    (I also have an entry in the Princeton Survey study, made weeks ago. We’ll see how that does.)

  • Katman

    My predictions:

    Obama 397
    McCain 141

    Obama takes NC, Indiana, Georgia, ND, Montana, and the Omaha electoral district (NE-02).

    Senate:
    Dems 56
    Reps 41
    Ind 2
    N/A 1 (Georgia runoff)

    House:
    Dems 263
    Reps 172

  • Frank

    Obama 353: Kerry + NC, VA, FL, CO, NV, OH, NM, IA.

  • Magic Dog

    On October 8th, I sent a long e-mail to a bunch of friends predicting Obama would would win 53-54% of the popular vote, and get 355-375EVs. I predicted the Democrats would have 59 Senate seats including Lieberman and Sanders, and would gain 25 in the House, “but I won’t be surprised if it’s higher.”

    On Oct. 29th, I made my “final prediction” of 53% and 375EVs for Obama, 61 Senate seats, and +34 in the House. I might have been a bit too optimistic about the Senate and the House. We’ll see.

    Going back further:

    On June 20th, I had e-mailed the same people, noting the strong historical correlation between presidential election results and the change in the national unemployment rate in the second quarter of an election year.

    Here is what I wrote:

    “Unless there’s some freak statistical artifact in June, I think this month’s economic numbers are going to be a disaster. It’s no wonder that my favorite electoral prediction sites, be they Democratic, Republican, or neutral, are all showing Obama with 300 to 350 electoral votes. States you never would have dreamed being competitive for Democrats are turning. Get this: McCain leads Obama by one ONE POINT in Georgia, and the latest Newsweek poll has him up by 15 points nationwide.”

    On July 4, after the June unemployment number came out, I wrote this:

    “The June unemployment statistics were released yesterday. The national unemployment rate stayed at 5.5%, which means unemployment started 2Q at 5.1% and ended 2Q at 5.5%. By historical standards, that’s a big increase and portends a significant victory for Barack Obama this fall.

    “The Republicans know how bad it is, too. At the moment, I’d say they’re set to lose everything Bush lost to Kerry in 2004, plus: OH, IA, CO, NM, NH and VA. If he does that, it’s 306-232 electoral votes. Obama has reasonable chances in IN, MO, NV, and FL. If he does those, it’s 360-178.

    “If it’s really big, he’ll win some or all of the following: NC, MT, ND, AK, and GA. That would be 399-139. Obama is getting ready for a big victory. Maybe that’s why he’s thinking of giving his nomination speech in a 75,000-seat stadium rather than a mere auditorium.

    “The Republicans know they’re in big trouble, which is why they’re flailing around so desperately trying to find angles of attack. I think it’s a matter of damage control. The way it looks now, the Republicans will lose at least 20 House seats and at least 5 Senate seats, in addition to the presidency. How much higher their losses go will be a matter of how well Obama runs his campaign.”

  • Sam Wang

    Everyone – So far, counting Bob’s original post, we have 25 entries, ranging from 281 to 453 EV (the second-highest is 407 EV).

    Calculating my very most favorite October thing in the world to calculate, that makes a median of Obama 364 +/-9 EV. Only a little above the Median EV Estimator, and equal to my Obama +1% prediction and to Bob’s prediction.

    This is interesting. I thought commenters around here had become a bunch of godless, Muslim, Rev. Wright-following, Ayers-hugging, toy-sharing, anti-mavericky non-mavericks. Yet amazingly, that all adds up to only +1% bias.

  • The Liberal Crab

    It’s going to be A LOT closer then people think. My guess is 302/236. Virginia, Colorado, Nevada, New Mexico, Iowa and Missouri.

    Although I keep wavering on Missouri.

    I think the ‘coal’ comments from Palin yesterday is going to affect that super tight race (see, this is the non-quantitative analysis that needs to be included beyond just the numbers). Florida is simply a guess that it will go to McCain – he’s been closing there and that’s a tough state for the dems.

  • blair alef

    Bruce & Steko. If there’s a polling bias that understated Obama’s strength, keep an eye on Louisiana. I have wierd hunch.

  • Bruce W

    Thanks, Steko for being a bigger outlier than me. But AK and TX? — that would be spectacularly funny.

    All reports are of a huge turnout and a lot of enthusiasm, which is good for O-Bam and bad for J Mac.

  • Steko

    My original prediction was 390 Obama but that got too popular so I diluted my vodka with some kool aid and came up with 453 (all swing, AZ, WV, AK, TX, Omaha and even Lincoln).

  • David in NM

    I’m in the 5% bias club. Using Dr. Wang’s polling bias chart, I’m guessing as follows:

    I expect McCain to win OH, FL, IN, MO, NC, ND, GA and AZ

    Obama will win PA, CO, NM and IA and hold NH (by less than the polls show)

    No call on VA or NV

    Obama at 273-291 EVs and by approximately 2% overall.

    Of course, the optimistic part of me REALLY hopes I’m HORRIBLY wrong and its an Obama landslide-I would SO enjoy that.

  • gprimos1

    Dr Wang,

    On the bias nomograph, does each red line represent 1/2 sigma from the black line expected value?

  • MB

    That One – 318
    Double Maverick – 220

    Obama wins VA, FL, PA, NM, CO, and NV but loses OH.

  • blair alef

    Hey Sam,

    Google Maps has real time, customizable, county by county election tallies coming tonight starting at 6:00 .
    http://maps.google.com/help/maps/elections/#2008_election

    Also, http://politicalwire.com/ has a great, concise rundown of what to watch for tonight

  • Greg

    I’d love to be wrong on the magnitude of victory, but who am I kidding? I simply want Obama-Biden to come out on top. In terms of prediction, this is more of a guess than anything Sam or other sites have posted. But I’m going with the following:

    Senate: +8 (sort of, see below) seats for Democrats, based on these outcomes: Hagan, Merkley, Shaheen, Begich, Udall, Udall, and Warner…however, I disagree with the counting of Lieberman as “independent,” and I now see him as a lost cause just like Zell Miller became. He’s on the (R) side of the aisle in my book. I also think Chambliss wins a runoff with a ton of Republican money backing him in a GOTV effort in Georgia. Final tally: D = 58, R = 41, I = 1

    House: I think there’s far too much for me to research, so I’ll play with house money and say +26 for the Democrats, giving (I believe) D = 259, R = 176.

    President: Obama/Biden 337, McCain/Palin 201
    I have no real basis for this other than having practically obsessed about this thing all year long, not to mention a little canvassing in limited free time (I don’t know how some of these volunteers are doing it!). I have IN, OH, VA, NC, NH, CO all going to Obama/Biden. I have MO, FL, ND going to McCain/Palin. Though this is about as scientific as if I were to bake a cake blind, it’s still fun to take a stand with some numbers.

    Thanks for the great coverage, Sam, and I look forward to getting the book as a much-hinted-for present this year!

  • Joseph Marshall

    I’m going to stick as close as possible to your meta-margin totals and call Obama 352-353. Having looked at the last few days of the Gallup tracking poll–where both “traditional” LV and “expanded” LV models are showing virtually identical scores, I think the amount of “hidden” Obama support is being overestimated.

    So I’m sticking with a proven approach, and I also want to say thank you for your time and effort to develop and present it.

  • Bruce W

    For most of the reasons cited by acorvid, above, I’m going to be an outlier (which, in another thread we’ve equated with “deeply crazy”).

    Obama 393, McCain 145;
    Popular vote +10 to +12 Obama

    My map is pretty much Sam’s map with the three toss-ups (ND, MO, ID) going to Obama, plus NC and an upset in GA.

  • Andrew Houck

    I agree with all the states in the common D364 model, but I think GA is going to flip as well, giving 379-159. Obama by 9 in the popular vote.

    For the Senate, 57-41-2, with Martin defeating Chambliss (now or in a runoff) but Franken falling to Coleman.

  • William Ockham

    Predicting the winner today is not that hard. What did you think 4-6 weeks ago?

    Obama 376, McCain 162
    Obama carrying Kerry states plus Florida, Ohio, North Carolina, Virginia, Indiana, Missouri, Colorado, New Mexico, Nevada and 1 EV in Nebraska.

    I made this prediction on October 1. I’ll stick with it. (No link to prove it, but I created my electoral college map as a pdf file with date/timestamp of 3:26pm 10/1/2008).

    For the popular vote, I said this on Sept. 18:

    Likely outcome: Obama 54%, McCain 43%, Others 3%.

    http://www.fivethirtyeight.com/2008/09/state-of-race-is-mccain-in-trouble.html#comment-4098428411888103915

  • Kevin

    Obama: 352
    McCain: 186

    Toss-ups go like this
    Obama wins North Dakota, Ohio, Virginia, Pennsylvania, Florida, Neveda, Missouri, Colorado, New Hampshire and New Mexico.

    McCain wins North Carolina, Indiana, Montana, Arizona and Georgia.

    California, Vote NO on Prop 8.
    Florida, Vote NO on Prop 2.
    Arizona, Vote NO on Prop 102.

  • Mark S.

    Obama 378 – I go with the current “Obama +2%” map.

    That gives Obama the “genuine” deciding states (CO, VA, OH, FL), the “looks-like-a-landslide” states (NC, IN, MO) and one “no way!” state (ND). But not MT, GA or AZ. Popular vote 53.5%-45.5%.

    Maybe it’s my heart overruling my head, but, I’ll go with the black+youth+can’t-take-any-more vote being larger than estimated. Plus I’d like to see the poll consensus be off by a bit. (Of course, they could be wrong in the other direction…)

    The Senate: Dems gain 8 (go Franken!) with Georgia in a runoff, but will go on to lose without Obama coattails.

  • blair alef

    Hey Brian, I think the biggest surprise might just be Louisiana but I’m not counting on in my my predictions.

  • JKD

    Obama 338, winning FL, VA, PA, & OH
    McCain 200, winning GA, NC, MO, IN, & ND

  • blair alef

    First, the PEC stateprobs.csv file is absolutly great. I reformatted the data as a scorecard for tonight and every one of my insecure friends breathed a sigh of relief.

    Second, I think the composit of voter turnout will be somewhat different than expected. I give all states with McCain leading by less than four points to Obama. This includes Montana & Georgia with the true toss up states but leaves out Arizona. 387 to 147.

  • Brian

    Obama 368, McCain 170
    “Battleground” states: PA, OH, VA, NC, FL, NM, CO, NV, NH, GA to Obama; AZ MT, ND, MO, and IN to McCain

    Obama with 7.5% popular vote margin

    Biggest surprise Georgia to Obama
    Last state called – Missouri for McCain by slimmest of margins

    Senate 57 D plus my Senator, Bernie Sanders, and Lieberman
    House 265 D

  • Evans

    Said it for days, 390+ EV for Obama due to a superior ground game, which will be the story tomorrow. Senate breaks 59-41 with a surprisingly close race in GA, where the AA vote has been drastically underestimated.

  • Dave

    I was predicting an Obama landslide — until I woke at 3 AM from a dream where he won by 2.4 percent. So that’s my unconscious’s final word, and I’m going with it.

  • RICHARD LANGFORD

    Barack will win all the states tied or leaning his direction. that is a total of around 407 EV

  • Paul Luther

    I’m going with Rove, too. And because I think that Obama has EV’s to spare, I don’t mind so much that MO goes red. I sick of all this ‘bellwhether’ talk.
    Missouri, get with the program!

  • Walter

    My prediction is that one of my students will give me a TV so I can watch The Daily Show tonight.

  • J.W. Hamner

    I’m with the Obama 364-173 crowd. While MO and NC appear to be 49-49 coinflips right now, I think the cell phone effect will get them both for Obama.

  • acorvid

    Obama ground game + Obama enthusiasm + reverse Bradley effect + Repuglican dis-spiritedness = Obama by 10-12% w/396 EVs,
    + 10 senate pick-ups.

    Dreaming? I don’t think so. Check this space on 11-5.

  • Keith

    obama 281
    mccain 257

    McCain takes IN, NC, FL, MO, PA.
    Ohio puts Obama over the top.

    Senate Dems plus 6
    House Dems plus 22

  • Michael Slavitch

    Bah. That’s what I deserve for putting in special characters.

  • Michael Slavitch

    Distributions with αreferred to< as being heavy tailed.

  • ThatSeattleGuy

    Here’s the prediction I put on the blog at 3BlueDudes.com a few days ago – repeating it here unchanged:

    EV: Obama 361, McCain 177
    PV: Obama 53.4, McCain 45.7, Others .9

    Sen: D+7 (D56, R41, I2)
    House: D+26 (D362, R173)

    /TSG/

  • Michael Slavitch

    I predict 375-163 unless there is a massive power-law anomaly in turnout distribution.

    If that is the case there is no mean and all bets are off.

    Distributions with α<2 as being heavy tailed, as they have infinite variance. Distributions with α<1 do not even have a mean, rendering much of undergraduate level statistical analysis (i.e. mean ± 3 standard deviations type results) inapplicable.

  • Michael

    I don’t predict outcomes. Too superstitious. I will however, predict that all the predictions of record turnout will come true. I voted at about 8 AM and over 600 people had already cast their ballots at my precinct, which had a total of just over 1800 votes for president in 2004.

  • Byron

    In the first, and only, instance of finding any common ground with Karl Rove–I will delightedly forecast 338 EV.

  • Watershed

    My first attempt at an election outcome prediction:

    Obama: 368
    McCain: 170

    “Toss up” state predictions:

    Obama will win Virginia, North Carolina, Georgia, ARIZONA, Ohio and Florida.

    McCain will win Indiana

  • Robert Vanderbei

    Hey. Is it Bruce Reznick from UIUC?

  • Bruce Reznick

    Hi, Bob. I’d be happy with the outcome you propose!