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

Election returns thread

November 4th, 2008, 7:30pm by Sam Wang

Here’s a thread to discuss returns, topics from the Geek’s Guide, and more.

7:59 pm: Watching returns on ABC. TV is an amazingly bad way to get returns. Go to or (In the end CNN is quite good.)

It’s hard to rely on most states’ returns because partial returns are often so misleading. I’d watch New Hampshire (CNN) (where Obama should win by 11%, and closed an hour ago. It’s a homogeneous, small state, so there won’t be late surprises. If his margin’s between 9 and 13%, that’s in line with my projections. Less than 9% – nailbiter (<310 EV). More than 13% – massive, red-ceiling-smashing blowout (>370 EV).

I’m personally interested in Indiana because it’s on a knife edge. It doesn’t matter which way it falls – but they closed 1-2 hours ago and I am curious.

People are asking me what to look for. I have no idea since I am not in much doubt. I’d say I will be convinced as soon as I see NH. Doubters will want to wait until PA, maybe 8:30 pm or so?

8:30 pm: Okay, not yet. All those states being called – SC, etc.? Fuhgeddaboutit. Not of interest. That’s what exit polls are for – to give media organizations the ability to call blowouts early in the evening. Look for surprises, of which there haven’t been any yet.

9:15 pm: Returns are slow, but no discrepancies from my projection yet. President: A number of routine states have been called. McCain is leading in GA, IN. Obama’s up by 12% in NH but with only 20% reporting. Senate: Wow, Chambliss might clear 50% in Georgia. Not surprising, but there was some hope. Democrats’ ceiling is now 59 votes (counting the independents). I am on eggshells over Minnesota. Come on, Minnesotans, start counting!

9:19 pm: What the hell is happening in Virginia?

9:53 pm: Ohio for Obama. It’s done. There’s no path for McCain. He could win NV, CO, NM, VA, NC, FL – and Obama would still be at 279 EV. But that won’t happen. Congratulations, President-elect Obama!

10:00 pm: OK, meta-analysis victory dance over. I’m with Paul Starr, editor of the American Prospect. He has word from the Obama campaign that they are optimistic about Virginia and North Carolina, and think they have a good shot in Indiana. Things are looking good for Obama. Speaking as a geek, I’m still on eggshells – I want to know how the knife-edge states will fall.

10:20 pm: Obama’s leading by a whisker in VA (86% reporting) and NC (75% reporting). McCain’s ahead in IN (86% reporting), MO (21% reporting, so unclear still), and ND (33% reporting, but a 16-point lead there). If these trends all hold up, then it looks like I called all the states correctly. However, I wasn’t expecting VA to be so close. I’ll tell you about the secret tiebreaking sauce tomorrow.

Oh, and by the way: Bradley effect? Turnout? Cell phones? Anti-Bradley effect? Overall, their net effect might be…zero.

11:23 pm: Except maybe not North Carolina? 11:38 pm Wait…

4:44 am: Well, that was all very exciting. Overall, a great night for Barack Obama – and for the Meta-Analysis. Thanks, everyone – some wrapup in the morning.

Tags: 2008 Election

68 Comments so far ↓

  • Matt McIrvin

    Obama might get Indiana too, though. It’s amazingly close.

  • Forrest

    georgia senate isnt over.. we have got to see how those early votes break… runoff still possible, and could be for the 60th seat after all is said and done.

  • Mike in Seattle

    Kudos Sam. You nailed it. What an amazing evening! Apparently the BBC is reporting that Washington’s 11 EVs put Obama over the top. That feels so good. Neurobiology of decision-making? Indecision? Well, my orbitofrontal cortex is on fire and my dopamine neurons are doing a victory lap! High 31 you geeky pundit you.

  • gprimos1

    Wow landslide victory for… the model! EV count right up there with the median estimator. So much for the unneeded extra variance provided by the more complicated prediction models.

  • omar

    finally went through my bias nomograph. Based on the nytimes numbers at this moment, Georgia, Arizona, and North Dakota all went for McCain by 5 to 8% more than predicted (all with > 90% reporting), and Pennsylvania went to Obama by 4% more than predicted (98% reporting). Otherwise, everything on the plot is within 3%. (not counting Montana…too early).

  • Eddie

    Good job– right in your bounds for sure. Us skeptics will have a much tougher time being believable in 2012!

    Go Mr. President-Elect!
    I’m so happy!

  • Eric T

    So until the results are complete, we can’t say for sure, but based on where we are now, I think I see an interesting pattern…

    It looks to me like the actual results tend to be more polarized than the polls were — maybe because undecided people tend to follow their friends and neighbors on election day?

    Just counting up states where the current count is more than 2% off from the polls, I get 10 red states getting redder (than the polls) vs 2 getting bluer, and 15 blue states getting bluer vs 4 getting redder.

    Any thoughts?

  • Hoosier

    Indiana is going BLUE! 364 would match Sam’s +1 Obama scenario….

  • Observer

    How important is GOTV:
    CNN is showing Obama ahead in NC with 100% of the precincts reporting by — 12,000 votes out of 4,000,000 cast!

  • Peter

    Thank you for all of your hard work, Sam.

    I’m asking for and recommending your book for Christmas and birthday presents. Thank you from someone who may not comment all that often, but checks the site daily.

  • Mike L

    Congrats Sam!
    You nailed the Electoral Vote and margin of victory which should be 6% + when west coast is counted.

  • Lorem

    So, Senate elections are still interesting!

    Georgia: early votes seem to have been counted now and with 99% reporting, Chambliss has 50.4%, so barring the remaining percent breaking 9:1 against him, he’s got it.

    Minnesota: with 99% reporting it’s still too close to call, although it looks to be leaning towards Franken.

    Oregon and Alaska: what in the world is happening here? These races aren’t even supposed to be competitive, yet the Republicans are posting slight leads in both right now with 70% and 81% precincts reporting respectively. The leads seem to have also held constant or expanded ever so slightly over the last two reporting jumps (which, like as not, means almost nothing).

  • Jon

    Pretty ironic good chap.

    You implemented your “personal prediction” bias and it throws-off the meta-margin in 2004 and skews the results.

    You choose not to implement your “personal prediction” bias into 2008….and it turns out it would have made the margin dead-accurate.

    Either way, between here, 538 and… gone are the days where B.S. and biased polls rule the roost. I seriously doubt we will have another election that isn’t accurately predicted before hand using the science you guys are throwing at the problem.

  • Sam Wang

    Jon – Totally right about the irony. This time around I was rather confident about the Obama +1% assumption, especially since it gave a small change, and was also in line with the trend up until the last few days anyway. But after being burned in 2004, I thought: don’t even go there.

    Thank you, everyone, for reading tonight – and all season.

  • Vicki Vance

    Thank YOU, Sam. Great work and a great site – it kept me sane the last couple of months (well, mostly sane, most of the time :)).

  • Walter

    Dr. Wang,

    Thank you, from all those of us who prefer science and reasoned analysis to spin and ideological drum-beating but don’t have time to compute it all at home. I think that sentence sums up the election result, too. :)

  • Nicholas J. Alcock

    Dear Prof Sam Wang,
    You said you would use a t-dist not a n-dist in 2012. You have since stated using a t-dist doesn’t alter swing states. No surprise there?

    I remember, you arguing that single polls were powerful:you stated this on your web-site.
    If they are so powetful why did you get NC wrong?

  • Nicholas J. Alcock

    Dear PEC,
    My comments have awaited moderation several times. Interestingly, they usually address Prof Sam. Wang’s methodology.

    Now, since Communist China holds most of the US Treasury’s debts i.e. trillions of dollars.
    Can you imagine Eisenhower or JFK/LBJ letting the
    Soviet Union hold in those days millions of dollars of US debt? Seriously, you are in hock to Commies? Haven’t you woken up to this fact?

Leave a Comment