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

A prediction on Biden v. Ryan

October 11th, 2012, 11:33pm by Sam Wang

Biden seems to have given Ryan a thrashing, cornering him on budget policy and besting him on social issues. Ryan came off as more knowledgeable on foreign policy than I was expecting. He appeared to basically agree with the Administration. Since the expectation on the right was that a buffoon (Biden) was up against a brilliant young Turk (Ryan), the outcome might be deflating.

I will make one prediction: whatever happens in the polls next will be assigned to the VP debate as a cause. Earlier today I pointed out that in key swing states, Romney’s bounce subsided rapidly – indeed, may have started doing so days ago. However, it takes days for polls to hit the news. In short, I expect the media to have a  case of post hoc, propter hoc. If and when things turn around, could the rooster (Biden) be about to get credit for the sunrise?

Update: sharp-eyed critic Fallows thinks Ryan did well, Biden did better.

Update 2: Reactions were a split decision among all viewers (44/48 for Biden/Ryan, CNN) and 50-31 among uncommitted voters (CBS). Some extra Republicans in the CNN sample; maybe they tuned in hoping for a second win. At the viewing party I attended in a Democratic-leaning venue, attendance was down by a factor of at least 5 from last week’s Presidential debate.

Tags: 2012 Election · President

72 Comments so far ↓

  • Arsalan

    It looks to me like Ohio, Virginia, and Florida are the key states. They could all go either way, and Obama needs to win just one of them to retain the presidency. 1/3. Come on, you can do it!

  • owlofminerva

    I doubt there are enough polls, but Wisconsin would be the state to watch to see if Biden actually moved the polls (highly doubtful, although I didn’t watch the debate).

    • Ralph Reinhold

      Ryan said several things that went against Wisconsin’s core. He may have helped swing more soft Romney supporters back to Obama than Biden did.

  • Mike

    All but one swing state… There’s that horrible Mason Dixon poll out from Florida. The internals make it sound like a big outlier, but it still counts in the averages… ugh! Taken 2 days ago, too!

    • Brian


      Outliers do not count when you’re taking a median instead of an “average”. That is why Dr. Wang’s analysis uses medians. Relax :)

    • Matt McIrvin

      Outliers still count when you’re taking a median, just not as much (the size of the most extreme outlier doesn’t matter, but the fact that it exists does).

    • Ralph Reinhold

      IMHO somebody should be running a site that uses Random Sample Consensus (RANSAC) or some other outlier detection scheme. I also think they should be running spectral analysis and higher order statistics to detect bias (frequency, skew and entropy should be consistent across all unbiased polls),. Votomatic uses Bayesian processes, but I don’t think that works any better on outliers than median.

      I plan to try that next time, if I am not involved in a series of deadlines.

  • John M S

    If Romney takes WI, NH, and IA, he won’t need OH……And RCP has all as Toss-ups…….also,….

    CNN’s poll of undecideds = Ryan 48% Biden 44%.

    • Mike

      That was a poll of debate watchers, which they admit is skewed 8 points to the GOP.

      CBS did the poll of undecideds, and they had Biden winning with over 50% of respondents. The HuffPo also had their own polling and Biden was crushing that one, too.

      Obama’s in a tough spot with the debate. CNN’s methodology is going to make it impossible for him to win.

    • Sam Wang

      I don’t think that poll is going to become the CW. Considering what is said on the right about Biden (buffoon) and Ryan (brilliant), presumably Team Romney came in expecting a blowout. Yet Ryan spent a lot of time on the defensive. Wind’s changing.

      As to combinations of states, I don’t care for that kind of fantasy baseball. You left out the other 2.3 quadrillion outcomes.

    • Mike

      He needs to win Nevada, too and all the other states like VA and NC.

    • NY Romneyite

      Disagree with Sam. Ryan went in with a different goal than the conventional wisdom suggested. Don’t forget, we politicos view Ryan as 1) conservative genius hero or 2) conservative evil grandma-pusher. Both are a little over the top, but that’s us with our partisan hats on.

      The Ryan pick was a signal from Romney to the base: get excited about me, I still like you guys. It enabled him to pivot to the center, which he has done somewhat convincingly (at least the polls say so) so far to the average voter, no matter what MSNBC says. Ryan had to get out of this debate looking like a serious and likable guy, with some ideas but not too many ideas. He did that. Leaves focus on Romney, who goes back up against Obama on Tuesday and is clearly the best debater of the four on the tickets.

      Don’t you think Ryan was prepared for a variety of Biden strategies? He intentionally kept his cool, while I think Biden’s strategy was to try to goad Ryan into getting angry, and being that evil caricature they want to put in low-info voters’ heads.

    • Sam Wang

      Despite the large change in the last week, this has been a polarized and immobile race. In that context, these events are about rallying the party faithful, not so much about swaying those elusive undecideds. Considering the last week of events, tonight was very energizing to Team Obama.

      I am unconvinced that the next few debates will do that much – though there is a grudge-match story line developing, always fun.

    • rollotomasi

      Ill chew my fingers off if romney wins any of those states.

  • Dave Kliman

    There was a poll that just came out with romney ahead in FL by 7 from oct 8-10…. what do you make of that?

  • Patrick Draut

    281 makes sense in terms of electoral math. Practically speaking, it is Obama carrying Ohio, Nevada, Wisconsin, Iowa and NH while losing Florida, Virginia, Colorado, and North Carolina.

    My question is, if we take off the table for now NC, Nevada, Wisconsin, Iowa and NH and assume they will stay red and blue respectively… we are left with four coin toss states.

    In order for Romney to win the general, he would need to sweep these states. If there is a 50/50 chance either candidate wins the state, would the likelihood of a Romney victory be 1/2^4?


    • NY Romneyite

      No, because the states will move in a correlated manner. If there is a 35% chance of Romney picking up a point or two nationally, he wins all 4 in that 35% of cases.

    • Patrick Draut

      That does not answer my question. I asked if there was a 50/50 chance in each state… Not a hypothetical scenario where Romney picks up points moving the odds of occurrence away from 50/50.

      Out of curiosity, how was the 35% chance of Romney picking up a point or two derived? Doesn’t sound very exact, or were you just using arbitrary numbers to show an example?

    • Some Body

      I’d add (in the capacity of the devil’s advocate), that it’s not just that one of the candidates may get a uniform swing; it’s also that the polls of all the states may get it slightly wrong in the same direction (e.g. because of a methodological issue that leads to sampling some part of the population incorrectly, or because of the Shy Torry Effect, or whatever). So, when the chances are 50/50 according to current polling, these are still not fully independent probabilities.

    • Sam Wang

      The evidence is that any such effect is quite small, i.e. less than 1%.

      The expected movement between now and Election Day is an SD of 1.8%, meaning that 68% of the time, the movement will be within +/-1.8%. For purposes of MM calculation, that would be spread across states fairly uniformly. However, changes in partisan enthusiasm will not be uniformly distributed. In other words, if the blue side gets really revved up, blue states get more of the benefit, and conversely for the red side. Given how extreme the race is, the more likely movement at the moment is in the blue direction – just as it was the other way around on Debate1 Eve.

  • mr oh

    “We weren’t told they wanted more security. We did not know they wanted more security there.” — Biden, speaking of the consulate in Libya

    But this was contradicted by State Department officials just this week, in testimony before a congressional panel and in unclassified cables released by a congressional committee

    • Obama 2012

      Trying to politicize Libya is absurd.

      Obama’s at fault for 4 Americans dying in Libya on 9/11? … if so was Bush at fault for 3000 Americans dying on US soil in 2001?

      I like what Biden said tonight about how usually we come together during these types of things… but now the disgusting Romney/Ryan ticket is doing anything they can to politicize these things… to attack the Obama foreign policy but when Biden calls Ryan out on it – he can’t really give a coherent version of what they would do different… all they want to do is complain.

      It’s just insane what they get away with.

      Imagine if the Democrats had tried to run against Bush on allowing 9/11 to happen in 2001 … the Republicans would have gone insane on them “anti American” … but really, this is the same type of tactic we see from Romney/Ryan on this today.

      there’s just no low that the GOP of today won’t stoop to.

      if we reward Romney/Ryan with the White House for this disgusting campaign… we deserve whatever hell they get us into.

    • Reason

      There goes that again. No. What he stated, correctly, is that he and the president did not know that they asked for more security. Which was true. Nor did Sec of State Clinton know. The internal investigation being done first found that the buck stopped with the Dep Sec of Foreign Ops, Ms. Lamb. Do not fall for that trap.

  • NY Romneyite

    Initial take from me: Biden in a squeaker.

    The more I think about this debate, I think smart people are going to think Biden won, and those same smart people are going to be surprised that the rest of the country is somewhere between total dead heat to lean Ryan.

    • Obama 2012

      CBS undecided voter poll disagrees with your assessment (and undecided voters are certainly not “smart people”)

      50% to Biden (31% to Ryan)

      but the real interesting thing to me is that Biden actually bettered Romney’s mark from last week (46%) in the same poll.

      this is an apples to apples comparison.

      the difference is: Romney only got his debate win by telling a continuous stream of lies. Biden got his win by refuting a continuous stream of lies and making clear the policy differences between Obama and Romney (when you can figure out what Romney’s stance is… at the moment)

    • Ohio Voter

      A Romneyite admitting a Biden win is like a +1/2% Rasmussen poll in favor of Obama. I’ll take it.

  • Insidious Pall

    The easiest way for Romney to win without Ohio is with IA and NV.

  • Joel

    Well, judging from my social media feeds (yes, this is anecdote not anecdata), Biden accomplished one thing that he needed to do: rally the Democrats. That’s really all that mattered here.

  • Michael Worley

    Good, close debate. This will make it about whose line of reasoning you prefer.

    • Ralph Reinhold

      I believe that this debate will shore up the party bases. There was a lot to take away from it for each side. Biden said a lot to make the Dems happy and Ryan said a lot to tick them off and vice versa.

      I think the problem with the undecided this time is not being able to chose one they are for when they don’t like either one. Some of those are based on religion and some on race and some on ideology. These people do not have one pragmatic bone in their body and are not able to make a choice of ‘the lesser of two evils’ because to them both are evil.

  • Obama 2012

    It was so awesome to see all of that Romney/Ryan “malarkey” called out for what it is. Since the news media won’t do their job, it’s up to Obama and Biden to do it.

    Biden did it tonight, now Obama has to do it Tuesday. Romney & Ryan are lying through their teeth. Repeatedly. They won’t give a straight answer on anything and won’t provide any details. This is completely unacceptable.

  • Obama 2012

    That’s the response Paul Ryan’s malarkey deserves.

    If you won’t give a straight answer, then you deserve to be laughed at.

  • Howie Weiner

    Biden did what he had to do tonight and did it well. He gave new life to the Obama campaign and did it in an engaging likeable real manner. Ryan seemed programmed and constantly on the defensive. I thought Biden scored biggest on the foreign policy sections especially Syria and Afghanistan. He made Romney-Ryan seem like they wanted to start another war and his statement, “we are out of Afghanistan in 2014, period,” was very decisive especially in comparison to Ryan’s waffling. But it was the discussion of abortion that may actually move some votes. Ryan basically said that they would try to legislate to make abortion illegal which not only contradicted Romney but I think scared a lot of people, especially women. It is this radical right agenda of neo-con war-mongering and social oppression that may tip some votes the Democratic way.
    The polls over the next few days will be interesting. The MM continues to fall, it should start to turn around by early next week if all I am thinking is correct.
    Then it is up to the President to come to his own defense. All eyes on next week’s debate!

  • Anbruch

    I watched the debate with the sound off and in terms of body posture it looked like Biden destroyed Ryan, who looked more tired and shell-shocked as the debate wore on. His tick to recover appeared to be short statements punctuated with prominent body gestures. I also found Ryan’s smirk far more irritating than Biden’s laugh. Biden’s body language by contrast was centered and assertive all night. YMMV.

    • Obama 2012

      Yup, GOP is trying to make this about Biden laughing at all of Ryan’s lies/BS (anything to divert from the issues) but Ryan was smirking all night himself and I don’t think it came off nearly as good as Biden’s responses which seemed like genuine reactions to the extreme nonsense being served up by Paul Ryan tonight.

  • Obama 2012

    you live in backwards land, I suppose.

    Ryan is the liar out of these two fellas, that’s what Biden was laughing at all night: all of Ryan’s absurd lies.

    the $716 billion Medicare lie that Romney/Ryan keep telling has been debunked numerous times. when you don’t cut benefits (and in fact; actually extend benefits) and extend the life of a program for an additional 8 years you are not cutting from the program.

    as far Afghanistan; give me a break. Biden clearly got the better of that exchange. Biden says we’re out in 2014; Ryan said… who knows what? He couldn’t give a clear answer. Just like he can’t answer how they’re going to pay for those $5 trillion in tax cuts without cutting deductions that help the middle class.

    That Romney/Ryan expect to win the White House without actually telling the voters what their plans for the country is insane.

    It’s crazy that this race is even close.

  • Obama 2012

    true – but Obama’s still in the lead. and I really think we’ve finally hit the bottom. if not Thursday then Friday… by Saturday I really think we’re going to start seeing an Obama bounce back.

    • Matt McIrvin

      If Romney’s really leading nationally (and essentially all the national polls say he’s at least 1 or 2 points ahead), I think the state polls will follow: EV count and Meta-Margin are going to go through zero before (and if) they bounce back. Obama has a swing-state advantage but it’s not as big as a 2-point gain in Meta-Margin.

    • Olav Grinde

      We shall see. If Obama has another unimpressive debate, then he’s in real trouble.

    • wheelers cat

      If Romneys leading nationally why havent the curves flipped yet?

    • Olav Grinde

      I suspect it’s because the election is not decided nationally. The election result is a cumulation of (mostly) winner-take-all, state-level races.

    • wheelers cat

      well…RAND is behaving exactly like Dr. Wang predicted today.
      leveling off 7 days after the initial effect.
      I really dont find much information in the shift graph.
      Such a small percentage of the respondents shift their votes, less than 1/2 of one percent, and the MOE is so wide– too noisy to really say anything.

    • William Ockham

      That’s because there is no shift. I know no one believes this, but the number of people who are both undecided about who to support AND will vote in the election, is indistinguishable from zero.

    • Doctor Gee

      Could you please explain your reasoning on this ??

    • William Ockham

      Examine longitudinal studies of individuals’ political views. Look at the amazing stability of the “PVI” of single year cohorts. For example, people who were born in 1960 (my birth year) have been consistently the most Republican cohort for the last 30 years. Pew has published some interesting data that shows the political preference by single year age cohorts in 2004, 2006, and 2008. The shape of the curves is remarkably similar.

      Take a close look a the relative stability of the Liberal/Moderate/Conservative (self-reported) split in the population of registered voters vs. the fairly large swings in the exit polls for conservatives and moderates (but not liberals). Compare this with the well-known problem of exit polled voters misreporting their vote and/or participation in the previous presidential election.

      Try to put all this data together in a coherent model. The only solution that I could come up with is that, in the course of an election campaign, all potential voters have already made up their minds (but a few don’t know it) about who they would vote for, but the real decision that everyone makes is whether or not to vote. Surpisingly, the marginal Republican voter is a self-described conservative without strong party affiliations while the marginal Democratic voter is a moderate. Liberals make up about 20% of the adult population and 20% of the actual electorate in every presidential election. Interestingly, moderates who vote Republican are about evenly represented in every presidential election. When the Democratic candidate wins, many more moderates who vote Democratic show up and correspondingly fewer conservatives show up.

    • wheelers cat

      William, you are ignoring the demographic timer timer and the biological basis of behavior.
      About half the population has organic conservative tendency, half has organic liberal tendency, due to genetic diversity.
      But the half of the pop that has organic conservative tendency PLUS dark-skin AND/OR female has been repulsed by the GOP policies.
      This is the last election where an all-white party (like the GOP is currently) has a chance at the WH because of demographic evolution.

    • William Ockham

      Not ignoring the demographics at all (I don’t know anything about the genetics argument you make).

      The “PVI” of a cohort seems to be determined by the party holding the presidency when that groups comes of age to vote. If the current president is thought of as a success (like Reagan and Roosevelt), that cohort tends to strongly favor that party. If that president ends up as a failure (measured by popularity), then that generation tends to favor the other party. I should note that family traditions will have a large impact as well.

      However, politics isn’t that important to most people. If a political view impinges on something that really matters to a person, that can cause a real shift in their view. The current Republican freakout over race and ethnicity is one example. Religious conversion is another.

      The most serious issue that the Republican party faces for the future is Texas. Without Texas, the party has no reasonable path to win the presidency. The median age of Hispanic citizens in Texas is 18. By 2024 at the very latest, a Republican party that loses Hispanics by 2-1 will be in trouble in Texas. Although I am a Texas Democrat, I think this is very dangerous for the country as a whole.

    • seemlessweb

      Dangerous in what sense? If you mean dangerous in the sense of a reactionary response to the structural demographic evolution that wheelers cat mentions (which is very real and just about here), I agree. If you mean it in a different sense I’m wondering what you are hinting at.

    • Matt McIrvin

      Some of the state polls still incorporate pre-first-debate data, don’t they? Especially for the bluer states that haven’t been as heavily polled.

  • Ram

    This was a wash! Probably Biden accomplished what Obama campaign wanted him- to stop or slow the Romney momentum. He undid some of the last week’s Obama’s NO SHOW by aggressively defending Obama and attacking Romney ticket. Ryan was impressive in his first major bout.

  • Matt

    Pollster’s model seems to be fairly responsive to changing events. In the national and the state polls it has bottomed out and maybe even reversed just a bit. Sam’s model seems to lag just a bit, but may well be more accurate. In any case, the idea that margin will keep falling is belied by the fact that the latest state polls aren’t any worse than earlier in the week and me be just slightly better for Obama.

    But I guess for some it is more fun to watch the line and extrapolate rather than think about the numbers that are actually driving it.

  • A New Jersey Farmer

    Romney will not win any of those states.

  • Bob D.

    If the polls do turn around, and if the media picks up the narrative (albeit post hoc ergo propter hoc,) I could imagine this propelling Biden to run for president in 2016 despite being 74 years old. I think we’ll need to wait and see how the narrative plays out, but if a myth is constructed that says Biden saved the election from the jaws of defeat with a masterful debate performance, I could see him running on that in 2016.

    • wheelers cat

      Never happen. Since Reagan’s [alzheimers] presidency, the trend is for the younger candidate to win. The only exception is the disputed election of Bush vs Gore, a difference of two years.
      The reason for this is 1) telegenics and 2) youth culture worship.

    • Matt McIrvin

      Biden could get away with doing what he did precisely because he’s not the presidential candidate (and also because he’s not a black man, and doesn’t have to walk the tightrope between appearing weak and scaring white voters).

      For 2016, I’m thinking Kirsten Gillibrand. Especially if Obama wins, and we want a stay-the-course type, but she might be good in any case.

    • wheelers cat

      Matt– 2016 is going to be the Battle for the Browns. Rubio is a lock for the GOP– its why he refused Romney’s VP slot.
      I like Julian Castro for the liberal ticket.
      Maybe share with Gillibrand.

  • Olav Grinde

    A 0.12% bounce?

    I’m taking heart in Sam’s calming words. Going to see where we’re at over the weekend. But at the moment, it looks spooky with previously Democratic safe states moving toward being in play.

    • Michael Worley

      You’ve got to compare it day over day. There usually is a small bounce in the morning update, even during the big Ryan drop (which may or may not be over)

    • Michael Worley

      Oops, meant the debate #1 drop

  • steve in colorado

    Biden certainly fired me up. And my wife (who is unpolitical) said of Biden: I like that guy! He’s funny.
    That said, it seems like Ryan was pushed to the ropes but held on, not quite getting blown out.

    The Obama ground campaign organizer here says that the Florida GOP vote scandal has basically killed Romney’s ground game, and that Obama canvassers are basically the only ones knocking on doors now. That’s certain to help quite a bit, so I wouldn’t count out Colorado quite yet.

  • William Ockham

    This is more for Sam than anyone else. I think I figured out why I missed this movement in the polls. Romney got essentially zero bounce from the RNC. This is pretty unusual, but this year’s RNC was unusually bad and it failed to fire up the Republican’s marginal voters.

    Because the DNC was so late this year, Obama was still at the peak of his post-convention bounce when the first debate happened. Romney is effectively getting his post-convention bounce from the first debate. But all of this is just noise. It’s a lot easier to tell pollsters that you are fired up and going to vote than it is to actually vote.

    The important thing to remember is that there aren’t any voters changing their mind about who they will vote for. Everything is just people moving in and out of the likely voter screens.

    If anyone doubts that, I have a thought experiment for you. Look at the movement reported by specific polling outfits (like the change reported by those two Pew polls taken three weeks apart). Compare the reported percentages of people expressing a preference between the two parties (i.e. ignore the undecideds). Now compare that the highs and lows that each of the two parties have received in the last 5 presidential elections. Do you really think that Obama went from being preferred by as high a percentage of the electorate as any Dem nominee in the last 30 years to doing worse than any nominee in the last 30 years? In the space of three weeks?

    National polls have always swung far more than actual results. Professor Wang’s system’s virtue is in smoothing that out by looking at the thing that really matters, state polls. Go take a look at which has one real advantage and that is its use of the consensus ‘fundamentals’ prediction as a Bayesian prior.

  • E L

    Is it time for James Fallows’ “Iron Rule of the Political Media: The story always changes”? It can only change in one direction after the beating Obama was given in the last 10 days. Biden’s job was to start to change that story. The political media should be ripe for change.

  • Brian

    PEC readers — this is off topic but if anyone is interested, I wrote a little extension for Google Chrome that, when clicked, will show you the latest Meta Margin, Obama EV, and Romney EV:

  • pechmerle

    Re reactions above to the debate itself:

    I think my fellow Dems — including Sam — are over-weighting the substance (on which Biden did better) and under-weighting the telegenics (laughing, interrupting, smirking, etc). On the latter, which is more important for the impact of a televised event, I think that Biden went much too far over the top. Many viewers will have felt (not necessarily thought, but felt) that Biden was disrespectful, not honoring the format, treating a calm, rational (-sounding, evasions notwithstanding) adult-appearing Ryan like a kid (inappropriately giving himself Biden authority from the age difference).

    My take: the post-debate bounce toward Obama-Biden will be there, but not as much as some of you seem to be expecting. (I wish it were otherwise.) We very involved types like a feisty, mix-it up political debate. The general public does not like that.

  • hoho

    Day after day all I hear from commenters here is that Obama has finally hit bottom and will begin rising,….meanwhile the ‘meta-margin’ continually drops.

  • Reason

    And back to you. See above. Biden did not lie. Ryan lied when he stated Obama did not call it an act of terrorism for 2 weeks. On a speech he gave the very next day, he called it terrorism. And he did it again the day after that. Before you drink the right kool aid, get your facts straight. This is all recent news.

  • keynes

    With the presidential election a little more than a month away, political science professors from the University of Colorado have updated their election forecast with more current economic conditions and are standing by their prediction that Mitt Romney will win.

    The professors — Kenneth Bickers of CU’s Boulder campus and Michael Berry of CU’s Denver campus — project that Romney will earn 330 of the total 538 Electoral College votes.

    “The economic fundamentals still suggest it’s a difficult terrain for any incumbent to run for re-election,” Bickers said in an interview.

    President Barack Obama is expected to receive 208 votes, which is down five from the initial prediction that the professors announced on Aug. 22.

Leave a Comment