Princeton Election Consortium

A first draft of electoral history. Since 2004

Meta-Margins for control: House D+2.8% Senate R+4.8% Find key elections near you!

One last look at the Debate #1 bounce

October 17th, 2012, 10:00am by Sam Wang


Just a quick note: one last look at the state poll meta-analysis as of 8AM. Note that an uptick has already started, therefore not caused by Debate2.

last look at state poll meta-analysis before Debate2 data start arriving

And the Meta-Margin…

Tags: 2012 Election · President

71 Comments so far ↓

  • Ram

    Sam…

    The election equilibrium appears to be MM of +3 for President Obama since May. After second debate, MM may just stay between +2 & +3 unless something turns up.

  • steve in colorado

    When Romney started harping on the Libya thing, I was half hoping Obama would challenge him to a $10000 bet and reach into his pocket and pull out $10000 of monopoly money.
    However, Candy smacking him down was just as satisfying and a lot less risky.
    And when the guy asked him the question about what he has done the last 4 years, I wish he’d have channelled Clinton and ask him personal questions, ask him if he has been having trouble these last 4 years and lay some emotion on him.
    But really, I can’t complain about the performance at all. It wasn’t a huge win for Obama by any means, but he definitely came out stronger.
    The problem with following the polls now (if they go up for Obama) is whether they go up because the 1st debate bounce is receding or whether it is because of the debate.

  • Amitabh Lath

    “Please proceed, governor”. That was masterful. Obama was the professor, and Romney looked like a blithering freshman.

  • steve in colorado

    BTW, wife and I filled out our mail in ballots last night. Probably help my brother-in-law with same this weekend.
    Just to be sure, I’m going to drop them off personally with the County Clerk rather than mailing them.

    • Steve16748

      Did you vote for legal recreational marijuana? Steve in Washington State is going to do that when he gets his mail in ballot this weekend. Every registered voter in Washington gets a mail in ballot. It encourages thoughtful, convenient participation. We are a long way from Florida up here.

    • steve in colorado

      Well, I was tempted, but voted no on a constitutional amendment to legalize marijuana. It would have severely pissed my wife off and I had a hard enough time convincing her to vote for Obama because she is rather socially conservative. And if she didn’t vote for Obama, she might have convinced her brother, sister and mother to vote against him too.
      But I definitely voted for the Campaign Finance amendment.

  • Dharmachord

    Excuse me if this has been thrown out there previously, but is this effectively Romney’s “convention” bounce? Since the GOP convention made such a poor effort at showcasing Romney, do you think the first debate was the first time a lot of people were actually really paying attention to him, and thus the over-sized debate bounce? And, thus, could there be a larger retraction in his bounce that wouldn’t be typically seen after a debate? Or will we even see that since the news cycles are so compressed this close to the election?

    • betsy teutsch

      I agree. There were lots of Republicans looking to feel good about their candidate, and Debate #1 finally gave them something to feel upbeat and pumped about. Amazingly, many of them are women. Go figure.

    • Steven S

      Yes – I think debate #1 was much like a new candidate entering the race – a moderate Massachusetts republican. This is similar to what happened during the Republican primaries when new “front-runners” rose and fell and similar to the Ryan entry. A new person produces a big jolt until the shininess wears off. So I think you’re right because that’s what conventions are supposed to do. At the debate, Romney wasn’t tethered to the radical right, but unfortunately he also wasn’t tethered to the truth as it relates to his previous campaigning.

  • Rieux

    “Ante hoc, ergo non propter hoc”?

  • E L

    I think Obama’s debate performance was strong enough to return the election to bouncing between 1.5 and 3% in Obama’s favor where it was stuck from early June till late August. In other word, all the conventions and debates and ads and wailing and great gnashing of teeth and rending of garments and sound and fury has signified nothing.

    • Anbruch

      On no basis other than gut feeling, I suspect it goes to at least O +4% before tightening back toward 3% in the last week. That all dependent on debate 3 being a relatively boring event and no other wild event happening before the election.

  • Marco

    I think the first debate finally gave Romney a base. They finally liked their candidate.
    But the base is now seeing Romney’s quick move to the center and they might be returning to the the lukewarm stage.
    It will be interesting to see if debate #2 fires up the other base, with a swing the other way.
    Oscillatory motion around equilibrium point: it appears that this election is in a quadratic potential…

  • Olav Grinde

    It’s also worth noting that we have one more debate to go. The third Presidential Debate, on the 22nd of October, will focus on foreign policy and national security. These are areas where President Obama enjoys far more voter confidence than Governor Romney — and, as last night clearly showed, these are topics where Romney is way out of his depth.

    There is an excellent chance that the third debate will be scored even more heavily in Obama’s favor.

    Sure, there could be a Black Swan event. The GOP’s/SuperPACs’ flooding the airwaves with anti-Obama ad could have a strong effect. And, yes, there is a microscopic chance that the President stumbles badly on leading the country…

    Barring these things, I would be very surprised if President Obama does not lead by at least 3% in the week before the election.

    • Amitabh Lath

      Olav, I agree with you that Obama will move up in the next few days. Or maybe that’s just what I want to see.

      I am still a little worried about the neck-snapping speed with which the median EV and meta-margin are moving post debate. If things move up to 310-ish on the median-EV (say, 3.25-ish on the meta-margin) all within a few days, then I would wonder if anything actually happened post-convention. Or was it all media-induced fluff.

    • Olav Grinde

      Amitabh, that’s an interesting question.

      Some pollsters, such as Rasmussen, are clearly part of the media — and I for one believe their main agenda is to drive the narrative.

      However, there comes a point where perception becomes reality!

      Polls influence (as well as reflect) voter enthusiasm. In turn, voter enthusiasm drives turnout. And turnout can decide close elections.

    • Ralph

      Last night, he mentioned taking out the perps on Benghazi. Biden mentioned it the week before. If they strike between Oct 22 and Nov 6, the Republicans will be whining about it, but I don’t think the public will buy their claiming ‘wag the dog.’

    • wheelers cat

      Amitabh, if my hypoth about asymmetrical political behavior (POC enthusiasm) is correct, we should expect a more extreme and immediate response from events that spark enthusiasm for republicans, and a slower response from events that enthuse democrats.
      Alsotoo (a palinism) the red robopoll effect (also asymmetrical) will bias the poll aggregates.

    • Steve Schlichtenmyer

      I agree with Amitabh that much of the narrative is driven by the media “fluff”. For example, in the days immediately preceding the first debate, Obama’s meta-margin was near or at its zenith, yet it is also true that, in the two days prior to that event, the national tracking numbers already began to decline for the president and the race tightened. As shifts in the nationwide sentiment reflected in national polls seem to precede their effect in swing-state polls, it seems logical that the bump which Gov. Romney ostensibly received from the first debate may have been a fait accompli (i.e. the polls were ALREADY moving in Romney’s direction. I see a similar effect after the second debate. Obama’s poll numbers, as Prof. Wang noted, already indicated an uptick towards Obama well before the debate. Media outlets have a lot to gain from this election cycle…with the Citizens united decision, unprecedented sums of money are being spent and it is in their best interests to attempt to drive the narrative. They will profit nicely by keeping folks dialed in to a close race. Their market share will rise and advertisers (and candidates) will spend more for commercial spots. I wonder what is the correlation between poll
      numbers, media expenditures, and market share.

    • Amitabh Lath

      I know little about pollsters, but there are ways to amplify effects already present in the data. Most polls have a 10% response rate. If I had a selection criteria that only let in 10% of signal, I would worry that this was a horribly biased sample.
      It could be these people are more likely to sway with the media narrative than the general voting population.

      Or maybe the R-leaning of this 10% is less likely to pick up the phone after a leaked video of Romney, and the D-leaning less likely after debate#1. It wouldn’t take much of an asymmetry to move the polls 5 points or so.

    • xian

      another event might be the capture or “taking out” of the Benghazi attackers

  • 538 Refugee

    Is there any evidence that early and mail voting closes the enthusiasm gap at all? That area where the Republicans generally get a boost? I live in Ohio and got the form in the mail for to cast an absentee ballot.

    • Pat

      Actually, is there a website which tracks the progress of early voting? Maybe a reliable spreadsheet that updates all the numbers provided by each state on how many people voted, and ideally a comparison to the total number of early voters in 2008 and the total number of voters in 2008. That would be great, though I wonder if these numbers are actually made public in a transparent manner by each state…

    • Sam Wang

      Yes, Michael McDonald at George Mason University runs one. here.

    • Ohio Voter

      Could be. No way really of telling, I guess, though mail voting/early voting greatly favors Democrats in Ohio (in 2008, something like 4 to 1).

    • Craigo

      In Iowa, where early voting has been going on for a while, 50% of all ballots have been cast by Democrats so far. That’s slightly behind this point in 2008, so maybe we shouldn’t expect another near 10-pt margin, but it is a positive indicator for Obama.

      In Ohio, Dems have requested 29.5% of all ballots, compared to 23% for the GOP. That margin is also lagging significantly compared to 2008. Ohio is probably on a knife’s edge right now.

      In Virginia, 55% of all ballot requests have been by women, which is line with Virginia’s 2008 exit polls (Obama won both men and women in that state, interestingly, though women with a larger majority). That’s a wash.

      Republicans are outpacing Dems in Florida by a small margin, but I can’t find comparison stats for 2008 yet.

      On balance, the swing states where stats are provided are narrowly in favor of Obama, though not as much as in 2008.

    • Craigo

      Actually, Virginia is probably a bit tighter than I indicated. Early voting turnout is outpacing 2008 quite a bit almost everywhere, but the only areas that have to yet to outdo their 2008 totals already are places where Obama needs to do well, like Arlington and Fairfax County and the City of Richmond.

    • Show Your Work

      Also witness the GOP’s attempt to stifle early voting in Ohio, which an appeals court had blocked. Just yesterday the Supreme Court refused to hear the case, so the reinstatement of early voting stands. This is widely viewed as a big gain for Obama. Always interesting when one party’s best strategy for victory is to try to reduce voter turnout. 47% indeed!

    • Joey Bagadonitz

      Craigo and others, re: Ohio early voting: See this post from GMU’s McDonald: http://www.huffingtonpost.com/michael-p-mcdonald/early-voting-rising_b_1962843.html. Apparently, party registration in Ohio is determined by the last primary you voted in. So, since there was a contested Republican primary but no Democratic primary, it sounds like it throws any cross-party comparisons off. Maybe the gap HAS closed, but it sounds like one should be cautious drawing any conclusions from current data.

  • AW

    Dr. Wang,

    Long time listener, first time caller. Excellent blog here.

    Did Mr. Romney just resurrect the 47% argument with his binders full of women comment? I didn’t think anything of it when I was watching the debates (didn’t even catch it really), but when I reread the transcript of the entire answer and it didn’t sound good. That and he stretched the truth about how the events actually happened.

    I don’t think it in itself was a gaffe of terrible proportions, but it can add to the general narrative, that started in the primaries, that Mr. Romney is out of touch and soft on women’s rights.

    These things have a way of taking on a life of their own and I’m interested in seeing what happens in the next few days.

    • Ohio Voter

      A lot of people are taking the “binders full of women” comment many, many different ways. None of them good.

    • Ralph

      What will hurt Romney the most on the binders full of women is that the women’s groups claim they forced them on Romney, not the other way around. However, what will hurt him on women more than anything is ‘going home to cook dinner.’

    • Olav Grinde

      Indeed. Expect more focus on that binder.

      http://bindersfullofwomen.com/

    • CH

      “Binders full of women” is the sticky meme, but whether this gains further traction and damages Romney depends on the extent to which the rest of Romney’s answer is publicized and discussed.

      “IF you want to have women in the workforce,” he said three times. If. As if this is something we’re still on the fence about in 2012.

      A lot of male pundits are indicating they don’t understand why “binders full of women” is so off, which may impact the extent to which it gets full airing. But it was incredibly revealing, incredibly offensive, and full of potential, I think, to move the poll numbers with women.

    • Show Your Work

      Methinks the “Binders full of irrelevant has-been politicians” just got a new entry!

    • Olav Grinde

      @SecureCare: Thank you so much for the link to the review of that binder. I see that over 3,500 people have found it useful — so surely this is one of the best office-supply products currently being offered through Amazon.

      Funny thing is, I was just looking for a good binder to keep track of my ongoing projects. That this one should be extra useful in the (unlikely) event that I should become a governor or CEO of a venture capital company — well, that is surely a great bonus feature!

  • Barb

    Romney reminds me of certain males of an older generation in my family when it comes to women. They try hard to master the newfangled lingo and are so pleased with themselves when they point out how good they are to the ladies. Why they even allow them to get off work early so they can do their duties at home! Bless their chivalrous, old-fashioned hearts, they are trying, but they just don’t get it.

    What they don’t understand is that attitudes towards women have already moved on from “let’s help the ladies out”, to “women are competent, period”. He thinks he scores points by pointing out how he rustled himself up some women, while we women are wondering why he didn’t already know a bunch of competent women to begin with. Finding competent women from a binder would be acceptable feminism in the 80s – just making the effort would get you brownie points. Not in 2012.

    • Ohio Voter

      My wife, who is pretty blasé when it comes to anything political, was appalled that Romney reduced his Chief of Staff to nothing more than a woman who needed to get home and feed her children.

      Additionally, as a man who mostly telecommutes, I was offended at the notion that my wife is to be expected to come home after her busy day at the office and cook dinner for my incapable self.

    • 538 Refugee

      Obviously Romney is all for women in the work place. They get paid less so it is good for business.

  • isthisrob

    Sam:

    Gallup has Romney up +6 LV and +2 RV. They seem to be way out there compared to everyone else.

    Your thoughts?

    • Matt

      I believe Sam’s answer came 2 days ago when he told us to ignore national polls and focus on the big picture of state polls.

    • steve in colorado

      Just saw this Gallup- confusing as they seem to have more non-partisan credibility than Ras and yet showing a much more republican result…

    • Ohio Voter

      In the words of Nate Silver,

      The first rule of poll analysis is that if a poll looks like an outlier, it usually is.

    • Olav Grinde

      Outliers happen.

    • Craigo

      This is not meant as a fully-fleshed critique, but I’ll note that Gallup’s likely-voter model is attitudinal-based, rather than demographic-based – i.e., they weight responses by the expressed likelihood of voting (and also past voting record, which is also a decent indicator).

      That’s much more methodologically sound than the party ID, which is simply absurd, but I personally wouldn’t do it for one simple reason – attitudinal questions are more error-prone than demographics. Very few respondents will think to lie about their age or race, but lots of people will falsely claim that they’re “certain” to vote, or that they voted in the previous election – see social-desirability bias. There’s a lot of uncertainty in that for my tastes.

      Gallup’s polling director noted that they use this approach because they do not want to try to predict the demographics of the electorate. He’s well-accomplished and I respect his opinions, but that’s simply wrong. Sex, age, and race vary little from cycle to cycle, and in predictable ways.

      Bottom-line is that “voter enthusiasm” is an inexact (according to Gallup, Democrats were significantly more enthusiastic than Republicans in 2004) and somewhat volatile predictor of the electorate. So if one side is relatively dispirited and one side is jazzed, that will swing the needle. Note that the entire sample is between the two presidential debates.

    • Green Caboose

      The internals of the Gallup poll show Romney up 22 points in the south and own 4-6 points in the other three regions.

      One challenge with national polls is that the proportion of respondants who actually show up is probably a lot less in non-swing states than in swing states, due to the focused GOTV efforts. So the national vote totals likely overstate the votes from non-swing states. Just how much is unknown, but it’s a safe bet that enthusiasm among the GOP base is building to a very high peak and we’re seeing that in the behavior of respondants.

    • Olav Grinde

      Green Caboose: “…it’s a safe bet that enthusiasm among the GOP base is building to a very high peak and we’re seeing that in the behavior of respondents.”

      No doubt. But I think Obama’s performance, and Biden’s take-down of Ryan, as well as Obama’s coming debate on foreign policy and national security, is sure to fire up the enthusiasm of Democrats.

      Also, don’t forget that Barack Obama is busy doing things as President — som of which may well get positive attention.

      Things are, for instance, continually happening behind the scenes in the Obama Administration’s highly effective war on terror, which has seen al-Qa’ida and its allies reduced to a shadow of their former selves. Don’t be surprised if the culprits behind the Benghazi attack are hunted down. It would also be significant if the USA were to eliminate yet another key al-Qa’ida figure such as Ayman al-Zawahiri, the Taliban’s Hakimullah Mehsud, or Sirajuddin Haqqani.
      .

      The Rand poll shows that the Republicans currently enjoy a 2.3 % advantage when it comes to “intention to vote”. I would not be surprised to see this gap closing in the weeks ahead.

      https://mmicdata.rand.org/alp/?page=election#intention-to-vote

  • Louis in NY

    Indeed, national polls are almost meaningless. Nevertheless, Green Goose is correct to point out the internals of the Gallup poll:

    Obama Romney Margin

    East 52 48 O+4
    Midwest 52 48 O+4
    South 39 61 R+22
    West 53 47 O+6

    • Craigo

      I agree with the internals point (though I’ll note that this is yesterday’s sample – the Romney +6 internals have not been posted yet), but I have to strongly disagree that national polls are meaningless. It’s a matter of knowing what you’re looking for, not dismissing everything because it’s not immediately obvious.

  • Louis in NY

    Indeed nationals are informational, however, the states EV are what matters. As such, Sam’s model adroitly speaks to this reality.

  • Rick

    One scary thing is a the new Marquette Law School poll out today. It shows O +1 49/48. On 9/30, the same poll was O +11, and on 9/16 it was O +14. That’s a very disturbing trend. Did they change their methodology because the previous two polls were outliers? Or is Wisconsin trending red in a very serious way?

    • Froggy

      Rick, the whole idea here is that we don’t sweat every single poll. WI has tightened up since the first debate, but Obama has led in the last 15+ polls there. Just relax.

    • Obama 2012

      Hometown son effect… I imagine some low info types are voting for romney/RYAN because he’s from Wisconsin.

      Might make Wisconsin closer than it otherwise would be, but still confident (especially after last night) that Obama will pull it out.

    • Mike

      And it was just after Ryan starred in the VP debate, too… Ryan bounce?

  • Joel

    Sam,

    I’m wagering that you’re someone who has sat in on graduate student committees. The whole Libya exchange reminded me of a floundering graduate student trying to bullshit his way through a point in his comprehensive exam. A tough committee just gives that student rope to hang himself with.

    Or in other words, “Please proceed, governor.”

  • Gregg

    As an avid Obama supporter I am of two minds: I follow Sam and Nate and understand polls and taking the long view. However, I think the trends showing Romney are very worrisome. I think a fair bit of Obama’s support was soft or ambivalent. His campaign made some huge mistakes that, I think, in hindsight, will be discussed in years to come. I believe the debate was the manifestation of those mistakes/beliefs/strategies. Obama was running on defining Romney negatively, to turn voters off. And in the process he failed to adequately create a clear/positive/forceful theme, meme that resonated. In fact, the debate highlighted the vacancy. They also overestimated the value of the multi-million dollar commercial orgy. Folks in the swing states are sick of them, confused by them and left numb. NPR profiled some FlA moms who articulated so well the confusion and disgust with the ad wars. These wars left Obama more vulnerable because it exposed a lack of coherent theme and purpose to his re-election. That’s why 70 million tuned in; they wanted the reality series and they got something mangled: Obama in a xanax coma and moderate Mitt. I hope Obama can undo the damage.

    • Reason

      I am new to this poll game and even I know that the Gallup poll has serious flaws. I am of the same hive mind with this site. That the only polls that one can get an objective view of the data, are the state polls. National polls are trending one way or the other. The Gallup poll claims Romney has a +22 advantage in the south. I do not buy that, but so what? The only real southern states in play are Virginia and Florida. Even with those, Romney still loses. This is a race of EVs, not PVs. 2000 showed us that. I see no reason to panic at this point. So if the national polls are still +R, will we all stay home or vote? That is the real issue.

    • Matt McIrvin

      If Obama were to really win without a single state in the South, that would be… really something. It hasn’t really happened since, arguably, 1908, though there were several years when the winner lost the entire Deep South.

      Obama could have won without VA, NC, and FL in 2008, though, and Thomas Schaller was arguing around that time that Democrats could actually write off the South and win. The shifts in electoral votes have made it slightly harder this cycle, but it’s still possible.

    • Matt McIrvin

      …I looked into the history of non-South victories some more: Coolidge won essentially without the South (but with Kentucky, Missouri and WV) in 1924, but the last winners who were shut out of the South including all of those states were James Garfield in 1880 and Benjamin Harrison in 1888, during the era when Democratic Party support was starkly Southern.

      As yesterday’s xkcd points out, that doesn’t mean it couldn’t happen again.

  • Ms. Jay Sheckley

    I’d like to see the VP debate on the metamargin chart, and I’d like to know if Sam thinks the pretty fish hook uptick could include a Biden-Ryan bounce yet

    • Matt McIrvin

      Since the uptick seems to be from a bunch of YouGov polls that predated the VP debate, I think the answer is no, or mostly no. Any effect from Biden-Ryan is going to appear shortly, though.

      The Marquette poll seems to suggest that the VP debate may have actually helped Romney/Ryan in Wisconsin, but that’s a very special case.

    • Olav Grinde

      Matt — so in other words, we are still waiting for the effect of the Vice Presidential Debate and the second Presidential Debate to kick in?

      It’s very interesting that we are seeing a significant bounce even in advance of this. In fact, from my point of view, most encouraging.

    • Matt McIrvin

      I think we are still waiting for that, yes.

      Bear in mind, I haven’t seen any evidence from national polls that the VP debate helped the ticket at all. (They usually don’t, but it’s an unusual year.) But it’s been a short time and there’s a lot of spread in the numbers.

  • Reason

    Yes. But the MU poll had one day of polling before the debate. I did see that they thought Ryan did better than Biden. But I see that depending on where the majority of the polling took place.

  • Reason

    Oh, and I am in Virginia. Do not count us out. We had over 400k new voter registrations this election. Most of them women. Most in Northern Virginia. Very encouraging.

  • wufwugy

    The Black Swan this year will be everybody discovering that most of the polls were wildly wrong. Obama will win NV CO IA OH VA and possibly NC by more than last time. Then everybody is going to magically remember that all these polls had awful archaic methodology that only accounts for a certain type of voter

  • ChrisD

    Just another poll, but the putative key state:

    OH SurveyUSA 10/12-15: O 45, R 42

    Survey USA had it O45, R44 a week ago, 10/5-8

  • Silas Barbosa Mariano

    I believe that Obama will still win, voters are very divided and a lot can still happen until the polls open.

Leave a Comment