Princeton Election Consortium

A first draft of electoral history. Since 2004

Presidential discussion thread

October 22nd, 2012, 3:00pm by Sam Wang


The Meta-Analysis is most sensitive to states with the largest influence in determining the Presidential election outcome. You can monitor your favorite state in the right-hand sidebar under The Power Of Your Vote, which ranks states according to individual voter power. Ohio’s usually near the top. Each entry links to the latest from Pollster.com, our data source. With the 3:00PM update come several new polls from Ohio (and Wisconsin, and Iowa…) and consequent movement in the Meta-Margin.

In other news…in case you missed it, here’s the Science Friday interview on NPR with me and Nate Silver.

I won’t be able to watch tonight’s debate (I am visiting Chappaqua to talk with parents and teachers about how children’s minds grow and learn). It is unlikely that debate #3 will have much effect – for example, see the 2004 EV history. The cake’s ingredients are mostly in. And the cake’s in the oven.

Tags: 2012 Election · President

229 Comments so far ↓

  • Venkat Ranganathan

    I think it is better we have moderated mails lest this also gets degenerated into other news sites. The moderation we have seen before has shown that both sides of the views were represented in the comments

  • Albert Ericson

    Never mind – I think I figured out the current EV estimate of 290 — it is not based on the Median EV graph data, but the blue EV outcome distribution chart, where the highest possible outcome currently is at 290 EV. Is this correct?

  • Albert Ericson

    Can someone explain to me, if the red bar stretches from around 280 to about 323, and the yellow bar stretches from around 260 to 347, why is the current electoral vote count projected to be at 290? Wouldn’t it be around 296 or 297 (halfway on the red bar)?

  • wheelers cat

    well, Sullivan is linking PEC again and even the NPR transcript.
    http://andrewsullivan.thedailybeast.com/2012/10/who-is-the-favorite.html
    Must mean Obama is out of danger for now.

  • ChrisD

    The number of comments in this thread has dropped by about 20 and the order of comments has been reshuffled. You can’t simply scroll to the bottom and find the latest non-reply comments. How about starting a new “Presidential discussion thread” every time the MM is updated?

  • Jack Rems

    re: Data Corruption

    Sam-
    I think you should start a separate thread on data corruption. A month ago I would laugh at the idea any pollster cared enough about the graph on PEC to try to jigger it, or that it was even possible for them to do so. But your profile (and 538′s) seems to be a lot higher now.
    There are hackers (on both sides) who will be messing with you in 2014 and 2016, if they’re not doing it very effectively now.

    • Brad Davis

      I think this is a great idea even though I don’t fully understand why either side (D or R) would have much interest in publishing a lot of heavily biased polling results in an attempt to affect the aggregate. Do polls really have that much impact on GOTV efforts? Is the general electorate that fickle? I think that this would lead to an interesting discussion on the interface between statistical political science (?? I don’t know quite how else to describe this) and neuroscience, although perhaps that would be best after the election is over.

    • Ms. Jay Sheckley

      Brad Davis- Ive learned here that people are more likely to vote and win than to defend a beleaguered candidate. So, yes the electorate is exactly that fickle. Not nice, but the motive is clear.

  • ChrisD

    Fwiw, here are some currently active pollsters’ house leans (from most Rep to most Dem, LV only), according to Simon Jackson at HuffPost:

    Gravis
    Rasmussen
    Mason-Dixon
    ARG
    YouGov
    WeAskAmerica
    Gallup
    UNH
    ========
    PPP
    CBS/Quinnipiac
    Ipsos
    SurveyUSA
    Marist

    Not listed are several pollsters now doing state polls, e.g., Suffolk, Angus Reid, and PulseOpinionResearch in OH.

    http://www.huffingtonpost.com/simon-jackman/skewed-polls-nonsense_b_1923107.html?utm_hp_ref=@pollster

  • Reason

    Yep. MM down .06 and 2 less EVs. Next two weeks are going to be hectic.

    • Michael Worley

      surprised its holding steady. NV is the only non-reach state not O+ 1 or below

      I consider WI and PA reach states.

  • Jun Talabucon

    Romnesia has just become pandemic. It has been diagnosed in Afghanistan, Iran, Iraq, Syria, Libya, Egypt, Israel, Russia and China.

    • Olav Grinde

      Obama could have benefited from using the word Romnesia yesterday, and making a TV ad based on his speech. Humor goes a long way in slipping past the viewers’/listeners’ armor to score a point.

    • Ms. Jay Sheckley

      Um I would have been funny to some, not others.

  • Stuart

    Any idea why PollTracker has PA at O+2 when the min poll since Oct 9 has been 3 (from Gravis)?

    http://polltracker.talkingpointsmemo.com/contests/pa-president-12

    • Michael Worley

      Susquehanna (R)
      (Sponsored by Republican State Committee of PA)
      Romney 49.0 Obama 45.0 7

    • bks

      Sorry, wrong state. Sorry about that. Have Ohio on my mind!?

      –bks

    • Joey Bagadonitz

      Yeah, Polltracker is strange to me, because their projections seem to shift around a lot. They say they use their judgment to reject polls about which “serious questions of methodology and/or integrity have been raised,” and thus reject Zogby and YouGov. Then look at AZ: the only three polls from October show R+9, O+2, and R+9 again. Yet they only show AZ as R+1.7, and thus in their “toss-up” column, partly because they just don’t include YouGov, which was the most recent R+9 result. This seems odd to me, but perhaps others who are smarter can speculate/explain why this approach might make sense.

  • Steve in Colorado

    The Rand piece that I’ve found most interesting is the shift between candidates graph. I noticed that the blue and red lines have finally crossed again (they crossed to red being over blue after the 1st debate and now they are back to blue over red). I find this to be the most encouraging sign, along with the stability of WI, NV, OH and IA

    • Olav Grinde

      I’ve also been looking at the “Intention to vote” graph. It’s vital that the Democrats close the gap.

  • bks

    There are three states to watch: Ohio, Ohio, and Ohio.

    –bks

    • Ohio Voter

      Pretty much right on. Each candidate has a difficult path to the presidency without Ohio, but if polling stays pretty much as is in the rest of the country (Nevada and Wisconsin stay in the Obama column), then Obama wraps this up with just Ohio.

    • Ms. Jay Sheckley

      I’m reading everywhere about new high Ohio poll numbers for the incumbant, while here he holds steady at a pale blue +1. Anyone have any idea if we need to wait a week to see these polls folded into PEC? Were these polls real outliers? Anyone?

  • ChrisD

    Sorry. My mistake. I think the effect on the average still holds, though.

  • ChrisD

    The noon update should drop the MM a bit given two new ARG polls in NH and NV, O +2 in each. (ARG is said to have a 1-pt Repub house lean.)

    NH: O +0.5 => tie
    NV: O +4 => O +3

    Then again, another swing state poll might show up in the next 15 min.

  • Ralph

    I was looking at the RAND state election forecast data for FL, OH and PA that is up today. (I won’t link because tomorrow it will be a different list.)

    I noticed the difference in the shapes of the curves but that there is quite a correlation between peaks and valleys of one state vs another. For example, October 18 is at least a slight peak in all three.

    It would be great to have access to their data base and do a cross correlation between states and between states and the national trends.

    • NC Obama Guy

      I was thinking about the state data on RAND. I would not spend too much time on it. If you think about it they only poll 3500 total people for the entire country. If you just simply divide by 50 to get a rough estimate (obviously different weights would apply to different sized states) that leaves on approximately 70 individuals polled per state. I do not think that is a large enough sample size to get any useful information. This is similar to when a national poll like Gallop makes a statement about the stats from battleground states just drawing from the normal national data. It is very poor data I would think. The thing that concerns me about RAND is the graph showing likelihood to vote. No matter how bad Romney performs it continues to rise…

    • Olav Grinde

      I think you’re right. I suspect most of the state-level swings we’re seeing in the Rand Poll is just noise, given the small sampling of any given state.

  • E L

    I keep waiting for James Fallow’s Iron Law of the Political Media, cited by Dr. Wang, to hit: “The story must change.” We’ve had “Romney momentum” for several weeks. A new story is due. I wonder what tack they’ll take.

    • Sam Champion

      The narrative will be that Obama bended but did not break. Resiliency in the upper midwest – OH, IA and Wisconsin – is a source of strength. A better GOTV effort with a superior on-the ground infrastructure keeps VA and FL close.

    • Ms. Jay Sheckley

      We have two bookstores. So I think in terms of narratives, but not sure how this plays as a _political_ narrative may weight toward the underdog, even if he is Lex Luthor, However: A new villain [or group villain] is an awfully strong element to ignore. After the second debate we were directed to consider Tagg Romney’s fist. Then we hear he has invested, through Bain, in Ohio voting machines. So those are real objects to focus on, for the larger story of “lost” registrations etc which promise to become stories with many, many faces as voting attempts to progress. International voting overseers have arrived too, mostly in unoffical capacities. It’ll be hard to look away.

    • Olav Grinde

      “International voting overseers have arrived too, mostly in unoffical capacities. It’ll be hard to look away.”

      I have confidence that Fox & Co will manage that difficult feat just fine. Unless perhaps they invite Amy Goodman, Fareed Zakaria or Bill Maher to offer their analysis.

    • Reason

      I am not sure they will. So far, all I see are the talking heads commenting on how Romney was at least a calmer peacenick. That is all the soccer moms need to believe. But, who can tell with the MSM.

  • Ram

    Here is a projection.

    http://research.uvu.edu/DeSart/forecasting/october.html

    The key is to understand first debate moved the race about 2-2.5%. For Romney to win, the race has to move another 1.o to 1.5%. It is not easy given approval rating and improved economic fundamentals. That is where incumbency matter! Second and Third debate matter!

  • Reason

    I get it. And I know we should be more concerned about the state to state polls, not national. But a +2 PPP and now a +4 Ras just fuels the MSM’s frenzy to help R. Like it or not, MSM can and has influence voter turnout. 2 weeks is a lifetime in politics.

    • Tractarian

      “Like it or not, MSM can and has influence voter turnout.”

      This smells like self-serving BS. Show me some proof.

    • Reason

      Tac, you only need to go back to all the elections since 2000. The media hyped on things Gore did. They made him more negative. The media did the same with John Kerry (swift boat ads and the media letting them go unchallenged). 2008, Basically realizing that Palin was just too dumb to be one heartbeat away from POTUS, they ripped her to shreds. Now, they need a horse race to get that ad revenue. They let Romney lie and claim he won. Come on now. Also, this is not a tit for tat site. Cool out.

  • Reason

    Not sure what the sailing metaphors mean regarding this.

    • wheelers cat

      Reason, I think the polls and pundits are trying to catch the wind, that is all I mean. They are watching the wind on the water and the telltales is all.
      ;)

  • shawnthesheep

    wheeler’s cat–

    Weather helm is actually when the boat has a tendency to steer into the wind when the rudder is neutral. Lee helm is when the boat steers off the wind. This is influenced by the rig design, sail shape, hull profile beneath the water, etc. But aside from pedantic rant about sailboats, I agree with your point.

    • Olav Grinde

      Speaking of which, have you heard about the CEO who mixed his metaphors in an important speech to his employees?

      “We have long stood at the edge of the abyss, but today our company is taking a decisive step forwards.”

    • Rick in Miami

      Ah, but you forgot to tell Wheeler that you can correct weather helm by letting out the mainsail or sheeting in the jib. He must know that! I like a little weather helm – beats lee helm (luffing vs. accidental gybe!)

    • wheelers cat

      lol im a grrl.
      i love Dr. Wang’s weather metaphors, so im just trying to keep up.
      so Rick and shawn, should Obama reef the mainsail and run up a drifter at this point?
      and shawn, im used to a keelboat sloop. you are absolutely right.
      ;)

    • Ms. Jay Sheckley

      Olav, what a fine concession speech that would make. This year, however, a losing candidate has a surprising likelihood of making an acceptance speech.

    • Ms. Jay Sheckley

      wheelers cat- good to know, and glad to meet you.

  • wheelers cat

    …or what YOU get when YOU sail too close to the wind.

  • Reason

    Any idea why PPP now has R up 2? This does not make any sense.

    • wheelers cat

      weather helm.
      that’s what you when you sail too close to the wind.

    • wheelers cat

      what you GET when you sail too close to the wind.

    • Tractarian

      PPP showing R+2 makes sense. Take that number and factor in the margin of sampling error. It is perfectly consistent with the tied (+/- 1) race that has been apparent since debate 2.

  • Some Body

    We don’t know if this is fully free, or still moderated to some degree (although, judging by the fact that comments now appear immediately, it seems that the thread is indeed unmoderated).

    But I’m always for lax moderation. This case too gives us some hints as to what’s going on in the spin room, some exercise in refuting bogus claims. I’d be more for deleting the lot of these comments if they’d turn into a flood or just lose any semblance of relevance to the discussion.

  • wheelers cat

    Again, I think the polls and pundits alike are spoofed by asymmetry.
    red and blue voters simply react differently to stimuli.
    RAND intention to vote
    https://mmicdata.rand.org/alp/?page=election#intention-to-vote
    A bad debate performance makes red voters more likely to vote for their candidate because they are already committed. A bad debate performance for Obama made blue voters question their allegiance.
    Romney HAD to do well on the first debate. He was circling the drain, at less than 2% in the nowcast. The media resuscitated him, but the biological basis of behavior is keeping him afloat.
    Since Obama is back above the freakout threshold (70.3%, 72.6%) today in the 538 forecast the poll truthers will be out in force.
    Batten the hatches and get ready for some weather helm.

    • Reason

      Well said Cat. This morning on MSNBC, they are all spinning to try and make Romney acceptable, while admitting he got his rear kicked.

  • lojo

    I wonder if there is any data about impact of MSM CW on votes. In Bush v. Gore, didn’t MSM totally ignore late Gore push? I wonder if that cost him votes. No idea if you get more votes with perception of being slightly behind, tied or slightly ahead. Seems like all would engender similiar behavior. Being way up or down might demotivate but unsure about impacts of others. I wonder if there is any research about media impact/voter perceptions of outcome on actual outcome (and how sensitive it is to spread of perceived outcome).

  • DaveM

    The “great news…for John McCain” reflex among the punditocracy began, I think, during the Lewinsky crisis, when the Inside the Beltway crowd was obsessed with the idea that the Clinton presidency could not survive the damage while both polling and anecdotal evidence strongly indicated that those OUTSIDE the Beltway didn’t much like Bill’s behavior but wanted their government to get on with governing.

    In fact, were the relevant data available, it might be eye-opening (or even jaw-dropping) to see public opinion charted against pundit opinion–does the latter lag the former, does it move contrapuntally, or does it show little or no correlation?

    Obviously, the tough task would be the conversion of pundit expressions of opinion into data points…

  • JimmHck

    The numbers are correct for independent voters. Look at the crosstabs.

  • Jay Bryant

    My mistake. Mr. Romney won many of the other questions, too. It’s very late here.

  • Jay Bryant

    On digging more deeply, I found the results that include the exact question: “Who did the debate make you more like to vote for?”

    It was 25 for each candidate.

    For all but one of the rest of the questions and the contest in general, Mr. Obama won.

    Here is my source: http://blog.timesunion.com/politicssource/cnn-poll-obama-wins-overall-romney-wins-many-issues/942/

  • Jay Bryant

    Here’s a link to the CNN poll: http://www.cnn.com/POLITICS/pollingcenter/polls/3287

    That shows 48 for Obama and 40 for Romney

    Why lie about something I can verify in five seconds?

  • Matt

    CBS has it as follows:

    Before the debate, 24 percent said they leaned to Obama and 20 percent to Romney, with 54 percent completely undecided. After the debate, 46 percent said they would support Obama, 32 percent Romney and 19 percent were undecided.

    So there’s that… but all of this is kind of silly tea leaf reading. The actual polls on the aggregate will give us a better picture toward the end of the week.

  • skmind

    Here’s how I saw debate #3: http://bit.ly/RhKtqa

    Obama won it handily, but I am not convinced that it matters much. His foreign policy record was pretty strong, and I think people already knew that he was the superior candidate on that front.

    He could have blown it, but he did not

  • Martin

    SWING AND A WHIFF!!!!! Better luck next time John Martin!

  • Jonathan Wu

    John Martin, I can’t help to put a remark about your posts, you are entitled to your opinion but not the fact.
    It’s inappropriate to netiquette here, sorry.

  • jay

    Ahh… probably the wrong forum to post this, but if only Obama gave the performance from his 2nd and 3rd debates in the 1st…
    I suppose this late-in-the-game nervousness is helping Sam’s and Nate’s page-view count.

    • Dean

      Jay, I understand how you feel, but we have to move forward. Right now, Obama won two of the three debates. Time will tell if this is enough to get him his second term. I believe that his performance tonight is good for his base, and from seeing the PPP swing state debate poll tonight, it’s encouraging to think this could bring more independents over to him. Every vote counts in a close race, and the president did an excellent job tonight in trying to get more support. We can’t ask for more at this point.

    • Ms. Jay Sheckley

      The first debate guidelines were to answer the question directing themselves to the audience, not the opponent. that’s why the president was seen looking down and why he answered as he did.
      He had reason to believe he “won” the first debate and on radio appeared to. A depressing event after which he was not given his due

  • Dean

    Obama won the debate in many polls, CNN, CBS, Yahoo! (informal, anyone can vote multiple times poll), PPP. Intrade picked Obama as the winner. CNN’s voter intent poll is very small in comparison to the overall polling. I don’t pretend to know how this will affect the overall race, but Obama won the debate, and it probably won’t hurt his chances and may even help them, especially in swing states, where it matters most.

  • Ross C

    I think you are making the best argument to return to moderated comments.

  • Florida Swing Voter

    Come on! The Romney team shouldn’t be spinning numbers around princeton.edu.

  • Ross C

    Given more thought, I suppose what all the commentators refer to as Romney’s “momentum” is the tightening of the race. For example, there was a major credible poll today (Quinnipiac?) that showed Obama up 5 in Ohio, but all the headlines I saw, including Talking Points Memo, were something like “Romney Closes Gap in Ohio,” because Obama was up by 10 in their last poll a month ago. I guess that is a more interesting story than “Electoral College Continues To Drift Away From Romney.”

    And just as, “This is good news…for John McCain!” didn’t help prevent a landslide against him, so all this bloviating won’t make much of a difference in the end. Still, it’s a good motivator to take nothing for granted and donate, call, knock on doors, etc. I’m going to phone bank from the Obama dashboard tomorrow night. You can call from home and hopefully make a difference.

  • Brian

    With all due respect, John King doesn’t know what he’s talking about.

  • Dean

    I care more about swing state polls than national polls, and the PPP debate poll tonight, for what it’s worth, is a very good poll for Obama.

  • Ross C

    You are unlikely to persuade this crowd with your cherry-picked numbers. One set was flat out wrong (PPP) and the other was pretty sad by any measure. You may have better luck in the Washington Post comments section.

  • Joey Bagadonitz

    CNN’s sample was more Republican than the electorate as a whole, and they still gave PBO a +8 on the “win,” whatever that’s worth. I’ll take it!

  • Dean

    The PPP swing state debate poll shows Obama won in every demographic, and with independents, he won 55/40. As far as independents’ voting intent, that also goes to Obama, 46/36, and for overall voters, it’s 51/45. I hope I have the numbers right, but I’ll post the link anyway:
    http://www.publicpolicypolling.com/main/2012/10/obama-wins-debate-by-11-points-in-swing-states.html#more

  • Matt

    Your numbers are wrong:

    From the actual poll:

    http://www.publicpolicypolling.com/pdf/2011/1022PostDebatePoll.pdf

    Did tonight’s debate make you more or less
    likely to vote for Barack Obama, or did it not
    make a difference?
    More likely 37%
    ……………………………………………….
    Less likely 31%
    ……………………………………………….
    Didn’t make a difference 30%
    ……………………………
    Not sure 2%
    ………………………………………………….
    Q5 Did tonight’s debate make you more or less
    likely to vote for Mitt Romney, or did it not
    make a difference?
    More likely 38%
    ……………………………………………….
    Less likely 35%
    ……………………………………………….
    Didn’t make a difference 26%
    ……………………………
    Not sure 1%

  • Matt

    I’m curious why so many pundits talk as if this election is Romney’s to lose or that he has the advantage. I mean much of the post debate coverage seems to be, yes Obama won, Romney was too cautious, but Romney will still win because he has the “momentum.” It’s like they live in an entirely different universe from those of us hitting refresh on both Nate and Sam’s blogs.

    I see little evidence of “momentum” and if any perhaps toward Obama a bit over the past week. I see Obama with a clear advantage and so the status quo would be that he wins… over 90% of the time according to Sam on an election held today. So what am I missing? Why the disconnect?

    • Ross C

      I couldn’t agree more. I turned off the TV after the debate when I heard Mark Shields of all people say that “Mitt Romney came into the debate with the wind at his back.” Do these talking heads only listen to each other? It’s at complete odds with reality, unless you only read the headlines about cherry-picked outliers.

      And thanks for correcting those false numbers which were supposedly from PPP. You beat me to it. I know PPP is supposed to be D-leaning, but they’ve had very R results recently. Which makes these numbers even better. And if their numbers are correct on swing state results (O 51, R45) then that is promising as can be.

    • Ross C

      Remember, “this is good news…for John McCain!”? It’s that all over again. Even the left-leaning pundits can’t help themselves. I think at the first sign of anything but a landslide, they try to cover their…bases…so they won’t have lost all credibility no matter who wins. Pathetic but predictable.

    • Joey Bagadonitz

      Two things: first, the media is not about empiricism, logic, reason, etc, and never has been. Empiricism just isn’t who they are. It’s why I don’t watch cable news. There are perhaps some exceptions, but not many.

      Second: The rational thing for them to do as businesses is to maintain the horserace narrative as long as they can. So that’s what they do. Romney with momentum is good business even if the reality is murky. The first point makes the second point easier, because many can’t or won’t fundamentally fact-check themselves, because they lack the cognitive or informational tools, or at least don’t bother to avail themselves of them.

      In other words, a mix of stupid and cynical. Just my $0.02.

    • Terry

      I agree.

      In many respects what passes for journalism these says hasn’t changed much from the 40′s. Instead of hawking ‘ink’ they push blogs and cable.

      It’s much easier to promote the common ‘wisdom’ than offer independent thinking. I’ve actually met with overt resistance when trying to ‘educate’ MSM types about PEC and 538. Not sexy enough I guess.

      What Sam and Nate do seems risky unless one believes the approach and the numbers.

      Even confidence in the meta-analytic approach does not sooth my GI tract. Mind over…..

    • Brian

      Totally agree. Both Silver (because he was hyped as a media wunderkind) and Dr Wang (for those of us in the know) have got the evidence they could easily turn to, and yet there’s Chuck Todd every night yammering about momentum for Romney….

    • mgallagher001

      Of course Chuck Todd (as well as the other pundits) are yammering on about Romney’s momentum. The networks and cable have a lot of money tied up in pundits and election coverage. If they came out and said “Obama is going to win hands down” or “Romney has this sewed up no matter what”, would you watch? They are in the $$$ making business and a good horse race prints $$$ for them.

    • Some Body

      Somehow, I feel more comfortable about the outcome of the election with Todd speaking about momentum for Romney and Fox News bragging it’s a done-deal Romney win than when it goes the other way around. Most people haven’t actually voted yet, have they?

    • Matt McIrvin

      TV pundits are mostly aged and conservative-leaning folk, and they also benefit from the perception of dramatic reversals. They did the same thing in 2008, but events were just too powerful a driver then for anyone to think McCain was winning by the end of the cycle. This time, the race really is pretty close, so they’re going to play it up all the more.

      I’d say more, but it gets ugly.

    • MathinGA

      We get bumbling Mark Shields and they get Brooks and Will. ARghhhh… give us some smart and well spojen left-leaning pundits, maybe then the people would get to understanding the issues from the left point of vew better..

  • Corrupted

    PublicPolicyPolling ‏@ppppolls

    Swing state indys think Obama was debate winner by a 55/40 margin. 90% of Dems think Obama won, 81% of GOP for Mitt:

    Most amazing part to me is that only 81% of GOP thought Mitt won. I’ve never seen them admit to losing anything.

  • Dean

    Maybe Obama only got a small bounce from debate #2, and it’s unknown what will happen for him after this debate. But maybe these last two debates will help shore up his tenuous electoral lead if not propel him to lead further. Maybe instead of a small bounce he can get a decent hold?

  • wheelers cat

    #horsesandbayonets trending on twitter.
    Romney looked unbright there.

    https://twitter.com/search?q=%23horsesandbayonets&src=tren

    105,767 tweets per minute
    https://twitter.com/princessofSTEM/status/260584657643520000