Princeton Election Consortium

A first draft of electoral history. Since 2004

Meta-Margins for control: House D+1.0% Senate R+4.2% Find key elections near you!

Ryanization of Senate races; a small Presidential bounce?

August 16th, 2012, 9:22am by Sam Wang


Yesterday I argued that Ryan’s VP selection would not so much affect the Presidential race as potentially tip the balance of Congressional power. Today, supporting evidence comes in the form of polls and messaging.

Senate candidates are very aware of the possibility. Senator Chuck Schumer, chair of the Democratic Senate Campaign Committee, is targeting Ryan’s fiscal competence. Elizabeth Warren (D), running for Senate in Massachusetts, hasalso chosen a new target…Paul Ryan. Heidi Heitkamp (D) in North Dakota is talking about consequences of the Ryan (and therefore GOP) budget for Medicare and the farm program. Their opponents, Scott Brown (R) and Rick Berg (R), give measured praise for Ryan. There will be more examples, but I think you can use the Google as well as I can. Even if voters are not swayed by this argument (which some claim), the party that takes the Senate will act as if they were swayed. Recall the Contract With America in 1995.

History of Popular Meta-Margin for Obama

In polls, you see that the Meta-margin has dropped rather sharply, from O+4.38% to O+2.64%. This apparent 1.7% drop is almost as large as the Sarah Palin bounce in 2008, which peaked at about a 2.0% change and lasted 2-3 weeks. For comparison, here is the 2008 Meta-margin time series (scroll down). I expect the Meta-margin to be accurate to within 0.5% (again, see 2008 to get an idea). Today’s change was driven by a load of Purple Strategies polls in OH, VA, FL, and other states. Their past results seem to be about on a par with Rasmussen, i.e. GOP-leaning.

A pollster-matched (i.e. before-and-after) estimate has the bounce at 0.2-1.7% (1-sigma CI, which I calculated as SD/sqrt(N)). Based on those numbers, the probability that there was any bounce (i.e. 0% or greater for Romney) is 87% (p=0.13, one-tailed test). This would not be enough for publication as a research result, but it’s probably true.

My guess is that the bounce is 1.0-1.5% — about one-third of the gap between Obama and Romney. For a definitive answer, wait a few more days.

Tags: 2012 Election · President · Senate

9 Comments so far ↓

  • The Political Omnivore

    Hello PEC,
    We have a review of the site up for our on-going digital politics series. Since I couldn’t find you on twitter, I’m writing you here:
    http://politicalomnivore.blogspot.com/2012/08/digital-politics-princeton-election.html

  • wheelers cat

    Rachel Maddow said every Republican has Paul Ryan as a running mate now, whether they want him or not.

  • Olav Grinde

    Well, I have at least two worries. The first one concerns a possible disparity in turnout, between fired-up right-wing voters and not-necessarily-voting liberals and moderates. This article by Lawrence R. Jacobs addresses how that could conceivably lead to a Romney-Ryan victory — polls not withstanding.
    http://www.cnn.com/2012/08/16/opinion/jacobs-romney-ryan/index.html?hpt=hp_t1

    • Sam Wang

      That’s a possibility, but most people who speculate about it are not really thinking about the available data, even political scientists. Silly pundits. Basically, enthusiasm is baked into the polls because pollsters make an attempt to measure it using likely-voter screens, which usually kick in around Labor Day.

      It is dangerous to speculate on enthusiasm, because enthusiasm looks different depending on where you sit. In 2004 I speculated about things like Democratic voter enthusiasm and how undecided voters would break. I added corresponding assumptions to my basic calculation. In the end, the polls themselves were accurate by themselves and I got a big surprise on Election Day.

      Based on my experience in 2004-2010, the bottom line is that voter enthusiasm is a factor that feeds the polls, which are in turn the true measurement. In other words, if you have polling meta-analysis, you have your answer. Of course, there still could be a day when you won’t like the answer…

  • Keith

    I saw a poll indicating that the president has a wide lead (43-20) among those least likely to vote. Any ideas about the methodology and/or likely accuracy of this?

  • Matt McIrvin

    I think that’s normal for a Democrat. It’s why higher turnout usually benefits them, and lower turnout usually benefits Republicans.

    It’s also why Republicans have an interest in making it harder to vote, and justify this by saying that if you can’t or won’t jump through the hoops you don’t really deserve it. This attitude in policy greatly helps Republicans.

    Younger people and poorer people vote less; single parents and people who have to work long hours vote less; retirees vote more, especially if they’re rich.

    There’s some worry about that many people who voted for Obama in 2008 are considering staying home this year because they think he’s got it in the bag. The thing is, the poll doesn’t explore whether they also said that at this point in the 2008 cycle. These people could probably be turned out if the Obama campaign bangs the drum about how he needs their votes.

  • Matt McIrvin

    Basically, enthusiasm is baked into the polls because pollsters make an attempt to measure it using likely-voter screens, which usually kick in around Labor Day.

    Since those screens don’t kick in for another week or two, presumably predictions based on summer polling like the one you’ve already made have some pro-Democratic systemic effect built in. Judging from past elections, it’s not huge, though I suppose this year’s more energetic voter-suppression efforts could make a difference.

  • Richard Vance

    Sam,

    Thanks for your work.

    Is there any correlation between money spent and the candidate that wins?

    Regards,
    Richard

    • Sam Wang

      Richard: yes, but it is not clear which causes which. Also, it’s not like the President of the United States has trouble getting a message out.

      The bigger role of money is probably in Senate and House races, where a flood of money can submerge the opposing candidate. This is why Karl Rove’s Crossroads GPS organization is focusing on them. A giant issue this campaign season. I wish it would get some press coverage.

Leave a Comment