Election Reflections

by Dan Hopkins on October 28, 2012 · 5 comments

in Blogs,Public opinion

I don’t have a twitter account (and can’t imagine how folks who do can find time for much else).  But here are four quick election-related reactions to recent events:

  • POLLS SHOULD REPORT RESPONSE RATES: With poll watching becoming something of a national obsession, and with The New York Times’ Nate Silver now the talk of the sidelines at kids’ soccer games, I have a simple suggestion.  A leading survey research journal, Public Opinion Quarterly, requires that authors report response rates for any surveys, using the standard definitions of the American Association for Public Opinion Research.  Why don’t we ask the same of public pollsters?  With response rates as low as they are, that knowledge might temper just a tad the frantic discussions of individual polling results.  At the very least, such a policy might shift the conversation toward the high-quality polls.
  • WHY SO LITTLE ATTENTION TO THE SENATE?:  With so many competitive Senate seats, strategic donors on both sides presumably have to make tough choices about the relative value they place on the Presidency versus a Senate seat.  Now, in comparing the value of a House seat and a Senate seat, one calculation is straightforward, since there are 4.35 times as many House seats and they are held for only 1/3 as long a period of time.  So if I ignore the probability of influencing the majority party, don’t discount the future at all, and add in a few more assumptions, I can guess that a Senate seat is worth 13.05 seats in the House.  But how does one evaluate the Presidency relative to the Senate?  I’m totally open to suggestions.  My suspicion is that however the number comes out, it will indicate that our coverage of this election is overly focused on the Presidency, to the exclusion of the 33 U.S. Senators who will likely remain in office two years beyond the 2013-2016 Presidential term.  Depending on the Senate’s composition, many of the presidential candidates’ plans might not survive until the inauguration, let alone be enacted into law.
  • INDEPENDENTS ARE A MOVING TARGET:  Over at the Washington Post, Chris Cillizza writes about President Obama’s lagging performance among independents.  The article is careful to distinguish covert partisans—the majority of independents whose behavior is reliably partisan—from pure independents.  But even so, independents are a moving target—and right now, more reliable Republican voters are adopting the term than are reliable Democratic voters.  In one September 2012 survey, 32% of Republican leaners and identifiers at first chose the term “independent,” while among the Democrats, the figure was just 24%.  In 2004, the pattern was reversed, with more Democrats at first calling themselves “independent.”  Such over-time shifts make it hard to talk about independents as a stable category.
  • PERSUASION BEATS MOBILIZATION, AT LEAST IN THEORY: We’re reaching the part of the election cycle where there will be a lot of talk about the importance of getting out the vote, a.k.a. turning out the base.  Still, in talking about the trade-off between persuading undecided voters and turning out the base, there is a simple but important bit of math that commentators often neglect: 1-(-1) > 1.  That is, if I persuade a voter to change her vote from one candidate to another, I simultaneously add a vote to my column and subtract a vote from my opponent’s column.  If I mobilize a voter who was for me but otherwise would have stayed home, I get a vote while leaving my opponent’s tally untouched.  Turning out voters might prove far easier than persuading them—but if you can do it, the math is in favor of persuading your opponent’s supporters to join your side.

{ 5 comments }

Marc H. October 28, 2012 at 11:47 pm

The persuasion argument relies on the assumption that 1) the undecided voter will actually vote and 2) the undecided voter if unpersuaded will vote for the other major party candidate. Some undecided voters who remain undecided through election day will simply skip the election as the costs of voting outweigh the benefits. Thus, if I persuade this voter to vote for my candidate, but without persuasion they would stay home, then the GOTV math is the same as the persuasion math. 1 = 1.

Additionally, some undecided voters who do vote may skip that portion of the ballot (somewhat unlikely in presidential contests) or vote for a third-party or write-in candidate to A) support that third-party candidate or B) signal a protest / frustration with the main party candidates. I followed this latter path in the North Carolina gubernatorial election. Again, persuading this type of voter will follow the same the math as a GOTV voter.

Lastly, there’s an assumption that it costs less in terms of time, effort, and resources for turning out at a base voter than persuading undecideds who will definitely vote. What this ratio is, I have no idea. But if a persuadable voter costs less than x2 a GOTV voter than persuasion may very well beat mobilization.

Where the math becomes strongest is to persuade a “leaning” voter for the other candidate to switch votes to your candidate. But I think that is a somewhat different conversation because a leaning voter is different than an undecided voter, at least from my point of view. I guess it partly depends on the definition of undecided.

Hmmmmm October 29, 2012 at 2:35 am

Yes, polls should report response rates in their top line reports, but not in the newspaper write up. First, there’s no evidence that response ates matter: http://www.aapor.org/Response_Rates_An_Overview1.htm

Second,if the public doesn’t understand margin of error, they are certainly not going to understand response rate. It will only undermine all polls, given the very low response rates we see today.

Hmmmmm October 29, 2012 at 10:11 am

let me amend that: there is no definitive evidence that response rates matter.

Andrew Schwartz October 29, 2012 at 9:38 am

Interesting observation on the partisan leanings of independents. I wonder if this phenomenon cycles, such that partisans of the party out of power are more likely to identify as independents…

RedGuyBlueState October 29, 2012 at 12:26 pm

I think there are three main answers (for those actually thinking about it, as opposed to just following the media emphasis) to your musings about the overemphasis on the presidential race. vs the senate:

1) The presidency is sine qua non for the GOP. There is no chance of repealing Obamacare without it.

2) The nominal checks and balances have become heavily tipped in favor of the executive branch. The regulatory agencies play an outsized role in our day-to-day lives, and Obama has made a lot of law with executive orders and activist appointments without all the fuss and bother associated with getting a bill through both houses of Congress.

3) And possibly the largest issue, the Supreme Court. The present court is very narrowly inclined (most of the time, pace Roberts appalling sophistry on Obamacare) to enforce limits on government. A second Obama term would almost surely tip the court the other way and eliminate the remaining rickety restraints on federal power.

Comments on this entry are closed.

Previous post:

Next post: