The Representativeness of Iowa Caucusgoers

I am here in Iowa with Lynn Vavreck. I’ll have more to report on our minor adventures later. But before the caucus takes place, it’s important to address a perennial concern: the unrepresentativeness of people who attend the caucus. This is a familiar refrain that typically involves claims about the high costs the caucus imposes on voters, the resulting low turnout, the domination by activists, etc.

Of course the caucus isn’t the least costly way to structure an election and caucusgoers are not a perfect random sample of the electorate. But claims about the unrepresentativeness of caucusgoers are generally overstated. (As are concerns about the unrepresentativeness of Iowa, but that’s another post. See here for a link to one piece of useful research, however.)

Here is some relevant data from a new and important book about the Iowa caucus, Why Iowa?, by political scientists David Redlawsk, Caroline Tolbert, and Todd Donovan.

  • Turnout in the Iowa caucus is low as a share of the total electorate (16% in 2008, which was a high-turnout year), but it is a much higher percentage of party members. Redlawsk and colleagues argue that because the Iowa caucus requires voters to register with a party, turnout should be measured as a fraction of party members. If you do so, 30% of Iowans caucused in 2008. This is by no means high, but it is much more in line with figures from other states’ primary elections.
  • In 2008, Iowa caucusgoers aren’t much more strongly partisan than registered party members generally. In an October 2007 poll, 50% of Republicans were “strong” Republicans. In a January 2008 post-caucus poll, 56% of Republican caucus-goers were “strong” Republicans. The comparable fractions among Democrats were 45% and 49%.  These are pretty small differences.
  • On the GOP side, you might think that caucusgoers were more religious or evangelical or concerned about social issues.  Not really.  In the October 2007 poll, 37% of registered Republicans called themselves “born again.”  In the January 2008 poll, 41% did.  Consider also the percentage who said that abortion was “very important” for their presidential vote: 41% of registered Republicans vs. 47% of Republican caucusgoers.  The same small difference emerges among those who prioritized gay marriage: 36% of registered Republicans vs. 33% of Republican caucusgoers.

One caveat: high caucus turnout in 2008 may have rendered caucusgoers more representative of party members generally.  If 2012 sees substantially lower turnout, caucusgoers could be less representative.  Another caveat: these findings don’t suggest that caucuses or sequential nominations processes are ideal overall.

Nevertheless, claims about unrepresentativeness Iowa caucusgoers tend to run well ahead of the data.

2 Responses to The Representativeness of Iowa Caucusgoers

  1. Rusty January 3, 2012 at 12:12 pm #

    Mitt Romney is mean to dogs. He scares us. Help me get my message out. *whine*

  2. paul g. January 3, 2012 at 10:13 pm #

    What is amazing about this Iowa meme is that what, to me, is the main point is completely avoided. It’s not about the representativeness of Iowa (though this matters) or about turnout (though this matters). The arguments of Lewis-Beck and Redlawsk, Tolbert, and Donovan are all essentially negative: EJ Dionne really isn’t that unrepresentative, the economy is really not that different, Iowans are really not that conservative.

    Where is the POSITIVE argument for structuring a presidential nomination system such that the same state is the first state, and that state uses a very unusual selection system (a caucus)?

    If you are going to argue that we need a “small” state to assure personal contact, then why Iowa? If you are going to argue that we need a caucus to foster deliberation, then why Iowa?

    When you layer even small levels of unrepresentativeness on economic, racial, and religious grounds (I’m not swayed by Michael’s factor analysis) on top, I remain skeptical of the Why Iowa argument. It’s simply not convincing.