Do GOP Voters Care About Electability?

Alex Lundry and I have a new post up at Model Politics.  In a YouGov survey from last week, we included an experiment.  After GOP voters had been asked which candidate they supported in the primary, we randomly assigned them to see Intrade probabilities for the GOP nomination, for the general election, or both.  Then we asked them a second time which candidate they supported.  The goal was to see whether knowing something about electability would change their preferences.  Indeed, it did:

In the graph we examine all respondents, regardless of which probabilities they saw.  In total, 35% of them changed their preference.  And unsurprisingly, they moved to the two candidates who did best on Intrade: Gingrich and Romney.

Our post has further details, such as:

  • Romney is helped by his general election probability much more than his nomination probability.
  • About 25% of Gingrich supporters defect to Romney after they’ve seen the general election numbers from Intrade.
  • Paul and Johnson voters tend to move to Romney.  Bachmann, Perry, and Santorum voters tend to move to Gingrich.

Ultimately, our results suggest that GOP voters are indeed willing to change their votes.  Moreover, they suggest that strategic considerations like electability can be persuasive.

See more in the post.

7 Responses to Do GOP Voters Care About Electability?

  1. Sebastian December 14, 2011 at 1:50 am #

    yay for slope-plots!

  2. shams December 14, 2011 at 3:27 am #

    Has this been done in the past, for Dems and Republicans alike? I would think it would be a more or less constant phenomenon.

  3. Adano December 14, 2011 at 11:34 am #

    Okay. But I think you miss an important point here.

    Maybe GOP voters care about electability (as you show), but maybe they don’t know enough about each candidate’s electability for that consideration to influence their decisions. After all, if they already knew which candidates were electable, then they already would have flocked to Romney and Gingrich even without the treatment.

    These aren’t hypothetical candidates you’re asking about. They are real candidates, in the final stage of a very long pre-primary battle. Votes know about as much about the candidates as they are ever going to know. If electability is going to matter, then voters should be aware of it by now. And if they were aware of it by now, then your treatment would have made no difference.

    • Michael Tesler December 14, 2011 at 1:23 pm #

      Bartels shows a substantial amount of post-Iowa learning, so I don’t think it’s fair to say, “If electability is going to matter, then voters should be aware of it by now. And if they were aware of it by now, then your treatment would have made no difference.” In fact, the treatment may more accurately reflect a post-Iowa political environment in which voters know a lot more and/or are paying much more attention to electability.

      • Adano December 14, 2011 at 7:15 pm #

        That’s probably true. The study shows that a clear signal about electability coming out of Iowa may have dramatic effects on voters. Of course, that signal may not look much look the Intrade numbers that voters were shown.

  4. Neil Malhotra December 14, 2011 at 8:22 pm #

    Cool post, John! A former graduate student of mine and I had a quick question about this.

    For the general election, isn’t the right price to use the probability of winning the general CONDITIONAL on winning the nomination? The general election price incorporates the probability of winning the nomination, which isn’t quite the concept of electability.

    The conditional probability can be calculated by dividing the general election price by the nomination price.

  5. John Sides December 14, 2011 at 10:21 pm #

    Neil: You’re right! That’s exactly what we should have done. Maybe we’ll try it again…