Jonathan Livengood writes:
I was reading a couple of your papers on voting (http://www.stat.columbia.edu/~gelman/research/published/rational_final6.pdf and http://www.stat.columbia.edu/~gelman/research/published/probdecisive2.pdf), and I wondered whether the results apply when people vote for third-party candidates. In part, I was wondering what it would mean in your model for a third-party vote to be decisive.
Is it rational (and under what conditions) to vote for a third-party candidate?
Since we are paying attention to some degree to what people say about their voting habits, I wonder what sense to make of the typical argument against voting for a third party. Namely: If I vote for a third-party, then I am voting against the two-party candidate that better represents my political views. If that candidate loses, then I will be responsible, since I would have voted for that person had I not voted for a third-party candidate.
Are people making a reasonable argument here or are they making an error?
My reply:
1. It can be instrumentally rational to vote for a 3rd party if they have a chance to win.
2. Because of the instability involved in multi-candidate elections, it can be difficult to assess this probability ahead of time.
This sort of decision arises in primary elections.
Livengood also sends along a paper (forthcoming in the philosophy journal Nous) that looks at voting and actual causation (roughly, causes of effects). He writes, “the paper picks on current theories of actual causation for saying weird things about three-candidate elections.” I haven’t looked at Livengood’s article in detail but I like that philosophers are using voting as an example. Voting is important!