"Campaign Effects in Multicandidate Races"
Andrew Gelman delves into a question I’ve been wondering about—does the political science literature about (the absence of discernable) campaign effects on election outcomes apply in more complicated races with multiple options? He says not really:
To put it another way, when the race is between candidate A and candidate B, you might as well vote for the candidate you prefer, no matter who happens to be in the lead—and the evidence is that people do just that. I’ve not seen evidence of any bandwagon effects in races between two candidates of opposing parties. But when the race is between A, B, and C, then, yes, your vote choice can definitely be affected by the possibility that C really has a chance of winning.
To put it yet another way: Elections are inherently more unstable when more than two candidates are involved.
That seems right on a theoretical level. If you look at something like the latest poll out of Canada you can see that a bunch of people who are telling pollsters they’ll vote Green or NDP are facing a tactical issue about whether it makes sense to cast a tactical vote for the Liberals. At the same time, there’s good reason to think that many Liberal voters would be turned off by explicit Liberal-NDP collaboration. So on its face, research about campaign effects in US presidential elections shouldn’t tell us much about this situation. But for a political scientist’s post on a political science topic, Gelman’s was uncharacteristically lacking in specific references to research on the subject—I wonder what’s out there, but couldn’t quite come up with the right combination of search terms to find anything relevant.