Who are the expressive voters?

Brennan and Hamlin (1998) predict that moderates are more likely to be expressive rather than instrumental voters, but do not test this hypothesis. Greene and Nelson (2002) claim to reject this, by finding that extremists are as likely to vote as moderates. We argue that Greene and Nelson's study was not a complete test of Brennan and Hamlin's hypothesis and we extend their analysis to provide a more thorough test. Our results imply that there is some evidence to suggest that extremist non-voters are less likely to be instrumentally motivated, providing some support for the predictions of Brennan and Hamlin.