The following paper is an argument that the expected utility of voting is (more or less) independent of the size of the voting population
Voting as a Rational Choice (PDF) by Edlin, Gelman, and Kaplan
I've often pointed out that it's an irrational conception of voting to think that a voting is only rational if you have a high chance of deciding the election. Close examination of this idea, I think, quickly shows it to be neither consistent with the reasons people have generally fought for suffrage nor consistent with any coherent notion of democracy. This paper doesn't make this argument; what the paper shows is that even if you do accept that calculation of 'decisiveness' is essential to rational voting, the addition of other interests -- in this case, how many people are potentially affected by the outcome of the election -- can still allow for the possibility that an election in which votes have a very low probability of being decisive can still be an election in which it is rational to vote. So, for instance, in a voting population where your vote is unlikely to be decisive, it still may be rational to vote if you believe that the outcome of the election will have a massive effect on the whole voting population. If you consider an election's outcome to be extraordinarily important, it can be very rational to vote even if you have little chance of deciding the outcome yourself.
Obviously, I don't think this is a real model of rational voting, since I don't think decisiveness has anything whatsoever to do with rational voting. But it's a very nice argument in that it shows that much more sophisticated and plausible models are available than those that simply rely on decisiveness. And given that a lot of people irrationally (in the 'normative' sense) do take decisiveness into account, this makes for a much better model of how voters in practice often do reason about voting. (For instance, unlike models based solely on decisiveness, the model in the paper gets the correlation of turnout to voting population running in the right direction: people are more likely to vote in larger elections, where their votes are less likely to be decisive, than in small elections, where their votes are more likely to be decisive. Likewise, it explains 'strategic' voting.)
