MBS 93-22
Random Utility Models for Approval Voting
Jean-Claude Falmagne, Michael Regenwetter
A random utility model for approval voting is described, based on the
notion that each voter has a personal ranking of the candidates, with the
random utilities being defined, in a standard manner, by the probability
distribution on the set of all rankings. The set of candidates chosen by
a voter is made of the top K candidates in the voter's personal ranking.
The size k of the chosen set and the ranking are assumed to be independent.
Two independent classes of necessary conditions are derived, which are
used to reject the model for some well known data.