In hypothetical scenarios involving two groups (nations or groups of workers), subjects voted on three proposals: one helped group A (their group), one helped B, and one helped both groups, more than the average of the first two but less than their maximum. When subjects voted for one proposal, most voted for the one that helped group A. This result is "parochial" because it helps the voter's own group even though it hurts the other group more. When voters could approve two proposals, they tended to approve the third proposal as well, and it was more likely to win. Approval voting can thus reduce the effect of parochialism, a bias toward one's own group, on election outcomes. In a second experiment, the authors replicated this effect using real-money payoffs. (Contains 2 tables, 1 figure and 4 footnotes.)