The central virtue at issue in recent philosophical discussions of the empirical adequacy of virtue ethics has been the virtue of compassion. Opponents of virtue ethics such as Gilbert Harman and John Doris argue that experimental results from social psychology concerning helping behavior are best explained not by appealing to so-called ‘global’ character traits like compassion, but rather by appealing to external situational forces or, at best, to highly individualized ‘local’ character traits. In response, a number of philosophers have argued that virtue ethics can accommodate the empirical results in question. My own view is that neither side of this debate is looking in the right direction. For there is an impressive array of evidence from the social psychology literature which suggests that many people do possess one or more robust global character traits pertaining to helping others in need. But at the same time, such traits are noticeably different from a traditional virtue like compassion.