Morality traits still dominate in forming impressions of others

Melnikoff and Bailey (M&B) demonstrate that liking of people who possess moral and immoral traits depends on one’s current goals (1). They show that a merciful juror is liked more than a merciless juror by subjects playing the role of a defense attorney but not a prosecutor (experiment 1); a dishonest spy is liked more than an honest spy when working for the CIA against ISIS, but not when working for ISIS (experiment 2); romantically unattached men implicitly like an infidelitous woman as much as a faithful woman (experiment 3); and individuals who have acted selfishly in a dictator game like an altruistic partner who is likely to punish them less than one who is unlikely to punish (experiment 4). They argue that these findings pose a serious challenge to what we have called the morality dominance hypothesis (MDH) (2, 3), which states that moral traits are the strongest contributors to, and always contribute positively to, global impressions of others.

M&B’s results qualify the MDH but do not undermine it as powerfully as they suggest.

Their primary experimental findings (experiments 1 and 2) examine warm feelings rather than the global impressions that are the focus of the MDH (see also ref. 4). In fact, on their measures of desired friendship—arguably much closer to a global impressions measure—subjects consistently preferred to be friends with moral rather than immoral targets, notwithstanding the fact that they sometimes had warm feelings toward immoral targets who aligned with their current goals. At best, M&B have shown that morality sometimes does not dominate in transient liking judgments, but it dominates in more enduring judgments of desired friendship or global impressions.

M&B’s first study is limited in using a trait, mercifulness, which relates more closely to warmth than to morality, as their manipulation checks showed. Their second study confounds honesty with task effectiveness—the agent who is more capable of dishonesty is the more effective spy—and since he is pursuing a praiseworthy goal (defeating ISIS), it stands to reason that people would like him more. Nonetheless, subjects still preferred to be friends with the honest spy. M&B’s third study is difficult to interpret because its implicit and explicit measures conflicted.

The results of M&B’s fourth study are important and substantiate a speculation that we made in ref. 3: “Perceivers who are pursuing morally questionable ends may not always prefer morality in others, particularly if it means that their goals do not align with that of the target … the unconditional positivity of morality may itself be conditional on the perceiver viewing themselves as moral” (pp. 1288–1289). We commend M&B for finding support for this idea, which qualifies the MDH as follows: Morality is judged positively in others, so long as one views oneself as pursuing morally good ends, and morality is judged positively independent of competence and sociability (whereas the reverse is not true). Given that most people view themselves as highly moral (5), the exception cases studied by M&B do not seriously undermine the generality of the MDH.