The impact of moral judgment on happiness attributions in children and adults

• In Event: The interplay of mental states and moral cognition

Abstract

Recent studies show that people’s judgments about whether an agent is happy can be influenced by their moral judgments (e.g., Phillips, De Freitas, Mott, Gruber, & Knobe, 2017). Specifically, even if an agent is described as having a lot of positive emotion, people are reluctant to say that the agent is happy if she is leading a morally bad life.

We explored this effect using both developmental studies and adult studies. In the developmental studies, we find that children also show the effect. In the adult studies, we find that the effect is driven by judgment about the true self. Taken together, these results seem to suggest that children’s attributions of happiness can be impacted by judgments about the true self.

In a series of developmental studies, we asked whether the effect of moral status would also arise for children. For example, in one study (N=120), participants were randomly assigned either to hear about an agent who is nice (e.g., sharing toys with others), an agent who is mean (e.g., stealing toys from others), or a ‘baseline’ agent with no further information. In all conditions, participants were told that the agent “feels really good.” Participants were asked to rate how happy the agent was on a 1-4 scale. Results are shown in Figure 1. There was a main effect of condition (β = -1.52, 95% CI = -1.86, -1.17) but no effect of age and no condition x age interaction.

To test the hypothesis that this effect is explained by true self judgments, we conducted a series of adult studies. In these studies, we expanded the scope of our inquiry to look not only at happiness attributions but also attributions of sadness, love and hatred. For example, in one study (N=603), we used a 2 x 4 design in which we independently manipulated emotion (happiness, sadness, love, hatred) and fittingness (whether the agent’s life merited the emotion vs. did not merit the emotion). We then measured both straightforward attributions of the emotion and also judgments about whether that emotion stemmed from the agent’s ‘true self’ vs. was just ‘on the surface.’ The manipulation of fittingness impacted true self judgments for happiness (b = 40.53, p < .001) and love (b = 17.36, p < .001) but not for sadness and hatred. Importantly, we found that same pattern for attributions of the actual emotions themselves. The manipulation impacted attributions only for happiness (b = 44.75, p < .001) and love (b = 21.63, p < .001), but not for sadness and hatred. An ANCOVA including true self judgments as a covariate showed that once true self judgment are included in the model, the impact of the manipulation decreases (from p2 = .04 to p2 = .01), and that there was a significant effect of true self, F(1, 592) = 38.8, p < .001, p2 .= 06, 95% CI [.03, .10].

Authors

• Joshua Knobe, Yale University
  Presenting Author

• Fan Yang

• Yarrow C Dunham, Yale University

• Michael Prinzing, Yale University