Madeleine Horner, Tadeg Quillien and Adam Moore
Affiliation: University of Edinburgh
Background: Empathy, the ability to relate and react to the emotional experiences of others, has been identified as a strong motivator for prosocial behavior. We explore people’s intuitive theory of empathy, focusing on the role it plays in the causal model people use to explain and predict prosocial behavior. We computationally formalize the hypothesis that people intuitively understand empathy as indexing the weight an actor puts on the welfare of the recipient (relative to the self) when deciding whether to help. Conversely, we also test if people can rationally ‘invert’ this intuitive theory to make inferences about an actor’s empathy, given their observable behavior. Methods: We test this proposal by asking participants (N = 150) to read a series of vignettes in which an actor has the opportunity to help a friend in need. We ask one group of participants (N = 50) to estimate the prior probability of someone feeling empathy in the described situation and to provide ratings of the cost and benefit of the helping actions. We ask another group (N = 100) to provide judgments about the likelihood of an empathetic actor helping, judgments about the marginal probability of an actor helping in a given situation (with no information about psychological state), and judgments about the posterior probability of an actor being empathetic given their actions in a situation. Results: We found participants consistently expect empathy to increase the probability that an actor would perform an action (β = 0.35, 95% CI [0.29, 0.41], SE = 0.03). Moreover, participants seem to expect empathetic actors to be more sensitive to the potential benefits of an action when deciding whether to help (β = 0.07, 95% CI [0.01,0.12], SE = 0.03). We find weak evidence that people are able to invert this ‘intuitive’ theory of emotion to infer if an actor is empathetic given their actions (β = 0.12, 95% CI [0.08,0.17], SE = 0.02, WAIC = 2407.9, LOO = 2409.3). The best fitting model indicated that participants relied more heavily on their prior beliefs about the expected level of empathy given a situation when judging if someone was empathetic (β = 0.18, 95% CI [0.12, 0.23], SE = 0.02, W AIC = 2374.0, LOO = 2375.5). Discussion: We provide empirical evidence that people systematically expect empathy to be associated with prosocial behavior, suggesting that people intuit empathy as indexing the weight that an actor puts on the welfare of the recipient when deciding to help. We also find that people have more fine-grained expectations of empathy: empathy increases the sensitivity to the potential benefits of an action when deciding to act (as opposed to unconditionally increasing the probability of helping). Although participants consistently associate empathy with prosocial behavior, we do not find that they use this expectation to infer empathy from observable behavior. One explanation for this may be that participants inferred actors were unlikely to behave as described unless they felt empathy, relying on a different causal model than was intended.