Examining empathic accuracy in a standardized task and in a naturalistic interaction: Associations, differences, and links with empathy

Empathic accuracy (EA) - the ability to infer others' emotions accurately - is typically conceptualized as a perceiver-level skill. This perspective implies that performance on different EA tasks should be correlated. Further, EA should be associated with trait empathy, given the theoretical similarity between the constructs, and that EA tasks are often used as behavioural measures of cognitive empathy. We examined the conceptualization of EA as a perceiver-level skill, and as a measure of perceivers' cognitive empathy. We recruited friend dyads (N = 137 dyads, M age = 19.61 years, SD age = 1.34 years) and tested associations and differences between EA measured with a personal task (rating the affect of a friend following a supportive interaction), and a standardized task (rating the affect of unknown targets discussing emotional events). Additionally, we examined associations between EA and cognitive and affective empathy. Analyses revealed low correspondence in EA between tasks and videos. EA was higher on each of the standard-task videos compared to in the personal task. Finally, greater self-report affective empathy, but not cognitive empathy, was linked to greater EA in both tasks. These findings challenge the notion that EA is a skill of the perceiver. Implications for conceptualizing and measuring empathic accuracy are discussed.