RESUMO
BACKGROUND: The lack of empathy is associated with several psychological and behavioral disorders, and it is important to assess this construct broadly, through multi-methods. OBJECTIVE: To conduct a psychometric analysis of the Brazilian version of the Multifaceted Empathy Test (MET), a computerized task that assesses emotional and cognitive empathy. METHODS: The samples were recruited from the community using the snowball method (phase 1: face-to-face; N = 142) and through social media (phase 2: online; N = 519). The participants completed the MET and the Interpersonal Reactivity Index (IRI) to assess the convergent validity between the instruments. To assess validity with correlated constructs (resilient coping and stress), the Brief Resilient Coping Scale and Perceived Stress Scale were used. A task was also implemented in the face-to-face application to assess facial emotions. The retest was applied 25 days later to a portion of the sample (face-to-face: N = 31; online: N = 102). RESULTS: It was observed adequate test-retest reliability for most items (ICC = 0.49-0.98), satisfactory infit and outfit indexes, discriminatory ability between sexes, weak convergent validity with empathy measures (r = 0.17-0.36), and correlate constructs (r = 0.12-0.46). MET presented good psychometric indicators, confirming its use in face-to-face/computer-based and online formats in clinical and research contexts. However, weaknesses were found regarding the cognitive subscale, demanding future studies to address larger samples to enable more robust conclusions concerning its adequacy. Further research on the instrument's internal structure can also contribute to its improvement.