Assessment of Trait Empathy Scales: Comparing the IRI and the TEQ on Convergent Validity with Sexual Prejudice, Gender Effects, and Temporal Stability

The Interpersonal Reactivity Index (IRI) is a widely used trait measure of empathy which suffers from low convergent validity. The Toronto Empathy Questionnaire (TEQ) shows improved convergent validity but lacks assessment of test-retest reliability and clarity as to the construct(s) it measures. We compared the TEQ and IRI convergent validities (N = 2693) with trait sexual prejudice, a construct opposed to empathy, using the Modern Homonegativity Scale (MHS). The TEQ outperformed the IRI in terms of convergent validity with MHS. The IRI empathic concern subscale was most predictive of both TEQ and MHS scores. However, its perspective-taking subscale predicted the TEQ but not MHS scores. Results support the view that the IRI fantasy and personal distress subscales are not empathic constructs. The TEQ increased predictive power for every model it was added to, suggesting another construct might be at play beyond empathic concern and perspective-taking. Participant gender predicted TEQ and MHS scores but did not moderate their association. However, longitudinal invariance analysis (N = 568) revealed mixed findings concerning all the scales, with TEQ possibly tapping into different constructs across time. Implications for research on empathy and its association with sexual prejudice are discussed.