Reliability of the empathy selection task, a novel behavioral measure of empathy avoidance

The empathy selection task is a novel behavioral paradigm designed to assess an individual's willingness to engage in empathy. Work with this task has demonstrated that people prefer to avoid empathy when some other activity is available, though individual differences that might predict performance on this task have been largely unexamined. Here, we assess the suitability of the empathy selection task for use in individual difference and experimental research by examining its reliability within and across testing sessions. We compare the reliability of summary scores on the empathy selection task (i.e., proportion of empathy choices) as an individual difference metric to that of two commonly used experimental tasks, the Stroop error rate and go/no-go commission rate. Next, we assess systematic changes at the item/trial level using generalized multilevel modeling which considers participants' individual performance variation. Across two samples (N = 89), we find that the empathy selection task is stable between testing sessions and has good/substantial test-retest reliability (ICCs = .65 and .67), suggesting that it is comparable or superior to other commonly used experimental tasks with respect to its ability to consistently rank individuals.