This is the authors' abstract. We don't add key points for this paper.
The Emotional Accuracy Test (EAT) assesses performance-based emotional understanding by comparing perceivers’ ratings with targets’ multidimensional self-reports. Prior research supports its convergent and construct validity; the present studies examined test–retest reliability. Study 1 (N = 27; mean interval = 25.52 days) repeated the same targets and showed rank-order stability, r = .58, 95% CI [.26, .79], and absolute agreement, ICC(A,1) = .59, 95% CI [.20, .80], with no systematic mean-level change. Study 2 (N = 62; mean interval = 17.00 days) used different target sets and showed rank-order stability, r = .69, 95% CI [.53, .80], and consistency, ICC(C,1) = .65, 95% CI [.49, .76]. Absolute agreement was lower, ICC(A,1) = .52, 95% CI [.37, .63], with higher Time 2 scores. Target set was confounded with occasion and scored composites differed in length (three vs. four videos), so the mean difference cannot be uniquely attributed to occasion. Adjusting for mean rating intensity reduced Study 2’s correlation to .49, a conservative sensitivity estimate because adjustment may also remove accuracy-relevant variance. Together with prior validity evidence, these findings provide additional psychometric support for using the EAT in group-level and correlational research.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Jacob Israelashvili (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: