Nonverbal emotion recognition and performance: differences matter differently

Purpose – This paper aims to explore and test the relationship between emotion recognition skill and assessment center performance after controlling for both general mental ability (GMA) and conscientiousness. It also seeks to test whether participant sex or race moderated these relationships. Design/methodology/approach – Using independent observers as raters, the paper tested 528 business students participating in a managerial assessment center, while they performed four distinct activities of: an in‐basket task; a team meeting for an executive hiring decision; a team meeting to discuss customer service initiatives; and an individual speech. Findings – Emotion recognition predicted assessment center performance uniquely over both GMA and conscientiousness, but results varied by race. Females were better at emotion recognition overall, but sex neither was related to assessment center performance nor moderated the relationship between it and emotion recognition. The paper also found that GMA moderated the emotion recognition/assessment performance link, as the former was important to performance only for people with low levels of GMA. Practical implications – The results seem to contradict those who argue that E‐IQ is an unqualified predictor of performance. Emotional recognition is not uniformly valuable; instead, it appears to benefit some groups more than others. Originality/value – The paper clarifies the emotional intelligence literature by providing further support for the predictive validity of emotion recognition in performance contexts, and by separating out how emotional recognition benefits certain population groups more.