Emotional Intelligence: What the Research Actually Shows

Emotional intelligence, broadly defined as the ability to recognize, understand, manage, and use emotions effectively, has been one of the most widely discussed psychological concepts in popular and organizational settings since Daniel Goleman's 1995 book popularized the term. The concept has generated significant research interest and significant criticism, and the gap between scientific findings and popular claims is large enough to require careful navigation. The theoretical framework for emotional intelligence has two main variants that differ in important ways. The ability model, developed by Peter Salovey, John Mayer, and David Caruso, defines emotional intelligence as a set of cognitive abilities involving the perception, integration, understanding, and management of emotions. On this model, emotional intelligence is measured with performance tasks rather than self-report, and it is empirically distinct from personality traits and general cognitive ability. The mixed model, associated primarily with Goleman's popular work and various commercial assessment tools, combines emotional abilities with personality traits, social skills, and motivational factors in ways that blur the conceptual boundaries. Research on the ability model of emotional intelligence, measured with the Mayer-Salovey-Caruso Emotional Intelligence Test, finds that EI as so measured is a genuine cognitive ability that is statistically distinct from personality and general intelligence, though correlated with both. Ability EI predicts several outcomes of interest, including social relationship quality, some aspects of academic and job performance, and wellbeing, though the effect sizes are typically modest and the incremental predictive validity over personality and general intelligence is limited. The more expansive mixed models of emotional intelligence have weaker scientific foundations. Self-report measures of emotional intelligence correlate highly with established personality traits, particularly agreeableness, conscientiousness, and emotional stability, raising questions about whether they are measuring something distinct from what personality research has already identified. Incremental validity of mixed-model EI measures over established personality measures is often minimal or nonexistent in rigorous studies. Goleman's claims about emotional intelligence in his 1995 book and subsequent writings substantially exceeded what the empirical literature supports. The claim that emotional intelligence accounts for more variance in life outcomes than IQ is not supported by the research evidence. The claim that the emotional competencies he identified can be reliably trained and that training produces lasting improvement in job performance has limited empirical support. The commercial EI training industry has been built on claims that are stronger than the science justifies. In educational settings, social and emotional learning programs have accumulated a reasonable evidence base as discussed in other contexts, though SEL programs are broader than what is typically meant by emotional intelligence training. The outcomes that SEL programs target, including emotion regulation, empathy, social problem-solving, and positive relationship skills, overlap with the emotional intelligence construct but are operationalized in terms of specific skills with clearer training protocols. In organizational settings, leadership research has explored the role of emotional intelligence in leadership effectiveness. Meta-analyses find positive but modest correlations between ability EI and various leadership outcomes. The intuitive appeal of the concept, that effective leaders need to understand and manage emotions well, is supported by qualitative observations and case studies as well as by the statistical relationships. The practical question of how to select for and develop these capacities in leaders is less clearly answered by the research. The most defensible conclusion from the research is that emotional perception, understanding, and regulation are genuinely important capacities that influence social functioning and wellbeing, that they can be assessed with reasonable reliability using ability-based measures, and that they have predictive validity for outcomes relevant to effective functioning. The grandiose claims of the popular EI literature, which overstated the uniqueness of the construct and its measurability and trainability, have not been sustained by rigorous empirical research. More modest claims are more honest and ultimately more useful.