What Is Emotional Intelligence? Science, Models, and Real-World Evidence

Origins of the Concept

The term "emotional intelligence" was introduced to academic psychology by researchers Peter Salovey and John D. Mayer in a 1990 paper in the journal Imagination, Cognition and Personality. They defined it as "the ability to monitor one's own and others' feelings and emotions, to discriminate among them, and to use this information to guide one's thinking and actions." The concept built on earlier work in social intelligence (Edward Thorndike, 1920) and multiple intelligences (Howard Gardner's interpersonal and intrapersonal intelligences, 1983).

The concept remained academic until Daniel Goleman's 1995 bestseller Emotional Intelligence: Why It Can Matter More Than IQ brought it to popular consciousness. Goleman's book made sweeping claims — that EI was more predictive of life success than IQ — that generated enormous public interest but also substantial scientific controversy. The popular and scientific understandings of emotional intelligence diverged significantly, and the gap matters for evaluating claims made in its name.

The Major Scientific Models

Emotional intelligence research has developed along two distinct lines with important differences:

1. The Ability Model (Mayer, Salovey, and Caruso)

The ability model treats emotional intelligence as a genuine cognitive ability — a form of intelligence concerned with emotional information processing. It defines four branches arranged from more basic to more complex:

1. Perceiving emotions: Accurately identifying emotions in faces, voices, images, and one's own physical states
2. Using emotions to facilitate thought: Harnessing emotions to direct cognitive resources — using mood to improve certain kinds of thinking (e.g., mild sadness facilitates attention to detail)
3. Understanding emotions: Knowledge of how emotions evolve, blend, and relate — understanding that anger can escalate to rage, or that complex situations produce mixed emotions
4. Managing emotions: Regulating one's own emotions and influencing others' emotions to serve goals

The ability model is measured by performance tasks — the Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT) — where items have objectively better or worse answers based on expert consensus and emotional knowledge. This makes it more like a traditional IQ test than a self-report questionnaire.

2. Mixed Models (Goleman and Bar-On)

Mixed models combine emotional abilities with personality traits and social competencies. Goleman's model includes self-awareness, self-regulation, motivation, empathy, and social skills. Reuven Bar-On's model includes emotional and social competencies, adaptive behaviors, and motivational components.

These models are typically measured by self-report questionnaires ("I am good at recognizing emotions in others") — raising methodological concerns, as people's self-assessments of emotional ability correlate poorly with actual performance on ability tests. Critics argue that mixed models conflate EI with personality factors already well-measured by established frameworks like the Big Five personality model, and lack discriminant validity.

What Does the Research Show?

Meta-analyses have sought to determine whether EI predicts important outcomes beyond what is already predicted by IQ and personality:

• Job performance: Ability-based EI predicts job performance with modest but significant incremental validity over cognitive ability and personality (Frank Schmidt and John Hunter's 2004 meta-analysis; Joseph and Newman's 2010 meta-analysis). Effect sizes are small to moderate and strongest for jobs with high emotional demands (e.g., sales, service, management).
• Leadership: Meta-analyses find positive relationships between EI and leadership effectiveness, particularly for transformational leadership behaviors, though effect sizes are moderate and the self-report vs. ability measure distinction matters significantly.
• Mental health and well-being: Higher EI is consistently associated with lower anxiety, depression, and alexithymia (difficulty identifying and describing emotions). This relationship is among the most robust in EI research.
• Relationship quality: EI predicts relationship satisfaction and social functioning, particularly emotion regulation aspects.

Goleman's original claim that EI accounts for 80% of life success (vs. 20% for IQ) has no empirical support and has been specifically criticized by Mayer, Salovey, and others as a dramatic overstatement. In controlled studies, variance explained by EI is considerably smaller, though real and meaningful.

EI vs. Personality

A persistent methodological concern in EI research is its overlap with established personality dimensions, particularly from the Big Five (extraversion, agreeableness, neuroticism, conscientiousness, openness). Self-report EI measures show substantial correlations with neuroticism (inversely) and agreeableness, raising the question: does EI add anything beyond what personality already captures?

Research using the MSCEIT ability measure consistently shows greater incremental validity (unique prediction beyond personality and IQ) than self-report measures, supporting the claim that ability-based EI is a distinct construct rather than a relabeling of known personality dimensions.

Emotional Intelligence in Practice

Several components of EI have particularly strong research backing for practical applications:

• Emotion recognition from faces and voices: Can be improved with targeted training; important for medical professionals, therapists, and negotiators
• Reappraisal (cognitive reappraisal of emotional events): Reinterpreting the meaning of an upsetting event to change its emotional impact — an evidence-based strategy for emotion regulation with strong support from James Gross's research at Stanford
• Emotional labeling: Naming emotions in words reduces amygdala activation (Matthew Lieberman, UCLA) — supporting the therapeutic value of emotional vocabulary
• Empathic accuracy: The ability to infer others' emotional states; teachable and improves with deliberate practice

The growing industry of EI training in workplaces has produced mixed results in rigorous evaluations. The most effective programs focus on specific, trainable competencies (emotion recognition, reappraisal techniques) rather than EI broadly, and use practice and feedback rather than simply providing information.