Three emotional intelligence models emerged in the 1990s, each claiming to measure and develop EQ. Daniel Goleman's mixed model (1995) connected emotional competencies to workplace performance, earning him the term's popularisation. Reuven Bar-On's EQ-i (1997) treated emotional intelligence as personality-like traits predicting wellbeing. Mayer, Salovey and Caruso's ability model positioned EQ as a cognitive skill set — like IQ, measurable through problem-solving tasks with objectively correct answers. Today they coexist, each with different research support and practical applications. This guide compares their origins, conceptual differences, measurement approaches, and what the science actually supports.
The Birth of a Construct: Mayer and Salovey, 1990
The term "emotional intelligence" first appeared in print in 1990, in an academic journal article by psychologists Peter Salovey and John D. Mayer. They were investigating why some people handle emotions more effectively than others — not in the pop-psychology sense of "being nice," but as a genuine ability: perceiving emotions accurately, understanding their causes and consequences, and using that information to guide behaviour.
Salovey and Mayer's 1990 definition was precise: emotional intelligence is the ability to perceive emotions in oneself and others, to understand what those emotions mean, and to manage emotions effectively. They positioned it explicitly as an ability, analogous to cognitive intelligence — a skill that could be measured objectively, not merely reported by questionnaire.
Critically, they did not claim it was a personality trait (like the later Bar-On model) or a collection of workplace competencies (like Goleman's). It was a specific cognitive ability.
Goleman's Popularisation: The Mixed Model, 1995
In 1995, psychologist and science journalist Daniel Goleman published Emotional Intelligence: Why It Can Matter More Than IQ. The book became a bestseller and made the term household currency. But Goleman's concept was broader than Salovey and Mayer's original framework.
Goleman grouped emotional intelligence into five domains, later refined to four:
- Self-awareness — recognising your own emotions and their impact
- Self-management — regulating emotions and impulses, managing stress
- Social awareness — reading others' emotions, empathy, organisational awareness
- Relationship management — influencing, communicating, conflict resolution, teamwork
- Motivation — (in some versions) drive to achieve, resilience
Goleman's innovation was connecting these competencies explicitly to workplace outcomes — leadership effectiveness, sales performance, team cohesion. His book argued that emotional skills predicted job success as powerfully as IQ, a claim that resonated with organisations and earned him speaking engagements globally.
However, Goleman's model was "mixed" — it bundled abilities (perceiving emotion), traits (empathy, motivation), and learned skills (communication) together without clear boundaries. Critics noted this made it harder to validate separately from personality tests like the Big Five.
Bar-On's Commercial Framework: The EQ-i, 1997
In 1997, psychologist Reuven Bar-On published the Emotional Quotient Inventory (EQ-i), the first commercial emotional intelligence assessment. Bar-On's model was also mixed, but arranged differently.
Bar-On's framework had five branches:
- Intrapersonal EQ — self-awareness, assertiveness, independence, self-regard, self-actualisation
- Interpersonal EQ — empathy, social responsibility, relationship building
- Stress management — stress tolerance, impulse control
- Adaptability — flexibility, reality testing, problem-solving
- General mood — optimism, happiness
Unlike Goleman's emphasis on workplace performance, Bar-On positioned emotional intelligence as contributing to overall psychological wellbeing and life satisfaction. His EQ-i was a self-report questionnaire — respondents answered roughly 130 questions about themselves on a scale, and their replies were scored into subscales.
The EQ-i became widely used in coaching, HR development, and clinical settings. Its main advantage was simplicity: measure yourself, get a score. Its main limitation: self-report bias — people often don't know (or won't admit) their true emotional patterns.
The Ability Model Refined: Mayer-Salovey-Caruso, 2002
In 2002, Peter Salovey, John Mayer, and psychologist David Caruso published the Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT), operationalising their original 1990 concept into a validated measurement tool.
The MSCEIT is organised around four branches:
- Perceiving emotions — accurately identifying emotions in faces, pictures, music, and design
- Using emotions — leveraging emotional states to facilitate thinking and creativity
- Understanding emotions — comprehending emotional language, blends, and transitions
- Managing emotions — regulating emotions in oneself and others
The critical difference: the MSCEIT is an ability test. Test-takers solve problems with objectively correct or defensible answers. For example, one item shows a person's face and asks "What is this person feeling?" Responses are scored against expert consensus or objective criteria — there's a right answer (or a better and worse answer), not just a subjective self-report.
This mirrors how IQ tests work — you solve a problem, your answer is scored as correct/incorrect or better/worse, and your score reflects your demonstrated ability. Because of this structure, the MSCEIT correlates more closely with cognitive ability measures and shows better discriminant validity from personality than the mixed models do.
Comparison Table: The Three Models at a Glance
| Dimension | Goleman (1995) | Bar-On (1997) | Mayer-Salovey-Caruso (2002) |
|---|---|---|---|
| Model type | Mixed (competencies + traits) | Mixed (traits + wellbeing factors) | Pure ability |
| Core claim | Emotional competencies drive workplace success | Emotional-social functioning predicts wellbeing | Emotional perception and management are cognitive abilities |
| Number of components | 4-5 competencies | 5 branches | 4 branches |
| Assessment method | 360-degree feedback (ECI / ESCI) | Self-report questionnaire (EQ-i) | Objective ability test (MSCEIT) |
| Answer format | Behavioural observations | Self-ratings (5-point scale) | Multiple choice with correct answers |
| Primary use | Leadership development, workplace coaching | HR screening, career coaching, clinical assessment | Research, ability diagnosis, cognitive profiling |
| Key strength | Practical workplace relevance | Broad emotional-social coverage, scalable | Objective scoring, cognitive precision |
| Key limitation | Overlaps with Big Five personality; lacks objective answers | Self-report bias; conflates ability with personality and mood | Narrow scope (ability-only); doesn't measure motivation or wellbeing |
How Research Distinguishes Them
Academic research has compared these models extensively, particularly since the early 2000s. Several patterns emerge:
Discriminant validity from personality
The Mayer-Salovey-Caruso ability model correlates more weakly with Big Five personality traits (openness, conscientiousness, extraversion, agreeableness, neuroticism) than the mixed models do. This suggests it's measuring something conceptually distinct — a cognitive ability rather than a personality trait. Goleman's and Bar-On's models, by contrast, overlap substantially with Big Five measures, particularly with openness (understanding emotions), extraversion (social awareness), and agreeableness (empathy). This isn't a flaw, but it does mean they're partly measuring existing personality dimensions rather than a novel construct.
Criterion validity for workplace outcomes
Goleman's model shows stronger correlations with job performance, leadership ratings, and team effectiveness in meta-analyses. This is because his domains (self-management, relationship building, social awareness) directly reference workplace competencies. The MSCEIT shows weaker links to job performance, though it does correlate with performance in roles requiring emotional labour (counselling, negotiation). Bar-On's EQ-i shows mixed results — it predicts some workplace outcomes but is often outperformed by simpler Big Five measures.
Relationship to cognitive ability
The MSCEIT correlates with general cognitive ability (g) at around r = 0.3 to 0.4, similar to correlations between IQ and broad personality measures. Goleman's and Bar-On's models show negligible or weak correlations with IQ, supporting the claim that they're not simply measuring intelligence under a different name. However, this also means the MSCEIT may be partially measuring IQ-related reasoning rather than purely emotional processing.
Test-retest reliability
All three models show adequate reliability (r ≥ 0.70), with the MSCEIT generally highest. However, reliability alone doesn't settle whether the construct is real or useful.
Practical Implications: Which Model to Use When
The choice between models depends on your purpose:
For organisational development and coaching
Goleman's framework is the standard. The 360-degree feedback format (the ECI and ESCI assessments) fits naturally into leadership coaching and team dynamics work. Organisations know how to act on Goleman results — they map to specific behaviours and competencies that can be trained.
For HR screening and personality profiling
Bar-On's EQ-i is simpler to administer and score. It's useful as part of a broader personality assessment or when you want to measure emotional wellbeing alongside job fit. However, it performs no better than Big Five assessments for most hiring decisions, and adds little beyond an agreeableness or neuroticism scale.
For research and cognitive diagnosis
The MSCEIT is the most defensible academically. If you need objective ability measurement — for instance, in neuroscience research, clinical assessment of emotional processing deficits, or validating training effects — the MSCEIT's problem-solving format gives you ground truth. Its weakness is that it doesn't measure the "softer" dimensions of emotional wellbeing or motivation that the mixed models include.
For self-knowledge
Any of the three can prompt useful reflection. Goleman's framework is most actionable for behaviour change. Bar-On's is broadest and most accessible. The MSCEIT is most intellectually rigorous but offers less practical guidance for self-improvement.
What the Science Actually Shows
After 35 years and thousands of studies, emotional intelligence is confirmed to be a real construct — but a more modest one than the 1990s hype suggested.
What's established: Emotional perception (facial recognition, tone discrimination) is a measurable ability with moderate stability. People do vary in their capacity to regulate emotions, and that variation matters for wellbeing and workplace functioning. Emotional skills can be trained, particularly in high-stakes domains (negotiations, customer service, therapy).
What's disputed: Whether emotional intelligence is fundamentally different from personality (particularly agreeableness and openness) or a higher-level name for existing traits. Whether EQ predicts job success better than simpler measures like conscientiousness. Whether a single score called "EQ" is meaningful, or whether subscales matter far more than the total.
The consensus: Emotional intelligence is useful — but its predictive power is smaller than early advocates claimed. It's not a replacement for IQ or personality assessment; it's a useful addition. The Mayer-Salovey-Caruso model is the most scientifically defensible. Goleman's framework is the most practically useful for organisations. Bar-On's is the most accessible but offers the least new information beyond Big Five personality.
The practical takeaway: develop your emotional skills (self-awareness, regulation, empathy, communication) because they improve wellbeing and workplace relationships. But don't expect emotional intelligence alone to transform your career or solve deep personality patterns — personality traits and cognitive ability matter just as much, if not more.
To explore your own emotional intelligence across these dimensions, try our free EQ test, which measures emotional perception, understanding, and regulation against norms.
Frequently Asked Questions
What's the difference between Goleman's and Mayer-Salovey-Caruso's models?
Goleman treats emotional intelligence as a collection of learned competencies (self-awareness, self-management, empathy, communication) that drive workplace success. Mayer-Salovey-Caruso treats it as a cognitive ability — like spatial reasoning or verbal comprehension — that can be measured through problem-solving tasks with correct answers. Goleman's model is broader and more practical for coaching; the MSCEIT model is more academically rigorous and objective.
Is emotional intelligence measured by self-report or by ability tests?
Both. Bar-On's EQ-i and Goleman's 360-degree feedback rely on self- or observer-report (you describe your own behaviour or others rate you). The MSCEIT is an ability test where you solve problems and your answers are scored as correct or incorrect, like an IQ test. Ability tests are less prone to self-deception bias but measure only cognitive aspects of emotion-handling; self-report measures capture broader life experience and wellbeing but can be inflated by overconfidence.
Can emotional intelligence be trained?
Yes, particularly specific skills like active listening, conflict de-escalation, and emotional regulation. Coaching and deliberate practice improve these competencies. However, training effects are moderate — personality traits like introversion or neuroticism don't shift much with training, and some emotional patterns reflect deep temperament that's harder to change. The more technical the emotional skill (facial emotion recognition, tone discrimination), the more trainable it is.
Is emotional intelligence more important than IQ for success?
No. Goleman's 1995 book claimed it might be, but research doesn't support that. Cognitive ability (IQ) and personality (particularly conscientiousness) are stronger predictors of job performance and life outcomes than emotional intelligence. Emotional intelligence matters — it improves relationships and wellbeing — but it's one factor among many, not a replacement for ability or character.
Which emotional intelligence test should I take?
If you want actionable feedback for personal or professional development, Goleman's framework (via coaching or 360-degree feedback) is most useful. If you want a quick self-assessment of emotional wellbeing, Bar-On's EQ-i is straightforward. If you want an objective measure of emotional perception and reasoning ability, the MSCEIT is the gold standard — though it requires professional administration. For a free, web-based emotional intelligence profile, our EQ dashboard offers a middle ground: it measures key dimensions without requiring a psychologist's interpretation.