Emotional intelligence: what it is, how it's measured, and how to train it

Everyone knows someone who is clearly book-smart and completely lost the moment a conversation turns emotional. And everyone knows the opposite too: people who never got great grades but always seem to know what to say when someone is struggling, or when it's time to say nothing at all. That gap is what the concept of emotional intelligence tries to explain.

The term has become so popular that today it gets used for almost anything, from justifying bad management decisions to selling weekend courses. It's worth separating what serious research actually shows from what got left behind as marketing.

Where the concept comes from

Emotional intelligence entered scientific psychology in 1990, in a paper by researchers Peter Salovey and John Mayer. The core idea was simple: beyond the intelligence that IQ tests measure, there is a specific capacity to process emotional information, your own and other people's, and to use that information to guide thought and behaviour.

The concept only reached the general public five years later, with the book Emotional Intelligence by psychologist and journalist Daniel Goleman, published in 1995. Goleman defined it as the capacity to recognise our own feelings and those of others, to motivate ourselves, and to manage emotions well in ourselves and in our relationships.

From there, the field split into two paths that remain separate to this day:

• The ability model, from Mayer and Salovey, which treats emotional intelligence as a measurable cognitive skill, similar to IQ.
• The mixed model, popularised by Goleman and by Reuven Bar-On, which bundles that skill together with personality traits and motivation, such as optimism, assertiveness or self-drive.

This split isn't just academic. It has direct consequences for how emotional intelligence gets measured, which is why different tests, on the same person, can produce very different readings.

The four-branch model

The most scientifically solid foundation is still Mayer and Salovey's original model, formalised in 1997, which organises emotional intelligence into four hierarchical abilities.

Notice that only the last branch, managing emotions, matches what most people picture when they hear "emotional intelligence". The other three are the foundation without which emotional management turns into guesswork. You can't manage well something you haven't yet noticed you're feeling.

How it's measured, and why the tests disagree with each other

There are now more than thirty different instruments for measuring emotional intelligence, split into three families:

1. Performance tests, such as the MSCEIT (Mayer-Salovey-Caruso Emotional Intelligence Test), which ask the person to solve concrete emotional problems, for instance identifying the emotion in a photograph or picking the most effective response to a conflict situation. There are right and wrong answers, just like in an IQ test.
2. Self-report questionnaires, such as Bar-On's EQ-i or the TEIQue, which ask the person to rate themselves on statements like "I handle frustration well".
3. Mixed competency models, used mostly in corporate settings, which combine self-assessment with personality and leadership traits.

The problem is that these three approaches don't measure exactly the same thing, and a person's results on a performance test and on a self-report questionnaire correlate only weakly with each other. One of the most cited findings in the literature is that nearly two-thirds of the variance in EQ-i scores is explained by personality (Big Five traits), emotional well-being and IQ, while for the MSCEIT those same three variables explain only about 14% of the variance. In other words, the self-report questionnaire is much closer to measuring personality than the performance test, which captures something more distinct.

This has a simple practical implication: if you take an emotional intelligence test online, the type of test matters as much as the score. A ten-minute self-rating questionnaire mostly tells you how you see yourself, not necessarily how you actually perform when facing a real emotional dilemma.

If the reliability of these tests interests you, our article on psychological personality tests explains the criteria that separate a valid test from an entertainment quiz, and they apply here in exactly the same way.

What emotional intelligence predicts, and what it doesn't

A landmark meta-analysis led by Ernest O'Boyle and colleagues, published in 2011 in the Journal of Organizational Behavior, pooled dozens of studies on emotional intelligence and job performance. The correlation found ranged from about 0.24 to 0.30, depending on the type of measure used. Source: O'Boyle et al., "The relation between emotional intelligence and job performance: A meta-analysis", 2011

That's a real correlation, but a modest one, and here's the detail that tends to disappear when the topic gets sold in corporate training: a substantial part of that correlation is explained by overlap with IQ and with already well-known personality traits, especially conscientiousness and emotional stability. Once those factors are statistically controlled for, emotional intelligence's unique contribution to predicting performance shrinks considerably.

This is also the most serious criticism the scientific community levels at the concept: the risk of it becoming a catch-all bucket for anything that isn't IQ or personality, while in practice overlapping heavily with both. That's not a reason to discard the concept, but it is a reason to be sceptical of anyone who promises that emotional intelligence, on its own, predicts career success, a healthy relationship, or a balanced life.

What can be said with more confidence is more modest, and still useful: the four abilities in the original model, perceiving, using, understanding and managing emotions, reliably help in situations involving conflict, negotiation and stress regulation, even if they aren't the whole explanation for anything.

Yes, it can be trained

The good news, and perhaps the reason it's still worth talking about this topic, is that emotional intelligence isn't a fixed trait like height. The evidence shows it can be developed through deliberate practice.

A meta-analysis by Victoria Mattingly and Kurt Kraiger, published in 2019 in the Human Resource Management Review, examined 58 studies involving formal emotional intelligence training programmes. It found a positive, moderate effect of training on emotional intelligence test scores, consistent regardless of participants' gender and the type of test used to measure progress. Source: Mattingly & Kraiger, "Can emotional intelligence be trained? A meta-analytical investigation", 2019

The duration and type of programme made a difference: longer, more structured training, with practice spaced out over time, tended to produce bigger gains than one-off sessions. In other words, this doesn't get solved with a single afternoon workshop, it gets solved with habit.

Concrete techniques, for each of the four abilities

Perceiving emotions

The simplest, most studied exercise is the emotion diary: at the end of the day, write down two or three moments and name the emotion felt in each one, as precisely as possible. Not "I felt bad", but "I felt excluded" or "I felt rushed, and that made me irritated". The precision of your emotional vocabulary, what researchers call emotional granularity, is itself a trainable skill.

Using emotions to facilitate thought

Before a task that requires creativity, notice your mood instead of ignoring it. Lighter emotional states tend to favour associative thinking and new ideas; more focused states favour analytical, detail-oriented tasks. Matching the right task to the right state, whenever your schedule allows it, is a practical way to use this ability.

Understanding emotions

When you feel a difficult emotion, ask yourself: what exactly triggered it, is this the first time I've felt this in this situation, and what usually happens next. Understanding the pattern, not just the isolated episode, is what separates reacting from anticipating.

Managing emotions

This is the ability that benefits most from concrete techniques, and psychology has a reference model for it, developed by researcher James Gross. His process model of emotion regulation identifies five points where you can intervene, from the most preventive to the most late-stage:

1. Situation selection: avoiding or seeking out situations, knowing in advance how they tend to affect you.
2. Situation modification: changing something about the environment before the emotion sets in (for example, changing the subject of a heating-up conversation).
3. Attentional deployment: redirecting your focus to a different aspect of the situation.
4. Cognitive change: reinterpreting the meaning of what's happening.
5. Response modulation: intervening after the emotion is already present, for instance through breathing or physical exercise.

A simple response-modulation technique, well supported by research, is naming the emotion out loud or in writing as soon as you feel it. Giving it a specific name activates brain regions linked to regulation and reduces the intensity of the automatic alarm response. It doesn't solve whatever caused the emotion, but it buys you a few precious seconds of distance before you react.

The limits worth taking seriously

Emotional intelligence isn't a magic variable that explains everything else. It doesn't replace therapy when there are real, persistent difficulties with emotional regulation, it isn't synonymous with being a morally "good" or "empathetic" person, and its overlap with personality means part of what the tests capture was already there before any training.

That said, what's left over after accounting for all that overlap is still real, and it's trainable with consistent practice. Understanding better what you feel, what other people feel, and what to do with that information, is worthwhile in itself, regardless of what any test measures.

This is also one of the threads that connects emotional intelligence to other self-knowledge topics we've explored, like the Big Five personality traits or attachment styles in relationships. No single one of these maps describes you completely on its own. Together, they start to build a more honest picture of how you work, alone and with others.

That's also why Agapone doesn't treat self-knowledge as a single, isolated test. How you perceive, use, understand and manage emotions feeds into your overall picture, alongside other dimensions, precisely because none of them, taken alone, is enough to explain who you are.