Emotional intelligence

Emotional intelligence is two things wearing one name, and almost every disagreement about whether it is real, measurable or useful comes from that.

The first is an ability: how accurately you can identify what an expression means, reason about how emotions blend and change, and use that information in a decision. Like any ability it is scored against a key — some answers are better than others — and it sits, statistically, close to the verbal end of general intelligence.

The second is a set of traits and self-beliefs: how emotionally steady, empathic, assertive and self-aware you report yourself to be. Scored by self-report, it sits close to the Big Five — so close that the interesting question is what is left once you subtract them.

These are not two measures of one thing. They correlate weakly, around r = .2 to .3, which is about what two genuinely different constructs look like. A person can score high on one and low on the other, and many do. Anyone who tells you what "your EQ" is without saying which of the two they measured has told you less than it sounds.

Peter Salovey and John Mayer coined the term in a 1990 journal article and meant the ability. Their model has four branches — perceiving emotion, using emotion to facilitate thought, understanding emotions, and managing them — and their instrument, the MSCEIT, is scored like a test with answers rather than like a questionnaire.

Then in 1995 Daniel Goleman published Emotional Intelligence: Why It Can Matter More Than IQ, which sold in the millions and is where nearly everyone met the idea. Goleman's version is much broader. It pulls in motivation, optimism, impulse control, empathy and social skill — a list that overlaps heavily with conscientiousness, emotional stability and extraversion — and it carried a headline claim, widely repeated as a number, that EQ accounts for far more of career success than IQ does. That claim did not come from the measurement literature, and Goleman has since distanced himself from the specific figures attached to his name.

A third strand arrived with K. V. Petrides and Adrian Furnham in 2001, who did something clarifying: they argued that self-reported EI is not a failed attempt to measure an ability at all, but a legitimate personality trait — "trait emotional intelligence" — and should be located inside the existing trait taxonomies rather than alongside cognitive ones. That reframing is now standard, and it is why researchers distinguish ability EI, trait EI, and the commercial hybrids usually called mixed EI.

So: same phrase, three research programmes, and a bestseller that made the term a household word while describing the version furthest from the original.

The zero-order numbers are less dramatic than the book and less dismal than the critics.

All three streams predict job performance, modestly. O'Boyle and colleagues (2011) estimated validities around .24 for performance-based ability EI, .30 for self-reported ability EI and .28 for mixed EI. Those are ordinary, respectable correlations in organisational psychology — not nothing, and not a rival to cognitive ability.

Ability EI matters more where the job is emotional. Joseph and Newman's 2010 meta-analysis found the relationship between EI and performance was considerably stronger in jobs with high emotional labour — nursing, sales, customer service, teaching — and close to nil in jobs with low emotional demands. This is one of the sturdier findings in the area and it is routinely dropped from training decks, because it implies the workshop is irrelevant to half the room.

Mixed EI predicts, but not for the reason it says. This is the sharpest result of the set. Joseph, Jin, Newman and O'Boyle (2015) asked why self-report EI questionnaires predict job performance so respectably, and decomposed them. The answer was that mixed EI measures are largely a compound of conscientiousness, emotional stability, extraversion, general self-efficacy and self-rated performance. Together those accounted for the bulk of the prediction. The measure works; it works as a mixture of things that already had names.

That is the difference between validity and incremental validity, and it is the whole argument. Does an EI score tell you anything you would not already know from a cognitive test and a Big Five profile? For mixed EI the honest answer is that the relationship falls to nil once those constructs are controlled, which is the finding of the paper itself rather than a critic's gloss on it. For ability EI the answer is: a little, reliably, in emotionally demanding work. Antonakis, Ashkanasy and Dasborough's 2009 paper is the most forceful version of the sceptical case, and their complaint is less that EI is fake than that the studies claiming it predicts leadership rarely ran the controls that would show it.

Almost every free online EQ test, including ours, is a self-report questionnaire. That places it in the trait and mixed camp, with the properties that follow.

It measures your beliefs about your emotional functioning. Not your functioning. The gap between the two is the entire subject of self-report questionnaires, and it is unusually wide here, because the people least able to read their own emotional states are the least equipped to report on that inability. Difficulty naming feelings has its own name — alexithymia — and the person with it does not necessarily know.

The items are transparent. I stay calm under pressure has an obviously approved answer. Scores drift upward in any setting where the result might be seen by someone else, which is the general problem of response styles in an unusually pure form, and it is why self-report EI is a poor basis for a hiring decision however good it is as a prompt for reflection.

A single EQ number hides the useful structure. The branch that most reliably predicts anything is understanding emotions — the vocabulary-and-inference branch — and it is not the branch people assume they are being scored on. A high overall score built on self-rated composure and a low one on emotional understanding is a common and actionable pattern that an average erases.

Our Workplace EQ test is worth taking for the section scores and worth reading with all of the above in mind: it is a structured self-assessment of workplace emotional habits, not a measurement of an ability, and the number at the top is the least informative thing on the page.

The commercial promise is that EQ, unlike IQ, can be raised. That is probably true in a narrow sense and the evidence is thinner than the promise.

The parts that do move under practice are specific and unglamorous. Emotional vocabulary moves: people who practise distinguishing irritated from anxious from depleted make better regulation choices, which is the finding behind emotional granularity. Noticing the body first moves; the entry on interoception covers why the signal usually arrives as a physical state before it arrives as a word. And the pause between feeling and acting is trainable in the ordinary way skills are, by rehearsal in low-stakes moments rather than by insight in high-stakes ones.

What does not obviously move is the underlying ability to read other people accurately, which behaves like most cognitive abilities: fairly stable, modestly improvable, not transformed by a day of training.

The practical version is narrower than "raise your EQ" and more useful. If your work is emotionally demanding, the ability side is worth real attention and the general trait score is mostly telling you about your personality, which you could have measured directly. If your work is not, the evidence says this is unlikely to be the lever you were told it is — and a specific problem, like how you behave in an argument you are losing, is better approached through the conflict style result than through a global EQ number.