Emotional intelligence: what the science actually says

“Emotional intelligence” is one of the few psychology terms that almost everyone has heard of. It shows up in job ads, parenting books and team-building workshops, usually shortened to “EQ”. But the popular version and the research version are not quite the same thing.

This guide walks through where the idea came from, the different ways researchers define and measure it, what the evidence says it predicts, and what a free online EQ test, including ours, can and cannot tell you.

Where the idea came from

The term entered academic psychology in 1990, when Peter Salovey and John Mayer proposed emotional intelligence as a set of skills: appraising and expressing emotion in yourself and others, regulating emotion, and using feelings to guide thinking and action (Salovey & Mayer, 1990). Their framing was modest and specific. Emotional intelligence was meant to sit alongside other kinds of intelligence, as a way of reasoning about emotional information.

Five years later, the science journalist Daniel Goleman turned the idea into a bestseller. His book’s subtitle claimed emotional intelligence could “matter more than IQ”, and it bundled together a broad set of qualities, including self-awareness, impulse control, persistence, motivation, empathy and social skill (Goleman, 1995).

The book did a lot to put emotions on the agenda in schools and workplaces. It also set off a wave of overclaiming. Popular articles and training programs began presenting EQ as the hidden key to success, often going well beyond what the research showed. Much of the scientific work since then has been an effort to pin down what emotional intelligence actually is, and how much it really adds.

Three ways to define emotional intelligence

Researchers broadly split into three camps. Knowing which one a test belongs to tells you a lot about what its score means.

Ability emotional intelligence

The ability approach stays closest to Salovey and Mayer’s original idea. Emotional intelligence is treated like a cognitive skill, measured with problems that have better and worse answers, much like an IQ test.

The best-known example is the MSCEIT (Mayer–Salovey–Caruso Emotional Intelligence Test). It covers four areas: perceiving emotions, using emotions to help thinking, understanding emotions, and managing emotions. Tasks include judging which emotions are present in faces or pictures, understanding how emotions blend and change, and rating how well different actions would work in emotional situations.

Because there is no single objective answer key for emotions, responses are scored against the consensus of a large sample or of emotion experts. In the test’s main validation study, these two scoring methods largely agreed, and the test showed reasonable reliability (Mayer et al., 2003).

Trait emotional intelligence

The trait approach, developed mainly by K. V. Petrides and Adrian Furnham, takes a different view. Instead of measuring emotional skill with right-or-wrong problems, it measures how people see their own emotional abilities and tendencies, through self-report questionnaires (Petrides & Furnham, 2001).

Petrides and Furnham drew together the facets that appeared across early EI models, including emotion perception, emotion regulation, empathy, impulse control, optimism, self-esteem and social competence. They argued that these self-perceptions are best understood as part of personality rather than as a form of intelligence. In this view, trait EI is a set of personality characteristics about emotion, closely linked to traits like emotional stability and extraversion (Petrides & Furnham, 2001).

Mixed models

“Mixed” models, including Goleman’s, combine emotional skills with a wide range of other qualities, such as motivation, optimism and social skill, usually measured by self-report. They are popular in workplace settings. Critics point out that because they mix so much together, it is hard to say what exactly is being measured, or how it differs from well-established personality traits (Joseph & Newman, 2010).

What emotional intelligence does and does not predict

The most useful evidence comes from meta-analyses, which pool results from many studies.

Job performance: a real but debated link

Two large meta-analyses looked at emotional intelligence and job performance. O’Boyle and colleagues found that EI was positively related to job performance, and that self-report and mixed measures added some prediction even after accounting for general cognitive ability and the Big Five personality traits (O’Boyle et al., 2011).

Joseph and Newman reached a more nuanced conclusion. For ability EI, the link with performance depended on the job. It was positive in jobs with high “emotional labor”, where managing feelings is part of the work, such as customer service or care roles. In jobs with low emotional labor, the link was weak or even negative (Joseph & Newman, 2010). They also proposed a “cascading” model: perceiving emotions helps you understand them, understanding helps you regulate them, and regulation is what connects most directly to performance.

So where does that leave the claim that EQ matters more than IQ? Neither meta-analysis supports it. A fair summary is that emotional intelligence has a measurable relationship with how people do at work, that the size of that relationship depends on how EI is measured and on the type of job, and that a good part of what self-report EI questionnaires capture overlaps with personality traits we already knew about.

Well-being and relationships

The clearest links between emotional skills and everyday life come from research on emotion regulation, which is the “managing” part of emotional intelligence. That research is worth looking at on its own.

Emotion regulation: reappraisal versus suppression

Emotion regulation means the ways people influence which emotions they have, when, and how they experience and express them. James Gross and Oliver John compared two common strategies (Gross & John, 2003):

• Reappraisal: changing how you think about a situation to change its emotional impact. For example, seeing a rejected application as practice for the next one, or reading a stranger’s rudeness as a sign of their bad day rather than a judgment of you.
• Suppression: holding back the outward expression of a feeling while it is happening, such as keeping a straight face while angry.

Across five studies, people who habitually used reappraisal reported more positive emotion, less negative emotion, better relationships and higher well-being. People who habitually used suppression showed the opposite pattern (Gross & John, 2003).

This does not mean suppression is always wrong. Sometimes keeping a calm face is exactly what a situation needs. But as a default habit, rethinking a situation tends to work better than bottling it up. This idea is built into the situational part of our test, described below.

Can emotional intelligence be improved?

This is where popular claims tend to run ahead of the evidence, so it is worth being careful.

A meta-analysis by Mattingly and Kraiger pooled 58 published and unpublished studies of emotional intelligence training. Overall, training had a moderate positive effect on EI scores, and the effect held across both ability-based and mixed measures (Mattingly & Kraiger, 2019).

There are important caveats:

• Published studies showed larger effects than unpublished dissertations, which suggests the published literature may paint a somewhat rosier picture.
• Many studies did not describe their training in enough detail to tell which ingredients mattered.
• Higher scores on an EI test after a course are not the same as lasting changes in behavior.

A reasonable reading is that some emotional skills, especially recognizing and naming feelings and practicing strategies like reappraisal, can probably be strengthened with practice and feedback. Claims that a short course will transform your EQ should be treated with caution.

The limits of self-report

Most free EQ tests, ours included, rely mainly on self-report: you rate statements about yourself. That has real strengths. It is quick, and it captures how you see your own emotional life, which matters in its own right. But it has known limits.

Self-flattery. People are not always accurate judges of their own emotional skills. Someone who thinks they read others well may not, and someone who worries they are cold may be warmer than they think.

Social desirability. Some answers simply sound better. When people have a reason to look good, self-report EI scores can shift a lot. In one experiment, people completed a self-report EI questionnaire and an ability EI test twice, once as if applying for a job and once honestly. The self-report measure was much easier to fake than the ability test (Day & Carroll, 2008).

Overlap with personality. Because trait EI is defined as self-perception, it shares a lot with personality measures such as emotional stability and extraversion (Petrides & Furnham, 2001). A high score partly reflects a generally positive, outgoing and calm disposition.

None of this makes self-report useless. It does mean a self-report EQ score is best read as “how I see myself right now”, not as an objective measurement.

How our emotional intelligence test works

Our free emotional intelligence test has two parts. All items are original; no items from the TEIQue, the MSCEIT or other published tests are reused.

Part 1: self-report (28 statements). You rate how much you agree with each statement on a 7-point scale, from “strongly disagree” to “strongly agree”. The statements are organized around the four broad factors of the trait EI model (Petrides & Furnham, 2001; Petrides, 2009), with seven statements each:

• Well-being: how positive you feel about your life and yourself.
• Self-control: managing strong feelings, urges and stress.
• Emotional awareness: recognizing and expressing your own and other people’s feelings.
• Sociability: feeling at ease socially, influencing a group’s mood and standing up for yourself.

Some statements are worded in reverse, so that agreeing with everything does not automatically push your score up.

Part 2: situations (8 scenarios). You read everyday situations, such as receiving harsh criticism or a friend repeatedly canceling plans, and choose the response you think would work best. Each option is scored from 0 to 3 based on expert judgment, in line with emotion-regulation research that generally favors rethinking a situation or addressing a problem calmly and directly (Gross & John, 2003). This part is reported separately, as a score out of 24.

Your results. You get:

• An EQ index from 0 to 100, based on your average across the 28 self-report statements.
• A percentile and a band (from “low” to “very high”), comparing your self-rating with other people’s. These use provisional norms, which are still being refined as more people take the test, so treat them as rough.
• A score for each of the four areas, plus notes on your relatively strongest area and room to grow.
• Your separate situations score.

Keep the limits above in mind. The EQ index reflects how you see yourself, which is not always how others see you. The situations part checks whether you can recognize effective responses, which is not the same as using them in the heat of the moment. The test is a tool for self-reflection, not an assessment for hiring, clinical or other high-stakes decisions.

Frequently asked questions

Is a free online EQ test accurate?

It can give you a useful snapshot of how you see your own emotional life, but it is not a precise measurement. Self-report scores are affected by mood, self-image and the wish to look good (Day & Carroll, 2008). For the most honest result, answer based on how you usually act, not how you would like to be.

Is EQ more important than IQ?

The research does not support that claim. Meta-analyses find that emotional intelligence relates to job performance, but the strength of that link depends on the job and on how EI is measured, and much of it overlaps with personality (Joseph & Newman, 2010; O’Boyle et al., 2011). Both matter, in different ways.

Can I raise my EQ?

Training studies suggest some improvement is possible, with a moderate average effect on test scores (Mattingly & Kraiger, 2019). Practical starting points include putting precise words to your feelings and practicing reappraisal, which is linked with better well-being than suppression (Gross & John, 2003). Expect gradual change, not a quick fix.

What is the difference between ability and trait emotional intelligence?

Ability EI is measured like an intelligence test, with problems that have better and worse answers (Mayer et al., 2003). Trait EI is measured with self-report and reflects how you perceive your own emotional tendencies, which places it closer to personality (Petrides & Furnham, 2001). Our test is mainly a trait EI measure, with a short situational section added.