Is Emotional Intelligence a Skill? What a 78,000-Person Meta-Analysis Shows

Is emotional intelligence a skill, or is it a fixed trait some people are simply born with? A 2022 meta-analysis in Frontiers in Psychology, pooling 253 effect sizes from 78,159 participants, treats it as the former. Across the pooled research, people who scored higher on emotional intelligence, however it was measured, also showed higher job performance, higher job satisfaction, and lower job stress. That pattern does not prove any one reader can develop the trait at the same speed as anyone else. It is evidence that the underlying thing behaves like a measurable skill, not a quantity fixed at birth.

Why "high EQ" means nothing on a CV

Emotional intelligence sits on almost every soft-skills list, and it carries the same problem as the rest of that list: nobody rates their own emotional intelligence as low, so the self-description tells a hiring manager nothing. What the research does differently is separate the label from three distinct ways of actually measuring it: ability-based tests scored like a cognitive exam, self-report questionnaires, and mixed models that blend emotional skill with broader traits like optimism and motivation. A claim is free. A tested score is not, which is the same distinction this site keeps drawing between a resume line and a checkable result.

What the meta-analysis actually found

The pooled correlation between emotional intelligence and job performance came out at ρ = 0.30 across all three measurement streams, with ability-based testing at 0.28, self-report at 0.33, and mixed models at 0.31.

The same meta-analysis found emotional intelligence correlated with job satisfaction at ρ = 0.29, with organizational commitment at ρ = 0.26, and negatively with job stress at ρ = -0.43, the strongest single relationship in the study. None of these numbers say a specific reader's next role hinges on the trait. They say that across tens of thousands of people, the pattern shows up consistently enough to stop treating the whole idea as unmeasurable.

Whether it can actually be trained

A trait fixed at birth would not respond much to training. A 2019 meta-analysis in Human Resource Management Review, reviewing 58 published and unpublished studies that trained emotional intelligence through a pre-post or treatment-control design, found a moderate positive training effect, and one that held up regardless of the participants' gender or whether the study used an ability-based or mixed-model measure of emotional intelligence. Effect sizes ran larger in published studies than in unpublished dissertations, a normal pattern in this kind of research and a reason to read the finding as real but not overstated. Put the two studies together and the honest read is a skill that starts unevenly distributed and moves with practice, closer to public speaking than to height.

Where this fits the six domains this site measures

The six non-technical skill domains this site checks treat emotional intelligence as behaviour, not adjective: reading a colleague's frustration before it derails a meeting, adjusting a message once it lands badly, staying steady when a plan changes. The case this site makes about listening is one narrow slice of the same territory, a specific behaviour with its own research base rather than a synonym for the broader trait. Emotional intelligence covers more ground, and the research above says the ground is real enough to score.

None of this makes the trait easy to fake past a trained rater. It means the opposite: it is one of the few soft-skill words with a real, checkable pattern of behaviour underneath it, the same standard this site applies everywhere else.