Does Emotional Intelligence Predict Success? What the Evidence Says

In the 1990s emotional intelligence went from an academic idea to a business slogan in a few years. Popular books suggested it might matter as much as, or more than, ordinary intelligence for getting on in life. Training courses, hiring tools and school programmes followed.

Researchers were left with a narrower and more testable question: if you measure emotional intelligence properly, does it predict how people do at work and at school, beyond what we can already predict from intelligence and personality? After three decades of studies, there is a reasonably clear answer. It is yes, a little.

Two different things share one name

The first obstacle is that emotional intelligence, usually shortened to EI, refers to at least two different things.

The original version, developed by Peter Salovey and John Mayer, treats EI as a set of abilities: perceiving emotions accurately, using them to help thinking, understanding how emotions work and change, and managing them in yourself and others. If EI is an ability, it should be measured like one, with questions that have better and worse answers. That is what the ability test built by Mayer, Salovey and David Caruso tries to do. In a 2003 study, 21 emotion researchers largely agreed with a standardisation sample of 2,112 people on which answers were best, which gave some confidence that the items were not arbitrary.

The second version treats EI as a trait, or as a mix of traits: how emotionally capable you believe you are, along with qualities such as optimism, self-esteem and sociability. This is usually measured with self-report questionnaires. In 2001 Konstantinos Petrides and Adrian Furnham argued that this kind of EI is best understood as part of personality, and their analyses located it among established personality factors.

In 2008 Mayer, Salovey and Caruso wrote that a schism had developed between these camps, and argued that much of what was sold as EI was really an eclectic mix of positive traits. For anyone reading the research, the practical lesson is simple: always ask which kind of EI a study measured, because the results differ.

Does it predict performance at work?

The most cited work evidence comes from two meta-analyses published a year apart.

In 2010 Dana Joseph and Daniel Newman proposed and tested a model in which the ability branches build on one another: perceiving emotions comes first, then understanding them, then regulating them, which finally affects job performance. They also showed how much these branches overlap with things we already measure, namely conscientiousness, cognitive ability and emotional stability. Their most interesting finding was that ability EI did not predict performance consistently. It was positively related to performance in jobs with high emotional labour, such as roles that require managing your feelings in front of customers, and negatively related in jobs with little emotional labour.

In 2011 Ernest O'Boyle and colleagues, with a larger set of studies, reported corrected correlations between EI and job performance of roughly 0.24 to 0.30 across the three main kinds of measure. They concluded that EI adds something beyond intelligence and the Big Five, especially for self-report and mixed measures.

Those numbers deserve context. Correlations around 0.25 are meaningful in psychology, but they describe trends across many people, not destiny for any individual. Self-report measures also share a problem: the person who rates their own emotional skills highly may be the same person who rates their own performance highly, which can inflate associations when performance is also self-reported.

Does it predict grades?

The largest review to date comes from school and university. In 2020 Carolyn MacCann and colleagues combined data from 42,529 students. The overall association between EI and academic performance was about 0.20. It was strongest for ability EI (0.24) and weakest for self-rated EI (0.12).

Then came the crucial step. After controlling for intelligence and the Big Five, ability EI explained an extra 1.7% of the variation in performance, self-rated EI 0.7% and mixed EI 2.3%. The branches of understanding and managing emotions did a little better, at 3.9% and 3.6%. Even so, EI came out as the third most important predictor, after intelligence and conscientiousness. The authors also noted that the data could not show causation: better students might develop emotional skills, rather than the other way round.

So, does it predict success?

Put together, the evidence supports a modest claim. EI, especially the ability of understanding and managing emotions, is related to doing well at work and in education. Part of that link comes from its overlap with intelligence and personality, and part appears to be its own. It does not support the claim that EI matters more than intelligence, and it does not show that EI training reliably causes better outcomes.

That is less exciting than the slogan, but it is more useful. It means emotional skills are one ingredient among several, and that they matter more in some settings than others.

What this means for you

• Know what your score measures. Our emotional intelligence test is organised around the four branches of the ability model, but it is a self-report questionnaire. It tells you how you see your own emotional skills, which is valuable, and it is not the same as a performance test.
• Look at the branches, not just the total. In the research, understanding and managing emotions carried more weight than the overall score.
• Consider your setting. If your work or studies involve a lot of emotional labour, these skills are more likely to make a difference.
• Round out the picture. Our social intelligence test looks at how you read social situations, and our empathy type test separates understanding what others feel from sharing it.

Emotional skills are real and they are worth developing. They are not a secret formula, and anyone selling them as one is ahead of the evidence.