EQ vs IQ at Work: What the Research Actually Shows About Emotional Intelligence and Performance

EQ vs IQ at Work: What the Research Actually Shows About Emotional Intelligence and Performance In 1995, Daniel Goleman published Emotional Intelligence, arguing that EQ — the ability to perceive, understand, manage, and use emotions — was a better predictor of life success than IQ. The book sold millions of copies and launched a multibilliondollar industry in emotional intelligence assessment and training. The claim that EQ matters more than IQ for workplace success has since been repeated in thousands of articles, corporate training programmes, and leadership development curricula. It has also been significantly overstated — and in some cases, misrepresented. Understanding what the research actually shows about emotional intelligence is important not because EQ is unimportant — it is genuinely important — but because the popular version of the research leads to interventions that are less effective than they could be. What the Research Actually Shows The most comprehensive metaanalysis of emotional intelligence research, published by Joseph and Newman in 2010 in the Journal of Applied Psychology, analysed 65 studies covering over 7,000 participants. The findings were nuanced. EQ was a significant predictor of job performance — but its predictive power varied substantially by job type. For jobs with high emotional labour demands — roles requiring significant interpersonal interaction, customer service, leadership, and teamwork — EQ was a strong predictor of performance. For jobs with low emotional labour demands — roles primarily involving technical or analytical work with limited interpersonal interaction — EQ added relatively little predictive power beyond IQ and conscientiousness. This finding is important because it challenges the universal claim that EQ matters more than IQ. For some jobs, in some contexts, EQ is a stronger predictor of performance than IQ. For others, IQ and specific technical skills remain the primary predictors. The relationship is not un