Human resource practitioners place value on selecting and training a more emotionally intelligent workforce. Despite this, research has yet to systematically investigate whether emotional intelligence can in fact be trained. This study addresses this question by conducting a meta-analysis to assess the effect of training on emotional intelligence, and whether effects are moderated by substantive and methodological moderators. We identified a total of 58 published and unpublished studies that included an emotional intelligence training program using either a pre-post or treatment-control design. We calculated Cohen's d to estimate the effect of formal training on emotional intelligence scores. The results showed a moderate positive effect for training, regardless of design. Effect sizes were larger for published studies than dissertations. Effect sizes were relatively robust over gender of participants, and type of EI measure (ability v. mixedmodel). Further, our effect sizes are in line with other meta-analytic studies of competency-based training programs. Implications for practice and future research on EI training are discussed.