Abstract
Emotional intelligence has been considered an important construct by schools and universities because of its theoretical importance and practical implications. Considerable resources and time have been spent by the educational institutions to develop the emotional skills of their students. The present meta-analysis aimed at studying the relationship of emotional intelligence, including its three theoretical models, with academic performance while accounting for age and gender as moderators. "Robumeta" package was used for the meta-analysis of multilevel random effects with robust variance estimation (RVE) in R version 4.0.3. Effect sizes were calculated and meta-regression analysis with RVE was used to assess the relationship with the moderator variables. A positive and significant overall relationship was found between emotional intelligence and academic performance (ρ = 0.19). The moderating effect of emotional intelligence with gender streams on emotional intelligence and academic performance's relationship was examined through meta-regression with robust variance estimation and sub-group analyses. Whereas the relationship between emotional intelligence and academic performance was not found to be moderated by age, it was found to be partially mediated by gender.