The relationship of trait EI with academic performance: A meta-analytic review

Year: 2012

Author: Perera, Harsha, Athanasou, James

Type of paper: Abstract refereed


Purpose. In the past decade, there has been considerable debate among scholars over whether trait emotional intelligence (trait EI) predicts academic performance (AP). Theoretically, trait EI may directly and indirectly, albeit modestly, influence AP; however, relations of trait EI with AP have been inconsistent across studies, raising the possibility that the observed relation may be moderated by methodological and theoretical factors. This meta-analysis aimed to assess the validity of trait EI for predicting AP and test moderators of the relation between the constructs.

Method. 74 effect sizes were drawn from 48 independent samples with a cumulative sample size of 10,292. Most analysed samples comprised tertiary students, but there were comparable aggregate samples from tertiary and secondary levels. Affective personality assessments were obtained in most of the primary samples via the Bar-On EQ-i family of instruments. The TEIQue and SEIS - and their variants - were also used to measure trait EI in multiple studies.

A meta-analysis with random-effects weights was used to estimate the inverse-variance weighted mean correlation between trait EI and AP. Homogeneity analysis was conducted to determine the significance and proportion of observed variance attributable to systematic between-samples variation. The moderating effects of gender, educational level, age, operationalisation vehicle, and transition status were tested using subgroup analysis and meta-regression.

Results. The meta-analysis revealed a modest weighted mean association between trait EI and AP ( unattenuated = .20). Homogeneity analysis revealed statistically significant and substantial heterogeneity in the effects, indicating that within-study error alone could not sufficiently explain the variability in the effects size estimates. Subgroup analysis showed that educational level moderated the correlation between trait EI and AP whereas gender, instrument and transition status had no moderating effects. Finally, meta-regression revealed a statistically significant moderating effect of age on the summary effect.

Conclusion. Consistent with trait EI theory, the meta-analysis yielded a favourable, modest validity for trait EI. The moderating effects of age and educational level, however, suggest that researchers must consider specific population characteristics in forwarding theoretical predictions about the effects of trait EI on AP. Notwithstanding the finding of significant moderation, much systematic variation between samples remains unexplained, raising the possibility that other moderators, such as general ability, level of stress exposure, subject area, and stability of measures may moderate this relationship. Limitations of this study, future directions and implications for education are discussed