Do EI Competencies Catalyze the Relationship between Intelligence and Learning Performance? (WITHDRAWN)

Abstract

The fundamental principle of emotional intelligence (EI) lies on the integration of emotion into cognitive processes that facilitate mental activities such as thought, memory and decision- making. Yet, prior research on EI has mainly focused on testing additive models that assume EI and cognitive abilities make independent contributions to performance; little attention has been devoted to investigating their potential interaction effects. We develop and test a task- dependent interaction model that reconciles the divergent findings in previous interaction studies. We posit that whenever tasks require the social cognitive domain and involve interpersonal interactions, EI and general cognitive ability (or g) are mutually reinforcing, such that the association between EI and g becomes stronger as g increases. Otherwise, in non- social tasks, a negative interaction between EI and g is expected. Using a behavioral approach to EI, also known as EI competencies, we collect observations of actual EI behavior as is seen by others who work and live with the person. Based on a sample of 849 MBA students including 23 nationalities, we test a hierarchical model that is contingent on the types of task (social vs. non- social) and rater (professional, personal and self). Our results reveal that aside a positive main effect on the classroom performance of professional executives, behavioral EI moderates the relationship of g and academic performance. Whereas we find evidence that in non-social tasks, behavioral EI has a stronger effect on MBA performance among candidates with low g, our data was short to support the hypothesis whereby, in social tasks, EI boosts the effect of cognitive abilities on performance. Moreover, while females had an advantage in social tasks, man had relatively higher performances in non- social tasks. Professional raters had a relatively small measurement error as compared to the other rater types, and thus produced the highest estimates in our model.