Abstract

The ability model of emotional intelligence (EI) specifies that four related abilities are involved: perceiving emotions, facilitating thought using emotions, understanding emotions, and managing them. Several performance-based assessments have been developed to measure those four abilities. Although some researchers find empirical support for the four abilities, others have argued that emotional intelligence divides into three abilities, two or even a single, unitary ability (Legree et al., 2014; Palmer, Gignac, Manocha, & Stough, 2005). We reanalyzed archival data from four ability tests of emotional intelligence, Ns = 503, 5000, 1000, and 2000, conducting item-level exploratory factor models of all four assessments for the first time. Based on those analyses, we suggest possible revisions of the 4-factor model to guide future research and assessment.

Overview of the present studies

In our research here, we analyzed the four assessments, the MEIS, MSCEIT, MSCEIT YRV, and MSCEIT-2, all at the level of their individual items—allowing their items to form factors with one another freely and apart from the tasks to which they had been assigned. Our purpose was forward-looking and focused on the question, “If one were to take a fresh look, what would be the nature and number of abilities one might ‘see’ in existing tests?”

Conclusions and implications for assessment

Theory is a powerful tool in the sciences but it is also important to adjust theories as new empirical findings improve our understanding. The theory of emotional intelligence has been associated with two key ideas: That emotional intelligence can be assessed as a mental ability using intelligence tests, and further, that the intelligence divides into four more specific abilities: accuracy at perceiving emotion, facilitating thought, understanding feelings, and emotion management. Of the first claim—that EI is an intelligence and can be assessed with intelligence tests—there is little doubt: Research from the MEIS forward demonstrates that the four-area model of emotional intelligence can be operationalized as a series of ability-based assessments (MacCann et al., 2014).

The research here, however, indicates that the four-area model could benefit from some reconsideration. The findings indicate the robustness of scores reflecting Understanding and Managing abilities, but that the Emotion Perception area may be more multi-dimensional than the original theory indicated, and that a “Connecting Emotional Features” score may better represent certain tasks than the earlier-conceived “Facilitating Thought” area. The findings make room, as well, for as-of-yet unclassified tasks (see Fig. 1). Collectively, the results encourage us to rethink EI in ways that have plain implications for future measures.

To the degree such new understanding closer approximates the reality of mental abilities in this area, it can inform applied practice so that the consultant and the clinician can provide people with more accurate feedback as to their abilities. A better understanding, such as this research provides, can help ensure that test scores convey distinctive meanings that are better understood than before—a key contributor to meeting the standards of optimal assessment (Joint_Committee, 2014; Sinharay et al., 2011). It is our hope that the findings obtained here, along with further research in the area, will help promote a better understanding of the nature of the ability model of emotional intelligence and its application.