Measuring Emotional Intelligence with the MSCEIT 2: Theory, Rationale, and Initial Findings

John D. Mayer ^1* David R. Caruso ² Peter Salovey ³ Iris Y. Lin ⁴ Braden J. Hansma ⁴ Joanna Solomon ⁴ Gill Sitarenios ⁴ Manolo Romero Escobar ⁴

• ¹ University of New Hampshire, Durham, United States
• ² Yale College, Yale University, New Haven, Connecticut, United States
• ³ Yale University, New Haven, Connecticut, United States
• ⁴ Multi-Health Systems, Toronto, Canada

The model of emotional intelligence as an ability has evolved in response to theoretical discussion and new research over the past 35 years. The revised model includes that emotional intelligence (EI) is a broad intelligence within the Cattell-Horn-Carroll (CHC) model of intelligence, and that more areas of problem solving are involved than originally detailed. An argument is made here that veridical scoring of EI is a sound procedure relative to scoring keys based on expert consensus or a single emotion theory. If EI fits present-day theories of intelligence (i.e., the CHC model), any subsidiary factor-based scales also should be highly correlated. These considerations led to a need to revise of the original Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT). The new MSCEIT 2 test was developed and tested across 5 studies: Two preliminary studies concerned the viability of new item sets (Study 1, N = 43) and enlisting the guidance of Ph.D. area experts to develop a veridical scoring key for each item (Study 2, N = 8). Next, a pilot study (N =523) and a normative study (N = 3000) each focused on the test's factor structure including whether a four-domain model continued to fit the data in a manner consistent with a cohesive broad intelligence. A fifth study (N = 221) examined the relation between the original and revised tests. The MSCEIT 2 is 33% shorter than the original, provides factor-supported subscale scores, and has good reliability at the overall level, with acceptable reliabilities for 3 of the 4 subscale scores, and adequate measurement precision across the range of most test-takers' abilities.