To measure such a construct as EI, it necessarily flows from a clear definition of what you are seeking to record or measure. EI is a psychological construct and as such is associated with a number of challenges. Constructs are the focus of theory, lacking an ability to be touched and evidenced in a range of individual differences. Thus, it is not the construct that is measured or recorded but its effects. This ambiguity leaves it open to multiple interpretations, and definitions. As hinted at above, differing models of EI indicate different measurements. Hence, it is important to maintain integrity of approach in any research study that aligns definitions with appropriate measures of behaviour, or actions. Studies have also indicated that there is a lack of correspondence between the varying EI measures suggesting that, they may not in fact be measuring the same thing. The lack of coherence in measurement of EI has resulted in criticism. In determining the credibility of any measure, any method of measuring EI should arguably address a number of important factors: Reliability, Validity, and the norming population (Jensen 2007 cited in Jensen et
al, 2007). Reliability refers to the consistency of a test. A common way of
determining this is through the Test-Re-Test approach and that of Internal Consistency. The premise behind Validity is that any test actually measures what it is intended to measure. Jensen, (2007) states that with EI measures, discriminant, and concurrent and predictive validity are the most commonly used approaches.
The Mayer, Salovey, Caruso Emotional Intelligence Test [MSCEIT®] (2002) (V2.0) is the current testing tool for the Mayer & Salovey Four branch model. Mayer & Salovey, (1995) have adopted an objective, performance based
assessment of EI, which echoes the type of testing used to measure traditional intelligence (IQ) (Mayer et al 2002, Mayer, et al, 2001; 2003). Brody (2004) argued that to compare EI ability measures with traditional IQ measures is unsound as the latter is qualitatively different. That there are differing measures of EI linked to different conceptualisations of the construct with arguably weak convergence evidence across them (e.g. .21 between MSCEIT and EQ-i) could be construed as undermining the credibility of the EI construct. However, such differences may also be supportive of qualitative differences between ability and personality-based conceptions.
Based on an ability model of EI, the test resembles a typical IQ test. Using a selection of ‘right’ or ‘wrong’ answers to a range of questions, the MSCEIT serves to provide an indexation of each of the four branches of EI in addition to providing a total emotional intelligence quotient (EIQ) (Mayer & et al, 2012). Comprising eight individual tasks, two of which aim to measure each branch component of the model. For example, perceiving emotions is measured by means of asking individuals to identify range of emotions in faces and landscapes. Facilitating emotions is measured by asking individuals to choose which emotions promote certain kinds of thoughts and actions. Understanding emotions is measured through use of blending exercises- which two emotions fit together to a certain descriptor e.g. “ which two emotions together are closest to contempt: a) sadness and fear or b) anger and disgust. How an individual manages their own emotions and the emotions of others can be measured by using a series of vignettes outlining social situations and asking them how they manage their emotional responses. (Mayer et al, 2002).
Although it appears in the literature as one of the better scientific measures of EI (Murphy et al 2006), critics have highlighted concerns- particularly around the issue of scoring as framed by expert and consensus methods. For example, does scoring replicate social conformity rather than individualized social competency? There are then concerns with the MSCEIT for example:
• Correlations between the MSCEIT and other EI measures
In answering these concerns, Mayer et al, (2012) draw attention to measures taken to ensure both a theoretical and reliable development of the MSCEIT. For example, they draw attention to their use of theories of emotion reasoning, the use of 21 ‘experts from the International Society for Research on Emotion in addition to a normalising group of some 5000 test-takers. Mayer et al (2002) state that expert and consensus scoring is highly correlated e.g. r = .93-.99. Again, Mayer et al. (2003) argue that in terms of reliability there is a viable correlation between the exert and consensus scores across a range for both total and branch EI scores, i.e. r = .68 for consensus scoring and .71 for expert scores. It is perhaps of value to note that Matthews et al. (2002) point to the reliability coefficients cited in cognitive ability tests have a range of .85- .95. Brackett & Mayer, (2004) observed that internal consistency of the MSCEIT is acceptable.
Matthews, (2004, 2007) have pointed out the high levels of correlation amongst the high level or experiential EI elements of the MSCEIT with an r = >.90. Although, in this same article, the correlations amongst the sub-scales, or the Four EI Branches is found to be marginal i.e. <. 60. This discrepancy was observed following the MSCEIT scores were obtained for the study group at the start of the research process. A number of the participants had achieved a high overall EI score and yet individual branch scores could range from low to medium within the same overall high score award. Such a differential is potentially explained n the MSCEIT manual in regard to user incompetence, or the user was being asked to answer in one context (e.g. employment) but answered from another (e.g. domestic setting). I raise the point here for the benefit of the reader in order to be transparent as to the scoring process and a potential issue with how scores are operationalised. Brackett & Mayer, (2003), and Roberts et al (2010) argue that the overall EI construct is reasonable and reliable, although the sub-scales or Branch scores require further work to firmly establish reliability coefficients. However, the ability-based model of EI is plausible from a scientific perspective and is meaningfully practical and I would
add offers up a way of engaging with and understanding the adaptive, or wisdom function of EI. Mayer, et al (2012) do acknowledge in response to Maul, (2012:405) that whilst it the MSCEIT does not measure every single emotional intelligence skill. It does measure:
“…important skills within each of the theory’s four branches rather than to measure exhaustively.”
Mayer, et al. (2012:407) highlight that based on correlations between the MSCEIT and other similar ability based EI models:
“...the new findings…make the case that the MSCEIT correlates meaningfully with a variety of ability based criteria of EI”
Schulte, et al (2004) stated that they found the MSCEIT to overlap substantially with what traditional intelligence as well as Big Five Personality factors. Murphy and Sideman, (2006) suggests that there are two possible interpretations of the value of MSCEIT. Firstly, that it is unlikely to add anything new of meaningful significance with regard to underlying human performance that one would not also get from ‘g’ or the Big Five factors. Secondly, ‘g’ and / or the Big Five factors do not necessarily add to what one would expect to predict based on the MSCEIT. When considering application of EI, it is worth noting Boniwell, (2012) who alludes to a dynamic relationship between positive psychology, mindfulness, and resilience as having an increasing influence on current perceptions of EI. Positive psychology is the study of optimal human functioning in order to promote those factors that enable individuals and communities to grow and thrive (Seligman & Csikszentmihalyi, 2000)