The purpose of this page is to introduce the ability model of emotional intelligence and the MSCEIT as a measure of it. A full history of the field is available elsewhere (Mayer, 2001). A brief introduction, however, can suffice here. In 1990, Mayer and Salovey published two articles on emotional intelligence. The first article (Salovey & Mayer, 1990) reviewed literature throughout the disciplines of psychology and psychiatry, artificial intelligence, and other areas, and concluded that there might exist a human ability fairly called emotional intelligence. The idea was that some people reasoned with emotions better than others, and also, that some people’s reasoning was more enhanced by emotions than others. The companion article (Mayer, DiPaolo, & Salovey, 1990) presented a first ability model of emotional intelligence — a suggestion that emotional intelligence, measured as a true intelligence, might exist. Since that time, Mayer, Salovey, and their colleagues refined their model of emotional intelligence (see Mayer & Salovey, 1997), and expended considerable efforts toward developing a high-quality ability measure in the area. The newly developed Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT) is the result of this theoretical and empirical research.

Many questions now arising about the MSCEIT were first studied empirically during the 1990’s. Here are a few questions that are often asked about measuring emotional intelligence and some quick answers. More details are provided below, and of course, these debates can also be followed in our publications on measuring emotional intelligence throughout the 1990’s and today (many of which are referenced below).

What do you mean, "ability model" of emotional intelligence?

When Dan Goleman popularized our own work on emotional intelligence, and the work of many others in related areas, he redefined emotional intelligence in a number of different ways. Since then, people have interpreted emotional intelligence in still more ways — some of which have little to do specifically with emotion and intelligence. To distinguish our model from others, we refer to it as an ability model of emotional intelligence. That is because it is centered on two things: The ability to reason with emotion, and the capacity of emotions to enhance thought.

Why is the ability model important scientifically?

The ability model of emotional intelligence defines a new psychological variable that is theoretically believed to be distinct from previously measured psychological qualities. There is now growing empirical evidence that tests specifically based on the theory measure something new.

What mental tasks measure emotional intelligence?

The abilities involved in emotional intelligence include "the ability to perceive emotions, to access and generate emotions so as to assist thought, to understand emotions and emotional knowledge, and to reflectively regulate emotions so as to promote emotional and intellectual growth " (Mayer & Salovey, 1997). These abilities are based both on an analysis of how intelligence operates and the sorts of knowledge structures and processes necessary to process emotional information.

I have heard there are some mental tasks on ability scales that don’t seem to fit my picture of what should be there. How did that come about?

Most people find the MSCEIT has "face validity" — that is, overall and individually, the tasks appear to measure EI. Still, some people find the test also includes some less obvious tasks, like reading emotions in landscapes "Pictures", or drawing analogies from emotions to physical sensations "Synesthesia." Some people have wondered how those tasks got there. The quick answer is this. First, for reasons beyond the scope of this posting, procedures for psychological measurement generally encourage measuring an ability with more than one task; therefore, several different tasks were employed in measuring each area of emotional intelligence. Second, researchers experimented with many different kinds of tasks in an attempt to measure all four hypothesized areas. Factor analysis indicated that a number of tasks, such as "Pictures" and "Synesthesia" ended up performing very well, and are parallel to the more obvious tasks. Because they worked so well in measuring EI, these tasks were retained in the final test.

How should it be scored?

What is the best criterion for a correct answer on an ability scale of EI? Mayer & Geher (1996) compared consensus and target scoring. Consensus criteria uses the modal group answer as the best answer. Target scoring uses the individual being evaluated (e.g., the person making a face) as the criterion. Briefly, it turned out that consensus outperformed target scoring. For one thing, most targets skew their emotions very positively, so as to appear socially desirable. Mayer, Caruso, & Salovey (1999) studied three criteria: Consensus, Target, and Expert criteria. Expert criteria uses emotions experts to set the correct answer. They found that Consensus and Expert scoring work best, followed by Target scoring. The new MSCEIT test can be scored either by consensus or by expert criterion. Most people get about the same score, no matter which way it is scored. Researchers are encouraged, however, to use both scoring methods and compare them.

Are the tests reliable?

The MSCEIT reliabilities are reported below. It is a highly reliable test, comparable to other tests on the market.
 
 

A Brief Introduction to the Four-Branch Ability Model of Emotional Intelligence

In this model, emotional intelligence is viewed as consisting of four separate components or branches.  Here is a summary of this four-branch model of emotional intelligence:

1) Perceiving and Identifying Emotions - the ability to recognize how you and those around you are feeling.

2) Facilitation of Thought - the ability to generate emotion, and then reason with this emotion.

3) Understanding Emotions - the ability to understand complex emotions and emotional "chains", how emotions transition from one stage to another.

4) Managing Emotions - the ability which allows you to manage emotions in your self and in others.
 
Further reading on the ability model:

Salovey, P., & Mayer, J.D. (1990). Emotional intelligence. Imagination, Cognition, and Personality, 9, 185-211.

Mayer, J.D., & Salovey, P. (1997). What is emotional intelligence? In P. Salovey & D. Sluyter (Eds.), Emotional development and emotional intelligence: Implications for educators (pp. 3-31). New York: Basic Books.
 

Different approaches to emotional intelligence are discussed in:
Mayer, J.D., Salovey, P., & Caruso, D. (2000). Models of emotional intelligence. In R.J. Sternberg (Ed.), The handbook of intelligence (pp. 396-420). New York: Cambridge University Press.
   

Different definitional approaches to emotional intelligence have also led to different measurement approaches. Please consult our brief descriptions of some of these measurement approaches.

Also see Mayer, J.D., Caruso, D. R., & Salovey, P. (2000). Selecting a measure of emotional intelligence: The case for ability testing. In R. Bar-On & J. D. A. Parker (Eds.). Handbook of emotional intelligence (pp. 320-342). New York: Jossey-Bass.
 

MSCEIT OVERVIEW

Description

The MSCEIT is an ability test of emotional intelligence designed for adult ages 17 years and older.  Normative data are from a sample of 5,000 individuals.

The MSCEIT consists of 141 items that yield a total emotional intelligence score, two Area scores, and four Branch scores.  The eight task-level scores are reported for research and qualitative use only.

The MSCEIT asks test takers to:

Identify the emotions expressed by a face or in designs.
Generate a mood and solve problems with that mood.
Define the causes of different emotions. Understand the progression of emotions.
Determine how to best include emotion in our thinking in situations that involve ourselves or other people.
 

Click here to view example MSCEIT items.

Administration

Paper and Pencil - Via a re-usable color test booklet and a one-time use answer sheet.
The answer sheet is mailed or faxed to the test publisher for scoring.

On-Line - The on-line version is scored automatically.

Testing time about 30 to 45 minutes.
 

Availability

The MSCEIT is available from Multi-Health Systems (www.mhs.com ).   MSCEIT prices are listed here.
 

Psychometrics

Scoring the MSCEIT

The MSCEIT is objectively scored. One reason why this is the case is that emotions have evolved over time as a complex, adaptive signaling system.

Many theorists agree that basic emotions have universal meaning - universal across cultures and even across certain species.
 

Consensus Scoring

Consensus scoring is based upon the agreement of a large number of people.  For example, if 70 percent of people felt that a photo was of a very happy person, then the best answer for the photo would be "happiness".

Consensus scoring "works" because of the evolutionary and social basis of emotion and its expression. In addition, a low score on such a test would be achieved by those people who are "off" or whose  perceptions of emotion are so unique as to cause them interpersonal problems.
 

Expert Scoring

Even though emotions convey information about social relationships, we have developed an alternative scoring method that relies upon expert judgment.  In the expert method, emotions experts determine which test answers are better, and which are worse. While our earlier test employed just two raters, the MSCEIT's expert scoring method employs 21 members of the International Society for Research on Emotions (ISRE).

If the two scoring methods were to differ radically, we would be faced with a major scoring issue.  Fortunately, the two methods yield very similar results, indicating that there are indeed better and worse answers on the MSCEIT.

In general, experts tend to agree more with each other, and to diverge from the general consensus, in those areas of emotional intelligence where the body of knowledge is better developed (i.e., perception and understanding).
 

Similar to Intelligence Tests?

Those of you familiar with the Wechsler scales of intelligence will realize that some Wechsler subtests (e.g., Comprehension) also utilize an expert scoring method.
 

Why Two Scoring Methods?

Multiple scoring methods have allowed us to determine whether it is possible to create an ability test of emotional intelligence that can be objectively scored. We need to demonstrate that there are better, and worse, answers on such tests. Given the nature of emotional information , we believe that there is adequate justification for both a general consensus scoring method as well as an expert method. Over time, however, it is likely that we will move to a single scoring method.
 

Additional Information

This is a complex and important issue. Please consult this recent article which describes scoring issues in some detail, including Scoring and Consensus issues.

MSCEIT Reliability

Internal Consistency

Internal consistency reliability for the MSCEIT V 2.0's normative sample are reported in the table below for both scoring methods.
   

Test - Retest

Brackett & Mayer (2001) found a test-retest reliability for the full-scale MSCEIT V2.0 of  .86, based on a sample of 62 people.
 

MSCEIT Validity

Face Validity

If a test appears to measure what it is supposed to measure, it has face validity.  One study explicitly examined the face validity of the MSCEIT in the workplace and concluded that "In general, the MSCEIT has good face validity" Pusey (2000).  We have additional information on face validity of the MSCEIT for researchers as well as for test-takers.
 

Content Validity

If a test's items are systematically drawn from the areas that the test is supposed to measure it is considered to have content validity. Remember that the MSCEIT is operationalizing the ability model of emotional intelligence. Therefore, the MSCEIT should measure the ability to identify emotions in persons and objects; the ability to generate emotion and use it to solve problems; the ability to understand emotional causes and complexity; and, the ability to manage emotion to enhance growth?
 

Factor Structure

Our factor analyses of the MSCEIT, based upon a sample of 1,985 test takers, are highly supportive of the four-branch model of emotional intelligence.  (Please ask us for a copy of our recent manuscript, Modeling and Measuring Emotional Intelligence with the MSCEIT V2.0.)
 

Discriminant Validity  - If a test correlates at very high levels with other tests, then it may lack discriminant validity.  This type of validity means that a test is measuring something relatively unique.

IQ  - Salovey, Mayer, Caruso, and Lopes (in press), in a sample of 97 participants, found nonsignificant correlations close to zero between the MSCEIT V2.0 and (self-reported) Verbal and Math SAT scores, as well as r = .15 (ns) with the Vocabulary scale of the WAIS-III.

Emotionality  -Salovey and colleagues (in press), found an intercorrelation between the MSCEIT and mood state of .08, ns, in a sample of 97 participants.

BarOn EQ-i  - The correlation between the MSCEIT RV1.1 EIQ and the BarOn EQ-i was .13 (ns) in a sample of 130 ethnically diverse students (Pellettieri, 2001).

TMMS  - The correlation between Trait Meta-Mood Scale (a scale of meta-experience of mood; see Mayer & Gaschke, 1988; Salovey et al., 1995) and the MSCEIT was .29 (p < .01) in a sample of 318 men and women (Gohm and Clore, 2001).  Salovey and colleagues (in press), in a sample of 97 participants, found correlations of .01 to .16 (all nonsignificant) with the TMMS.

NEO Personality Inventory  - Salovey and colleagues (in press), in a sample of 97 participants, correlated the MSCEIT V2.0 with the NEO PI and other scales. Its correlations were r = -.13 (ns) with Neuroticism, r = .04 (ns) with Extroversion, r = .33 (p < .05) with Agreeableness, r = -.23 (p < .05) with Openness, and r = .25 (p < .05) with Conscientiousness.
 

Concurrent Validity

What does emotional intelligence, and the MSCEIT, predict? Contrary to the claims in the popular press, we are certain that emotional intelligence is not "twice as important as IQ". Indeed, we know of no psychological variable that is that powerful a predictor. The MSCEIT will likely predict important outcomes, but at levels that one usually obtains in psychological research.

There are a number of studies that are in the field, but those that have been completed suggest that the MSCEIT offers additional predictive validity for outcomes such as pro-social behavior, deviancy, and academic performance (see Mayer et al., 2002b).

We again remind researchers and practitioners that the applied use of emotional intelligence tests must proceed with great caution.
   

Additional Information

The purpose of this page is to introduce certain background concepts regarding the MSCEIT. Additional information can be obtained from the MSCEIT publisher, in the MSCEIT Manual, through a series of articles, and also, on-line at this site (www.emotionaliq.org).

A list of references is also available.