APA NLM

tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

Psychological Assessment © 2015 American Psychological Association

2015, Vol. 27, No. 4, 000 1040-3590/15/$12.00 http://dx.doi.org/10.1037/pas0000269

The Factor Structure and Psychometric Properties of the Spanish Version

AQ:1-3 of the Mayer-Salovey-Caruso Emotional Intelligence Test

AQ: au Manuel Sanchez-Garcia Natalio Extremera and Pablo Fernandez-Berrocal

University of Huelva University of Malaga

This research examined evidence regarding the reliability and validity of scores on the Spanish version
of the Mayer-Salovey-Caruso Emotional Intelligence Test, Version 2.0 (MSCEIT; Mayer, Salovey, &
Caruso, 2002). In Study 1, we found a close convergence of the Spanish consensus scores and the general
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

and expert consensus scores determined with Mayer, Salovey, Caruso, and Sitarenios (2003) data. The
This document is copyrighted by the American Psychological Association or one of its allied publishers.

MSCEIT also demonstrated adequate evidence of reliability of test scores as estimated by internal
AQ:4, 5 consistency and test–retest correlation after 12 weeks. Confirmatory factor analysis supported a 3-level
higher factor model with 8 manifest variables (task scores), 4 first-level factors (corresponding to the
4-branch model of Mayer & Salovey [1997], with 2 tasks for each branch), 2 second-level factors
AQ: 6 (experiential and strategic areas, with 2 branches for each area), and 1 third-level factor (overall
emotional intelligence [EI]), and multigroup analyses supported MSCEIT cross-gender invariance. Study
2 found evidence for the discriminant validity of scores on the MSCEIT subscales, which were
differentially related to personality and self-reported EI. Study 3 provided evidence of the incremental
validity of scores on the MSCEIT, which added significant variance to the prospective prediction of
psychological well-being after controlling for personality traits. The psychometric properties of the
Spanish MSCEIT are similar to those of the original English version, supporting its use for assessing
emotional abilities in the Spanish population.

Keywords: psychometric properties, emotional intelligence, MSCEIT, discriminant validity, incremental

validity

Since Mayer and Salovey published their emotional intelligence mixed model), EI is defined as a constellation of emotional self-
(EI) construct in 1990 (Salovey & Mayer, 1990), much research perceptions that are located in the lower levels of personality
has shown that the abilities to perceive, assimilate, understand, and hierarchies (Petrides & Furnham, 2001).
regulate the emotions in oneself and others are important predic- Based on these different theoretical models, researchers have
tors of personal and social functioning (for a review, see Mayer, designed different assessment tools to evaluate EI. Therefore,
Roberts, & Barsade, 2008). Currently, there are two main theoret- trait models involve a constellation of emotion-related self-
ical approaches in EI research. Mayer and Salovey (1997) concep- perceptions and dispositions and should be measured via self-
tualize EI as the ability to accurately perceive, assess, and express report questionnaires, whereas ability models operationalize EI
emotions; the ability to access and generate sensations that facil- as a cognitive ability and should be measured via performance
itate thought; the ability to understand emotions and emotional tests. A performance test is characterized as presenting different
knowledge; and the ability to regulate emotions to promote emo- emotional tasks and decision-making problems to the partici-
tional and intellectual growth. Based on this ability model, EI is pants, as opposed to self-reports that measure emotional self-
considered similar to cognitive intelligence and includes the ca- efficacy.
pacity to process emotional information (Mayer & Salovey, 1997). Although both approaches have shown empirical evidence for
On the other hand, based on the EI trait model (also known as the predicting relevant personal outcomes, Mayer and Salovey (1997)
argue that EI, as another kind of intelligence, is best described as
a set of intelligences and, therefore, is best measured by ability-
based tools (Mayer, Salovey, & Caruso, 2000). Following this
Manuel Sanchez-Garcia, Department of Clinical, Experimental, and reasoning, different instruments for measuring EI have been de-
Social Psychology, Faculty of Education Sciences, University of Huelva; veloped in recent years, such as the Multi-Emotional Intelligence
Natalio Extremera, Department of Social Psychology, Faculty of Psychol- Scale (Mayer, Caruso, & Salovey, 1999), the Situational Test of AQ: 7
ogy, University of Malaga; Pablo Fernandez-Berrocal, Department of Emotional Understanding (MacCann & Roberts, 2008), and the
Basic Psychology, Faculty of Psychology, University of Malaga.
Test of Emotional Intelligence (Blickle, Momm, Liu, Witzki, &
This research was supported in part by projects SEJ-07325, PSI2012-
Steinmayr, 2011). However, the EI performance test that is the
37490 and PSI2012-38813 (Spain).
Correspondence concerning this article should be addressed to Manuel most widely accepted in the research community is the one devel-
Sanchez-Garcia, Department of Clinical, Experimental and Social Psychol- oped by Mayer and colleagues, known as the Mayer-Salovey-
ogy, Faculty of Education Sciences, University of Huelva, Avd. Fuerzas Caruso Emotional Intelligence Test (MSCEIT; Mayer, Salovey, &
Armadas, S/N, 21007 Huelva, Spain. E-mail: msanchez@uhu.es Caruso, 2002; Mayer, Salovey, Caruso, & Sitarenios, 2003), which

1
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

2 SANCHEZ-GARCIA, EXTREMERA, AND FERNANDEZ-BERROCAL

was designed to assess the four-branch EI model (Mayer & Sa- Each of the eight MSCEIT tasks, in turn, comprises an item
lovey, 1997). parcel; some items refer to the same stimulus, whereas others are
The MSCEIT is composed of 141 items divided into eight tasks free-standing. For example, a parcel structure appears when an
(two for each branch) that are designed to measure the four individual is shown a face (the face task) and is asked to evaluate
branches of EI: perceiving emotions, using emotions to facilitate the presence of five different emotions on that face through a
thought, understanding emotions, and managing emotions (Mayer graduated, five-choice response format (e.g., from no happiness to
& Salovey, 1997). In general, the MSCEIT yields 15 scores: a total extremely happy). The five items related to those emotions com-
score, two area scores (experiential and strategic), four branch pose a parcel because they refer to the same face, although each
scores (corresponding to the four-branch model), and eight task item refers to a different emotion. However, other items require a
F1 scores (see Figure 1). response to a stimulus and are, in that sense, independent, although
AQ: 8 The first two factors in the model, Perceiving and Facilitating, they also have five answer choices (e.g., judging a behavior’s
form part of the experiential area because of their close relation to effectiveness in managing emotions from 1 ⫽ very ineffective to
sensations. Thus, the Perception factor has to do with the ability to 5 ⫽ very effective; Mayer et al., 2003).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

perceive emotions in others, whereas the Facilitating factor refers

This document is copyrighted by the American Psychological Association or one of its allied publishers.

to being able to use emotions to improve one’s own thought

Scoring Methods
process. On the other hand, the Understanding and Managing
factors are part of the strategy area because they involve the ability The MSCEIT is objective because some answers are better and
to evaluate and plan action based on the information provided by others are worse, as determined by general or expert consensus
sensations and emotions. The understanding factor refers to the scoring. General consensus scores reflect the proportion of people
ability to know how one’s emotions change and also how other in the normative sample who endorse each MSCEIT test item.
people’s emotions change, altering their behavior over time. The Expert norms were obtained from a sample of 21 members of the
managing factor refers to the ability to integrate logic and emo- International Society for Research on Emotions who provided their
tions to more effectively make decisions. expert judgment on each item.
Each of the factors or branches is evaluated by means of two Thus, the score assigned to each item’s answer choice is based
tasks. The ability to perceive emotions is determined by perceiving on the proportion of persons who chose it as correct, whether by
emotions in facial expressions and photographs; the Emotional general or expert consensus. For instance, in the scores generated
Facilitating factor is measured by sensations and facilitation tasks, by general consensus, if the proportion of choices related to
and the ability to understand emotions is assessed through one task happiness was .10, .10, .05, .20, and .55 in the standardized
for blending emotions and another for combining emotional sample, then a person who chooses 5, extremely happy, would
changes or transformations. Finally, the ability to manage emo- score .55, whereas a person who chooses 1, unhappy, would score
tions is evaluated through an emotional management task and an .10. The scores for the expert consensus were similarly based on
emotional relations task. each answer choice percentage, as selected by experts.

Figure 1. Diagram of the Mayer-Salovey-Caruso hierarchical model. Fit indices and standardized factor
saturations.
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

SPANISH MSCEIT: PSYCHOMETRIC PROPERTIES 3

The convergence of the two scoring procedures with other best-fitting model was a modified two-factor model proposed by
procedures developed in other contexts, cultures, or languages Eack et al. (2010).
determines the usefulness of the scores. In this sense, analyses of
MSCEIT standardization data (n ⫽ 2,112) by Mayer et al. (2003) Evidence of Convergent and Discriminant Validity
demonstrated a closer convergence between expert and general
consensus scoring methods (r ⫽ .908). Palmer, Gignac, Manocha, MSCEIT areas, factors, and tasks are related to each other,
and Stough (2005), using an Australian sample, also found a strong although they are functionally different and do not conceptually or
relationship (r ⫽ .993) between scores determined using American empirically overlap (Mayer et al., 2003; Palmer et al., 2005).
consensus weights and scores using Australian consensus weights. However, the MSCEIT itself is relatively independent of other
However, other studies have questioned the adequacy of this personality constructs, insofar as it shows moderate correlations
scoring system, especially with regard to general consensus (see, with some measures of general cognitive skill that include non-
e.g., Davies, Stankov, & Roberts, 1998; Keele & Bell, 2009; verbal intelligence and, especially, verbal intelligence (Kong,
Roberts, Zeidner, & Matthews, 2001). 2014; Roberts et al., 2001; Van Rooy, Viswesvaran, & Pluta,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

2005). Furthermore, the MSCEIT scores show evidence of dis-

This document is copyrighted by the American Psychological Association or one of its allied publishers.

criminant validity with respect to other self-reported EI measures,

Factor Structure analytical intelligence measures, and several well-being and per-
Nevertheless, the factor structure underlying MSCEIT scoring sonality variables (Brackett & Mayer, 2003; Brackett, Rivers,
has been at the center of intense controversy in the scientific Shiffman, Lerner, & Salovey, 2006; Rossen & Kranzler, 2009).
literature. This structure determines the number of dimensions that
must be measured and interpreted (Brackett & Salovey, 2006). Evidence of Predictive and Incremental Validity
Several studies have attempted to provide evidence in favor of the
Low scores on the MSCEIT have primarily been associated with
theoretical, four-branch structure. Some researchers have claimed
the male gender and with high illegal drug and alcohol use, violent
that their studies support a four-factor MSCEIT structure that is
behavior, low academic performance, and poor relations with
consistent with the four-branch model (e.g., Brackett, Rivers, &
friends (Brackett, Mayer, & Warner, 2004; Rossen & Kranzler,
Salovey, 2011; Day & Carroll, 2004; Mayer et al., 2002, 2003).
2009), as well as with higher depression scores (Goldenberg,
Whereas others have doubts concerning the suitability of this
Matheson, & Mantler, 2006; Salguero, Extremera, & Fernández-
factor structure, in most cases, these researchers have used a
Berrocal, 2012) and stronger desires for vengeance for interper-
college student population (Fan, Jackson, Yang, Tang, & Zhang,
sonal offenses (Rey & Extremera, 2014). There are many reasons
2010; Gardner & Qualter, 2011; Gignac, 2005; Palmer et al., 2005;
for believing that EI plays an important role in predicting one’s
Roberts et al., 2006; Rode et al., 2008; Rossen, Kranzler, &
subjective sense of well-being and positive mental health. In fact,
Algina, 2008) and proposed alternative one-, two-, three-, or
some studies have found positive relationships between the
four-factor models. Despite the proliferation of models, the hier-
MSCEIT and psychological well-being (PWB) and life satisfaction
archical model (i.e., the MSCEIT’s implied theoretical structure of
using experimental (Brackett et al., 2006) and prospective designs
eight tasks, four branches, two areas, and general EI) has been
(Extremera, Ruiz-Aranda, Pineda-Galan, & Salguero, 2011). AQ: 9
tested on a few occasions and yet is almost never proposed as the
best model for later analyses, even though Mayer et al. (2002)
thought that this type of test best determined the clustering of Gender Comparisons
MSCEIT scores according to theory. There is ample evidence of gender differences in MSCEIT
scores in the literature (Brackett & Mayer, 2003; Brackett et al.,
Factorial Invariance 2004, 2006; Fernández-Berrocal, Cabello, Castillo, & Extremera,
2012; Palmer et al., 2005; Salguero et al., 2012). Moreover, the
One of the research goals most tested in studies using MSCEIT MSCEIT manual (Mayer et al., 2002) and recent meta-analytical
scores is the comparison of target groups (cultures, gender, age, studies suggest that women score higher on the four MSCEIT
etc.). A preliminary step in these analyses should be the verifica- factors than men do (Joseph & Newman, 2010) and that gender
tion of the equivalence of MSCEIT measures among those target might act as a moderator in EI studies (Kong, 2014).
groups (Meredith, 1993). Thus, to provide more evidence of the
suitability of the MSCEIT measurements, the factorial invariance
Overview of Studies
of its scores among different groups has been tested, although in no
case has the invariance of the hierarchical model been tested. For The MSCEIT was adapted into Spanish using back translation
example, Gardner and Qualter (2011) tested invariance in two age with a sample of 946 high school and university students (Extrem-
groups (young vs. older adults) but used a three-factor model (the era, Fernández-Berrocal, & Salovey, 2006), and it showed prelim-
experiential area and the understanding and management inary evidence that the reliability of the total scores (by area,
branches). Karim and Weisz (2010) tested the cross-cultural in- branches, and subscales) is adequate when using Spanish samples.
variance (Pakistani-French students) of the correlated four-factor However, evidence of convergence between the Spanish and orig-
(branch) model. In a study in which schizophrenia and normative inal English consensus methods, the factor structure underlying the
samples were compared, Eack, Pogue-Geile, Greeno, and Kesha- scores, gender invariance, score stability (test–retest), and other
van (2009) could not find factorial invariance. According to their evidence of validity based on the relationship with other variables
data, the four-factor model is best used for the normative sample, in the Spanish version of the MSCEIT have still not been explored
whereas for the sample of the population with schizophrenia, the or analyzed.
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

4 SANCHEZ-GARCIA, EXTREMERA, AND FERNANDEZ-BERROCAL

The main purposes of this research were to analyze the factor Further evidence of validity was collected using multigroup
structure of the Spanish MSCEIT scores, its reliability, the test– analysis as a function of gender. For the invariance comparison,
retest stability, and the level of convergence between the Spanish we followed the procedure applied by Wright, Lukowitsky, Pincus,
consensus scores and the general and expert consensus scores and Conroy (2010), which was based on the recommendations of
determined with Mayer et al. (2003) data and gender invariance Chen, Sousa, and West (2005). To do this, we first separately
(Study 1). We also wanted to evaluate evidence of its discriminant tested the hierarchical model shown in Figure 1 for men and
validity by examining the relationship of its scores and conceptu- women to determine whether the model fit each group. In the
ally distinguishing between personality traits and self-reported EI multigroup analysis, we began by specifying a baseline model
scores (Study 2). Finally, we examined evidence of the predictive (Model 1) of configurational invariance in which no parameter was
and incremental validity of the Spanish MSCEIT scores to explain restricted, and it was assumed that the factor structure was the
the significant variance in prospective (2-month) predictions of same for both groups. Afterward, restrictions were added to this
PWB by controlling for the relevant predictors of PWB, such as baseline model, verifying its fit and comparing this fit to the model
personality traits (Study 3). that preceded it, in the following order: (a) invariance of the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

first-order factor loadings (Model 2), (b) invariance of the second-

This document is copyrighted by the American Psychological Association or one of its allied publishers.

order factor loadings (Model 3), (c) invariance of the third-order

Study 1
factor loadings (Model 4), (d) error variance equivalents (Model
This first study had several purposes. First, we wanted to pro- 5), and (e) equivalence of first- and second-order factor distur-
vide proof of the convergence of these scores with those generated bances (Model 6). Invariance was assumed when each of the
AQ: 10 by the general and expert consensus of Multi-Health Services. models fit and when its fit was not significantly worse than its
Once convergence was demonstrated (i.e., if they were in agree- predecessor. For model comparisons, we used the chi-square dif- AQ: 15
ment), we would proceed to the psychometric analysis of our data. ference and change in the CFI. If there was any discrepancy, we
To do so, we analyzed the evidence for the reliability and validity preferred the change in the CFI; we considered a change ⱕ.01 as
of scores from the Spanish version of the MSCEIT. To prove the being indicative of invariance (Cheung & Rensvold, 2002; Rast,
reliability of the scores, we estimated the coefficients of internal Zimprich, Van Boxtel, & Jolles, 2009; Wright et al., 2010).
consistency and their stability. Evidence for validity came from (a)
the factor structure (by which we tested the hierarchical model by
Results
Mayer et al., 2002); (b) the cross-gender factorial invariance of this
model, and whether invariance was demonstrated; and (c) the Convergence of scoring criteria. We found a close relation-
analysis of the differences in MSCEIT scores by gender. ship between scores assigned to the item responses using the
Spanish consensus weights and scores using the American general
(r ⫽ .93) and expert (r ⫽ .85) consensus.
Method
Furthermore, the generated scores for the people in our sample
AQ: 11 Participants and procedure. The sample consisted of 3,448 were based on different criteria (American consensus, experts, and
Spanish college and high school students (69% females) ranging in Spanish consensus) and are very highly and significantly corre-
age from 17 to 76 years (M ⫽ 24.48, SD ⫽ 10.67). The partici- lated (see Table 1). All the correlations are above .90, except those AQ:16,
T1
pants were first given a set of brief written and verbal instructions related to the facilitating branches and tasks.
and were informed of their anonymity. Next, the MSCEIT v. 2.0 Descriptive statistics and reliability. Table 1 shows the
was administered to the participants in groups of 30 to 50. Finally, means, standard deviations, and reliability of the scores generated
the participants were debriefed and thanked for their participation. from the scoring criteria of the Spanish normative sample. The 15
AQ: 12 Measures: Emotional intelligence test. The Spanish version scores varied from 0.04 (the lowest score, which was for the
of the MSCEIT v. 2.0 (Extremera & Fernández-Berrocal, 2009; pictures task; mean of 0.46) to 0.63 (the highest, corresponding to
AQ:13, Mayer et al., 2002) was administered. the faces task; M ⫽ 0.50). For the total EI score, the values varied
14
Statistical procedures. Non-CFA or factorial invariance anal- from 0.16 to 0.54 (M ⫽ 0.43).
yses were conducted using IBM SPSS Statistics Version 19 soft- Concerning the internal consistency of the scores in our sample,
ware. the total MSCEIT has a reliability of .94, with area reliabilities of
To analyze the test’s factor structure, we conducted a maximum .92 for the experiential area and .98 for the strategic area. The
likelihood confirmatory factor analysis (CFA) of the hierarchical reliability estimates of branch scores varied from .71 (Facilitating)
model (three-level higher factor model) by Mayer et al. (2002; see to .92 (Perceiving).
Figure 1) with EQS 6.0 software (Bentler, 1995). The scores that Score stability evidence for scale reliability, which was esti-
were directly used as observable indicators of the factors were the mated based on the correlation coefficient of the test–retest scores,
scores on the tasks, not on the original items. Similar to Mayer et was very good. All the correlations were above .7 except for the
al. (2002), and following the recommendations of Cattell and relationships task. For this test, the equality of means and vari-
Burdsal (1975) and Hughey and Burdsal (1982), we combined the ances between the two test applications was also verified, which
scores on the test item parcels, which shared characteristics that enabled these high correlations to be interpreted as evidence of the
differentiated the tasks. To determine the overall model fit along scores’ reliability and stability.
with the chi square, we used indexes that were less sensitive to Evidence of validity based on the internal scale structure:
sample size, including the non-normed fit index (⬎ .90), the Factor structure. Evidence of validity based on the internal
comparative fit index (CFI ⬎ .90), and the root mean square error structure of the Spanish version of the MSCEIT would be found if,
of approximation (⬍ .08; Kline, 2005). based on the data, we were able to reproduce the theoretical model
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

SPANISH MSCEIT: PSYCHOMETRIC PROPERTIES 5

Table 1
Descriptive Means and Standard Deviations, Correlations of the Spanish Consensus Scores with
the U.S. General and Expert Consensus, and Reliability (Internal Consistency and Test–Retest)
of the Spanish Consensus Scores

Correlations
N ⫽ 3,448
Reliability internal Reliability
M (SD) U.S. Expert consistency test–retest
Group of scales Scales N ⫽ 3,448 consensus consensus N ⫽ 3,448 N ⫽ 349

Global Total .43 (.06) .98 .97 .94 .81

Areas A. Experiential .45 (.07) .96 .96 .92 .81
B. Strategic .41 (.07) .99 .97 .87 .82
Branches 1. Perceiving .47 (.10) .99 .98 .92 .83
2. Facilitating .42 (.07) .85 .85 .71 .82
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

3. Understanding .45 (.08) .97 .96 .80 .83

This document is copyrighted by the American Psychological Association or one of its allied publishers.

4. Managing .37 (.07) .99 .95 .83 .81

Tasks
Branch 1 A. Faces .50 (.11) .99 .98 .79 .76
E. Pictures .46 (.11) .99 .98 .88 .88
Branch 2 B. Facilitating .44 (.08) .70 .68 .64 .79
F. Sensations .39 (.08) .95 .95 .66 .80
Branch 3 C. Changes .45 (.08) .96 .93 .69 .86
G. Blends .46 (.10) .96 .98 .66 .80
Branch 4 D. Management .38 (.07) .98 .91 .67 .86
H. Relations .37 (.09) .99 .93 .69 .63
Note. Cronbach’s alpha coefficient was used to calculate the internal consistency of the tasks. For the rest of
the scores, the Spearman-Brown coefficient (split-half) was calculated. U.S. general and expert consensus scores
were determined with Mayer et al. (2003) data.

of eight tasks, four branches, two areas, and an overall score (see configurational invariance model also had good fit indices, which
Figure 1). However, this model showed identification and solution shows that the described factor structure is maintained for men and
AQ: 17 problems for the involved equations. According to Bollen (1989), women.
in a second-order CFA, the higher order factors must have at least As restrictions were added to the various models, the factorial
three factors in the lower order. In this model, however, only two invariance of the MSCEIT scores between the groups was con-
lower order factors appeared for each higher order factor. To solve firmed. Although the change in chi square (⌬␹2) was significantly AQ: 18
this problem of model identification, Rossen et al. (2008) sug- different in some cases, the differences between CFIs in the nested
gested setting the factor saturations of the third-order factor over models were never higher than .01. In practice, in the last model,
the second-order equal to one another or setting the variances of the standardized factorial saturations were the same for men and
error of the second-order factors equal to one another. The second women. Therefore, we may conclude that the MSCEIT is a robust
option was chosen for use in this research. Once the parameters instrument that has strong invariance between men and women, as
were estimated, an inappropriate solution appeared (showing a demonstrated by their comparable scores.
negative variance of error of the first-order factor in the facilitating Evidence based on the relationship with other variables: EI
branch); therefore, following the recommendation of Chen, Bollen, and gender. In our sample, women scored higher than men on
Paxton, Curran, and Kirby (2001), we set the variance of error to all tasks, branches, areas, and total score. The differences between
.0001. the mean scores are statistically significant (p ⬍ .001), and the
Once the identification problems were resolved, the fit indices effect size (Cohen, 1988), in general, may be qualified as medium
revealed that the proposed model was adequate. Furthermore, all (Cohen’s d from 0.36 to 0.56). The widest differences are seen in
the factor saturations were relevant, significant, and suitable for the branches, the areas, and the total score (see Table 3). AQ:19,
T3
the theoretical model. The second- and third-order factor model
adequately reproduced the task correlations.
Discussion
Evidence of validity based on the internal scale structure:
Factorial invariance by gender. Before proceeding to the anal- The results of Study 1 demonstrate that the Spanish version of
ysis of the gender differences in the MSCEIT scores, we needed to the MSCEIT is structurally and functionally very similar to the
determine whether the MSCEIT scores were equivalent across original English version, and that the criteria for scoring seem
men and women; thus, we tested the MSCEIT cross-gender mea- adequate and very close to those used by Mayer et al. (2002, 2003).
surement equivalence. However, factorial invariance can also be Furthermore, with regard to the Spanish version and the sample of
interpreted as more evidence for the validity of the test’s under- Spanish subjects, this study provides evidence for the scale’s score
lying factor structure. The results of the invariance and the fit of reliability (the scale’s internal consistency coefficients and stabil-
T2 the different models are shown in Table 2. ity) and validity, which is based on the internal test structure.
The fit of the hierarchical model is adequate for both men and The estimates for the reliability of scores on the Spanish version
women. Therefore, we proceeded to study the invariance. The first of the test used for our sample are satisfactory. These values are
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

6 SANCHEZ-GARCIA, EXTREMERA, AND FERNANDEZ-BERROCAL

Table 2
Fit Indices of the EI Model by Gender and Multigroup Analysis

Model comparisons
Model df ␹ 2
NNFI CFI RMSEA Models ⌬df ⌬␹2 p ⌬CFI

Women (n ⫽ 2,391) 16 163.65 .94 .96 .062 — — — — —

Men (n ⫽ 1,053) 16 114.61 .94 .97 .077 — — — — —

Multigroup analysis: Factorial invariance by gender (N ⫽ 2,391/1,053)

Model 1 32 278.25 .94 .97 .047
Model 2 36 300.86 .94 .96 .046 2–1 4 22.61 .000 .01
Model 3 38 302.86 .94 .96 .045 3–2 2 2.00 .368 .00
Model 4 40 409.96 .93 .95 .052 4–3 2 107.10 .000 .01
Model 5 48 515.63 .92 .94 .053 5–4 8 105.67 .000 .01
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Model 6 51 614.46 .91 .93 .056 6–5 4 98.83 .000 .01

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Note. • • •.

very close to those found in other studies using the original English Study 2
version (see, e.g., Brackett & Mayer, 2003; Mayer et al., 2002,
2003; Palmer et al., 2005; Rossen et al., 2008), even for popula- As other studies have shown, MSCEIT scores are clearly dif-
tions whose original language is not English (e.g., Karim & Weisz, ferentiated from EI measurements based on self-report scales (e.g.,
2010) or when using versions in other languages (e.g., Iliescu, Ilie, Brackett & Mayer, 2003; Brackett et al., 2006) and from other
Ispas, & Ion, 2013). personality characteristics (Rey & Extremera, 2014). The purpose
The CFA performed on the hierarchical model proposed by of this second study was to demonstrate that the four EI branches
Mayer et al. (2002) and the invariance analyses provide empirical from the Spanish version of the MSCEIT are distinguishable from
evidence in favor of the theoretical EI model based on the hierar- personality variables and self-report measures of EI, and, further-
chical structure of eight tasks, four branches, two areas, and a total more, that these self-report measures are less separable from
EI score. This model, although it had good fit indices, had not been personality measures. We therefore attempted to provide evidence
previously used as the baseline model of analysis (see, e.g., Fan et of discriminant validity when comparing the Big Five and Trait
al., 2010; Gardner & Qualter, 2011; and Rode et al., 2008). We Meta-Mood Scale (TMMS; Salovey et al., 1995) scales. AQ: 20

also verified that the MSCEIT scores have the same interpretation
for men and women, which enables statistical comparisons of the Method
two groups.
Thus, as evidence of validity is based on its relationship with Participants and procedure. The participants were 621
other variables, we analyzed the relationship between MSCEIT Spanish undergraduate students (79% women) with ages ranging
scores and gender. We found that our results support those of other from 17 to 54 years (M ⫽ 22.20, SD ⫽ 4.66). The questionnaires
researchers: Women scored higher on the scale than men. As were administered with written instructions and in paper-and-
theoretically and from previous research it was expected that pencil format. The participants completed the questionnaires in
women would score higher, we interpret this difference as further groups and received course credit for their participation.
evidence for the adequacy of the Spanish MSCEIT. Measures.
Emotional intelligence test. The Spanish version of the MS-
CEIT v. 2.0 (Extremera & Fernández-Berrocal, 2009; Mayer et al.,
Table 3 2002) was administered.
Comparison of Means of Men (1,053) and Women (2,391) Big Five Inventory 44. (BFI-44; Benet-Martínez & John,
1998). The BFI-44 is a 44-item, self-report inventory designed to
Standard Mean assess the Big Five factors of personality: Extraversion, Agree-
••• ••• Mean deviation comparison
ableness, Conscientiousness, Neuroticism, and Openness to Expe-
Total Men .4071 .07239 p ⬍ .001 rience. The BFI-44 scales have shown substantial internal consis-
Women .4401 .05239 Cohen’s d ⫽ .56 tency, retest reliability, and clear factor structure. We used the
Experiential Men .4239 .08364 p ⬍ .001
Spanish version of the BFI-44, which has psychometric properties
Women .4560 .06346 Cohen’s d ⫽ .46
Strategic Men .3903 .07623 p ⬍ .001 similar to those of the English version (Benet-Martínez & John,
Women .4241 .05706 Cohen’s d ⫽ .53 1998). In Study 2, we found Cronbach alphas of .85, .62, .79, .83,
Perceiving Men .4495 .10986 p ⬍ .001 and .78 for the Big Five dimensions of Extraversion, Agreeable-
Women .4832 .08571 Cohen’s d ⫽ .36 ness, Conscientiousness, Neuroticism, and Openness to Experi-
Facilitating Men .3983 .08031 p ⬍ .001
Women .4289 .06358 Cohen’s d ⫽ .44 ence, respectively.
Understanding Men .4295 .09008 p ⬍ .001 TMMS (Salovey et al., 1995). We used the modified Spanish
Women .4636 .06864 Cohen’s d ⫽ .45 version of the TMMS (Fernandez-Berrocal, Extremera, & Ramos,
Managing Men .3511 .07920 p ⬍ .001 2004). This instrument is composed of 24 items and provides an
Women .3847 .06430 Cohen’s d ⫽ .49
indicator of perceived EI. The respondents were asked to rate their
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

SPANISH MSCEIT: PSYCHOMETRIC PROPERTIES 7

agreement with each of the items on a 5-point Likert-type scale This factorial solution is maintained almost intact when the
ranging from 1 (very much agree) to 5 (very much disagree). The analyses are separately completed by sex. The MSCEIT factor is
scale is made up of three subfactors: Attention to Feelings, Emo- always maintained separately from any other factorial combina-
tional Clarity, and Mood Repair. Attention to Feelings, assessed by tion. For women, although a fourth factor appears, the same
the first eight items, is the degree to which people believe they pay structure is always maintained: MSCEIT forms a separate factor.
attention to their feelings (e.g., “I think about my mood con-
stantly”); Emotional Clarity, evaluated by the subsequent eight Discussion
items, refers to how people believe they perceive their emotions In this study, we attempted to provide evidence for the conver-
(e.g., “I frequently make mistakes about my sensations”); and gent and discriminant validity of scores on the MSCEIT. We also
Mood Repair, assessed by the remaining eight items, refers to found that there was a closer relationship between personality and
people’s belief in their capacity to interrupt negative moods and TMMS measures than between the personality scales and the
prolong positive ones (e.g., “Although I sometimes feel sad, I MSCEIT.
usually have an optimistic outlook”). In this study, we found First, we interpreted the clustering of the four branches into a
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Cronbach alphas of .87, 88, and .85 for the TMMS Attention, single factor as evidence of convergent validity. These data coin-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Clarity, and Repair dimensions, respectively. cide with those found by Brackett and Mayer (2003), Iliescu et al.
Statistical procedures. We performed three principal-axis ex- (2013), and Koven and Max (2014). In these studies, the basic
ploratory factor analyses with direct oblimin rotation on the whole evidence for convergence was found in the strong branch correla-
sample and on subsamples of men and women for the five Big Five tions.
scales, the three TMMS dimensions, and the four MSCEIT Concerning discriminant evidence, the data in this study agree
branches using the SPSS program for Windows (Version 19). largely with the conclusions of Van Rooy et al. (2005), in which
Multiple criteria, including an examination of the scree plot, ap- the EI measures based on mixed models overlap more with per-
plication of the Kaiser rule (eigenvalues ⱖ1), and parallel analysis sonality measures than with the ability model. Specifically, several
(O’Connor, 2000), were used to identify the optimal number of MSCEIT studies using the Big Five (see, e.g., Brackett et al., 2006;
factors to retain. Iliescu et al., 2013; Rode et al., 2008) and the NEO-BFFI (War- AQ: 24
wick & Nettelbeck, 2004; Webb et al., 2013) as personality mea-
Results sures support this conclusion. Concerning self-report EI scales,
In the first analysis of the whole sample, three factors were some authors use the TMMS (Koven & Max, 2014; Warwick &
AQ: 21 obtained. The eigenvalues of the first three factors were 2.6, 1.7, Nettelbeck, 2004), as we did, or the SREIS (Brackett et al., 2006;
and 1.4, which explain up to a total of 48% of variance. After Webb et al., 2013, who also used the EQ-i). Both in our study and
rotation, the remaining factor structure is clearly interpretable, in those mentioned, a strong relationship between personality
with saturations over ⫾.40 (except for Conscientiousness; see characteristics and self-report measures is found, whereas the
AQ:22, Table 4). Factor 1 is made up of four BF and two TMMS dimen- relationships of these two measures with the MSCEIT scores are
23, T4
sions, Factor 2 is composed of four MSCEIT branches, and Factor irrelevant. Moreover, our results are congruent with those of
3 is made up of Neuroticism (BF, which is also saturated at ⫾.4 McCrae’s (2000) work, according to which the Big Five covers
with the first factor) and Attention (TMMS). Our results also show most of what is measured by mixing the conceptions of EI.
that the branches of MSCEIT are largely distinct from the person-
Study 3
ality measures rather than the TMMS subscales; for example,
Factor 2 of the MSCEIT correlates .22 and ⫺.02 with Factors 1 The aim of this study was to determine the extent to which the
and 3, respectively. MSCEIT would make significant, nonredundant contributions to

Table 4
Factor Solution of Measures of Personality and Emotional Intelligence: TMMS and MSCEIT
Using Principal Axis Factoring With Oblique Rotation (Pattern Matrix)

Full sample Men Women

••• ••• I II III I II III I II III IV

Big Five Extraversion .51 .59 .64

Openness .45 .43 .52 ⫺.26
Agreeableness .42 .48 .43 ⫺.28
Conscientiousness .29 .37 ⫺.37
Neuroticism ⫺.46 .70 ⫺.59 ⫺.34 .82 .39
TMMS Repair .67 .63 .57 ⫺.53
Clarity .62 .63 .45 ⫺.76
Attention .64 .78 .55
MSCEIT Managing .54 .66 .47
Perceiving .53 .47 .56
Understanding .50 .70 .43
Facilitating .50 .47 .55
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

8 SANCHEZ-GARCIA, EXTREMERA, AND FERNANDEZ-BERROCAL

prospective (2-month) predictions of the criterion variable (PWB) (van Dierendonck, Díaz, Rodríguez-Carvajal, Blanco, & Moreno-
beyond what is accounted for by the influence of personality traits Jiménez, 2008). The reliability of the composite PWB score (␣ ⫽
(Big Five dimensions). .88) was high.
Statistical procedures. We conducted three hierarchical re-
Method gression analyses (on the complete sample and by gender) using
SPSS for Windows (Version 19). In addition to the Big Five and
Participants and procedure. The original sample included the overall EI regression parameters over PWB, we analyzed and
502 Spanish undergraduate students who were enrolled in psychol- interpreted the percentages of variance explained and the semipar-
ogy and health science courses, volunteered for this study, and tial correlations of PWB with overall EI while controlling for the
earned extra course credit for participating. Of this original sam- Big Five dimensions. According to Cohen (1988), in the social and
ple, 465 undergraduate students (74% women) were aged 18 to 52 behavioral sciences, a large effect explains 25% of the variance, a
years (M ⫽ 21.31, SD ⫽ 4.94); they took the tests at two separate medium effect is approximately 9%, and a small effect is approx-
times, which were made available for later analyses. The question- imately 1%. To correctly interpret the semipartial correlations, we
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

naires were completed in paper-and-pencil format in class under followed the recommendation of Hunsley and Meyer (2003). Ac-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

the supervision of research assistants. The MSCEIT and a measure cording to these authors, semipartial correlations are a very suit-
of the Big Five personality traits were completed at Time 1 able measure of a variable’s contribution to criterion prediction.
(approximately 1 month into the term); 2 months later, the partic- Values from .15 to .20 represent an important contribution to the
ipants were tested on a measure of PWB. prediction, taking into account both the number of variables al-
Measures. ready included and the relationships between them.
Emotional intelligence test. The Spanish version of the MS-
CEIT v. 2.0 (Extremera & Fernández-Berrocal, 2009; Mayer et al.,
Results
AQ: 25 2002) was administered.
BFI-44 (Benet-Martínez & John, 1998). The BFI-44 is a For our regression equations, we first included the Big Five
44-item, self-report inventory designed to assess the Big Five dimensions and then the total MSCEIT (see Table 5). T5
factors of personality: Extraversion, Agreeableness, Conscien- In the regression model, the five personality variables explained
tiousness, Neuroticism, and Openness to Experience. The BFI-44 a relevant percentage of the variance. The variable with the most
scales have shown substantial internal consistency, retest reliabil- weight was Extraversion, followed by Neuroticism (for men, the
ity, and clear factor structure. We used the Spanish version of the inverse was found). The third variable with the most weight, EI,
BFI-44, which has psychometric properties that are similar to those was entered in the second step.
in the English version (Benet-Martínez & John, 1998). In this The increase in the percentage of variance explained by adding
study (Study 3), we found Cronbach alphas of .82, .64, .76, .80, EI is significant. The semipartial correlation of PWB with the
and .75 for each of the Big Five dimensions of Extraversion, overall EI, controlling for the Big Five dimensions, was .18 for the
Agreeableness, Conscientiousness, Neuroticism, and Openness to whole sample, .19 for men, and .17 for women. As EI enters
Experience, respectively. the regression equation in second place after the Big Five vari-
PWB test (Ryff, 1989). This scale measures six dimensions of ables, its contribution to predicting PWB may be considered sub-
PWB: self-acceptance, environmental mastery, purpose in life, stantial and relevant. We also took into consideration that for the
positive relations with others, personal growth, and autonomy. In whole sample, the only effect of MSCEIT on PWB was 6%
this study, we used the shorter version of the Spanish scale of PWB (.252 ⫽ 0.06), which can be considered a medium-sized effect.

Table 5
Hierarchical Regression Model for Predicting Well-Being (PWB, Time 2, After 2 Months) Based
on the Big Five Dimensions and the Total EI Score (Big Five and EI Measured at Time 1)

All (N ⫽ 465) Men (n ⫽ 120) Women (n ⫽ 345)

Models ␤ R 2
⌬R 2
␤ R 2
⌬R 2
␤ R2 ⌬R2

Step 1: Big Five traits .25 .25ⴱⴱⴱ .19 .19ⴱⴱⴱ .29 .29ⴱⴱⴱ
Extraversion .31ⴱⴱⴱ .23ⴱ .33ⴱⴱⴱ
Agreeableness .04 ⫺.06 .08
Conscientiousness .14ⴱⴱ .01 .17ⴱⴱ
Neuroticism ⫺.22ⴱⴱⴱ ⫺.25ⴱ ⫺.23ⴱⴱⴱ
Openness .10ⴱ .18 .08
Step 2: Big Five ⫹ EI .28 .03ⴱⴱⴱ .22 .03ⴱ .32 .03ⴱⴱⴱ
Extraversion .30ⴱⴱⴱ .21ⴱ .32ⴱⴱⴱ
Agreeableness .01 ⫺.10 .05
Conscientiousness .14ⴱⴱ .01 .18ⴱⴱⴱ
Neuroticism ⫺.22ⴱⴱⴱ ⫺.23ⴱ ⫺.22ⴱⴱⴱ
Openness .10ⴱ .17 .08
Total EI .18ⴱⴱⴱ .20ⴱ .18ⴱⴱⴱ
Note. • • •.
ⴱ ⴱⴱ ⴱⴱⴱ
p ⬍ .05. p ⬍ .01. p ⬍ .001.
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

SPANISH MSCEIT: PSYCHOMETRIC PROPERTIES 9

Discussion and Salovey’s (1997) hierarchical model with eight tasks, four
branches, two areas, and an overall EI score. Moreover, multi-
Beyond the predictive ability of the personality traits as mea- group analyses supported MSCEIT’s cross-gender invariance.
sured by the Big Five, the MSCEIT’s ability to predict PWB The last analysis in Study 1 and those in Studies 2 and 3 referred
provides empirical evidence that the Spanish version of the MS- to a different source for evidence of validity based on the relation-
CEIT measures a construct that is related to other psychological ship with other variables. This approach involves the degree to
variables. This would be expected according to the original theo- which these relationships are consistent with the construct under-
retical models, which we should interpret as evidence of the scale’s lying the proposed test interpretations (American Educational Re-
validity. As some researchers have noted, in the study of new
search Association et al., 2014). Thus, we found that women
theoretical constructs, such as EI, even findings that account for
scored higher than men on the Spanish version of the MSCEIT, as
small amounts of variance independent of other well-known per-
was also the case in the original English version (Brackett et al.,
sonality variables should be viewed as reasonable contributions to
2004, 2006; Mayer et al., 2002). Furthermore, the Spanish version
the understanding of the mechanism involved in human function-
leads to scores that are differentiable from instruments generated
ing (Mayer et al., 2000).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

from the trait model and other scales that measure personality
Other studies based on the original version of the MSCEIT
This document is copyrighted by the American Psychological Association or one of its allied publishers.

variables (Brackett et al., 2006; Webb et al., 2013). Finally, our

provide empirically similar findings. For example, the studies by
data suggest that the scores from the Spanish version of the
Brackett et al. (2006) and Higgs and Dulewicz (2008), as well as
MSCEIT predict measurements of well-being reasonably well, as
the review by Carmeli, Yitzhak-Halevy, and Weisberg (2009),
did the original English version (Brackett et al., 2006).
show a close relationship between EI and PWB.
Despite the interesting contributions of this study, several lim-
General Discussion itations must be mentioned. First, we demonstrated the relationship
of the MSCEIT to a series of personality variables. Nevertheless,
The MSCEIT, which has become one of the most widely used as we have already mentioned, the MSCEIT has also been used in
instruments for measuring EI as a human ability, has demonstrated other spheres, such as education and labor. Future studies should
excellent psychometric properties in normal and clinical samples provide evidence for the construct and predictive validity of scores
(Eack et al., 2009; Karim & Weisz, 2010; Mayer et al., 2000, 2003; on the Spanish version in these contexts. Second, because of the
Mayer, Salovey, & Caruso, 2012; Palmer et al., 2005). Current close relationship found between gender and MSCEIT scores, new
research on EI has shown that high MSCEIT scores are positively studies are needed to describe the test’s functioning in samples in
related to healthy personal and social functioning, and to academic which the proportion of men to women is more balanced or, if
and professional success in different contexts (see Mayer, Roberts, applicable, in samples that are predominantly men. Finally, a
& Barsade, 2008; Mayer, Salovey, & Caruso, 2008, for reviews). similar process should be completed for the age variable because,
Because of its wide applicability, the MSCEIT has been translated as the MSCEIT authors (Mayer et al., 2002, 2003) noted, EI is a
into different languages, including Spanish (Extremera & type of intelligence that changes and develops with age. The
Fernández-Berrocal, 2009). The usual practice is to follow the samples used in our study were predominantly young, and thus
International Test Commission (2005) guidelines, providing sev- these results would have to be replicated with older population
eral different tests of agreement of the original and translated samples to check their evolution throughout the human life cycle.
items. However, the evidence of validity based on test content is In summary, the factor structure and the psychometric properties
insufficient, and there are others that also provide justification of scores on the Spanish version of MSCEIT are similar to those
for the use of certain tests and decision making based on test of the original English version and support its use for assessing EI
scores. It is therefore necessary to provide more evidence abilities in the Spanish population.
justifying MSCEIT use and supporting the inferences that are
derived from its scores (American Educational Research Asso-
ciation, American Psychological Association, & National References
Council on Measurement in Education, 2014).
American Educational Research Association, American Psychological As-
The aim of this research was to analyze the factor structure and sociation, & National Council on Measurement in Education. (2014).
psychometric properties of the Spanish version of the MSCEIT Standards for educational and psychological testing. Washington, DC:
(Extremera & Fernández-Berrocal, 2009) by means of several Author.
reliability analyses of its scores and their usefulness, based on Benet-Martínez, V., & John, O. P. (1998). Los Cinco Grandes across
evidence of validity found from the internal structure of the scale cultures and ethnic groups: Multitrait multimethod analyses of the Big
or its relationships to other relevant variables. In three studies Five in Spanish and English. Journal of Personality and Social Psychol-
using large participant samples, our results have supported the use ogy, 75, 729 –750. http://dx.doi.org/10.1037/0022-3514.75.3.729
of the Spanish version of the MSCEIT as a suitable instrument for Bentler, P. M. (1995). EQS structural equations program manual. Encino,
measuring EI in the Spanish population. CA: Multivariate Software.
Blickle, G., Momm, T. S., Liu, S., Witzki, A., & Steinmayr, R. (2011).
In Study 1, we demonstrated close convergence between the
Construct validation of the test of emotional intelligence (TEMINT): A
Spanish consensus method and the original English scoring
two-study investigation. European Journal of Psychological Assess-
method that makes use of both general and expert consensus ment, 27, 282–289. http://dx.doi.org/10.1027/1015-5759/a000075
(Mayer et al., 2002, 2003) and found that scores on the MSCEIT Bollen, K. A. (1989). Structural equations with latent variables. New
showed excellent psychometric properties with high reliability and York, NY: Wiley. http://dx.doi.org/10.1002/9781118619179
test–retest stability. Using CFA analysis, we also confirmed that Brackett, M. A., & Mayer, J. D. (2003). Convergent, discriminant, and
the factor structure of the Spanish version is congruent with Mayer incremental validity of competing measures of emotional intelligence.
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

10 SANCHEZ-GARCIA, EXTREMERA, AND FERNANDEZ-BERROCAL

Personality and Social Psychology Bulletin, 29, 1147–1158. http://dx Extremera, N., Ruiz-Aranda, D., Pineda-Galan, C., & Salguero, J. M.
.doi.org/10.1177/0146167203254596 (2011). Emotional intelligence and its relation with hedonic and eudai-
Brackett, M. A., Mayer, J. D., & Warner, R. M. (2004). Emotional monic well-being: A prospective study. Personality and Individual Dif-
intelligence and its relation to everyday behaviour. Personality and ferences, 51, 11–16. http://dx.doi.org/10.1016/j.paid.2011.02.029
Individual Differences, 36, 1387–1402. http://dx.doi.org/10.1016/ Fan, H., Jackson, T., Yang, X., Tang, W., & Zhang, J. (2010). The factor
S0191-8869(03)00236-8 structure of the Mayer–Salovey–Caruso emotional intelligence test V 2.0
Brackett, M. A., Rivers, S. E., & Salovey, P. (2011). Emotional intelli- (MSCEIT): A meta-analytic structural equation modeling approach.
gence: Implications for personal, social, academic, and workplace suc- Personality and Individual Differences, 48, 781–785. http://dx.doi.org/
cess. Social and Personality Psychology Compass, 5, 88 –103. http://dx 10.1016/j.paid.2010.02.004
.doi.org/10.1111/j.1751-9004.2010.00334.x Fernández-Berrocal, P., Cabello, R., Castillo, R., & Extremera, N. (2012).
Brackett, M. A., Rivers, S. E., Shiffman, S., Lerner, N., & Salovey, P. Gender differences in emotional intelligence: The mediating effect of
(2006). Relating emotional abilities to social functioning: A comparison age. Behavioral Psychology/Psicología Conductual, 20, 77– 89.
of self-report and performance measures of emotional intelligence. Jour- Fernandez-Berrocal, P., Extremera, N., & Ramos, N. (2004). Validity and
nal of Personality and Social Psychology, 91, 780 –795. http://dx.doi reliability of the Spanish modified version of the Trait Meta-Mood
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

.org/10.1037/0022-3514.91.4.780 Scale. Psychological Reports, 94(3, Pt. 1), 751–755. http://dx.doi.org/

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Brackett, M. A., & Salovey, P. (2006). Measuring emotional intelligence 10.2466/pr0.94.3.751-755

with the Mayer-Salovery-Caruso emotional intelligence test (MSCEIT). Gardner, K. J., & Qualter, P. (2011). Factor structure, measurement invari-
Psicothema, 18(Suppl.), 34 – 41. ance and structural invariance of the MSCEIT v2.0. Personality and
Carmeli, A., Yitzhak-Halevy, M., & Weisberg, J. (2009). The relationship Individual Differences, 51, 492– 496. http://dx.doi.org/10.1016/j.paid
between emotional intelligence and psychological wellbeing. Journal of .2011.05.004
Managerial Psychology, 24, 66 –78. http://dx.doi.org/10.1108/ Gignac, G. E. (2005). Evaluating the MSCEIT V2.0 via CFA: Comment on
02683940910922546 Mayer et al. Emotion, 5, 233–235, 2003. http://dx.doi.org/10.1037/1528-
Cattell, R. B., & Burdsal, C. A., Jr. (1975). The radial parcel double 3542.5.2.233
factoring design: A solution to the item-vs-parcel controversy. Multi- Goldenberg, I., Matheson, K., & Mantler, J. (2006). The assessment of
variate Behavioral Research, 10, 165–179. http://dx.doi.org/10.1207/ emotional intelligence: A comparison of performance-based and self-
s15327906mbr1002_3 report methodologies. Journal of Personality Assessment, 86, 33– 45.
Chen, F., Bollen, K. A., Paxton, P., Curran, P. J., & Kirby, J. B. (2001). http://dx.doi.org/10.1207/s15327752jpa8601_05
Improper solutions in structural equation models: Causes, consequences, Higgs, S., & Dulewicz, V. (2008). Emotional intelligence, well-being and
and strategies. Sociological Methods & Research, 29, 468 –508. http:// personality: An empirical study of their interrelationship. (Working
dx.doi.org/10.1177/0049124101029004003 Paper Series- Management, M-08 – 08). University of Southampton,
Chen, F. F., Sousa, K. H., & West, S. G. (2005). Testing measurement Southampton, UK.
invariance of second-order factor models. Structural Equation Model- Hughey, J., & Burdsal, C. (1982). 16 PF-E structure using radial parcels
ing, 12, 471– 492. http://dx.doi.org/10.1207/s15328007sem1203_7 “versus” items. Journal of General Psychology, 107, 107–119. http://dx
Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit .doi.org/10.1080/00221309.1982.9709913
indexes for testing measurement invariance. Structural Equation Mod- Hunsley, J., & Meyer, G. J. (2003). The incremental validity of psycho-
eling, 9, 233–255. http://dx.doi.org/10.1207/S15328007SEM0902_5 logical testing and assessment: Conceptual, methodological, and statis-
Cohen, J. (1988). Statistical power analysis for the behavioral sciences tical issues. Psychological Assessment, 15, 446 – 455. http://dx.doi.org/
(2nd ed.). Hillsdale, NJ: Erlbaum. 10.1037/1040-3590.15.4.446
Davies, M., Stankov, L., & Roberts, R. D. (1998). Emotional intelligence: Iliescu, D., Ilie, A., Ispas, D., & Ion, A. (2013). Examining the psycho-
In search of an elusive construct. Journal of Personality and Social metric properties of the Mayer-Salovey-Caruso Emotional Intelligence
Psychology, 75, 989 –1015. http://dx.doi.org/10.1037/0022-3514.75.4 Test: Findings from an eastern European culture. European Journal of
.989 Psychological Assessment, 29, 121–128. http://dx.doi.org/10.1027/1015-
Day, A. L., & Carroll, S. A. (2004). Using an ability-based measure of 5759/a000132
emotional intelligence to predict individual performance, group perfor- International Test Commission. (2005). International guidelines on test
mance, and group citizenship behaviours. Personality and Individual adaptation. Retrieved from http://www.intestcom.org AQ: 27
Differences, 36, 1443–1458. http://dx.doi.org/10.1016/S0191- Joseph, D. L., & Newman, D. A. (2010). Emotional intelligence: An
8869(03)00240-X integrative meta-analysis and cascading model. Journal of Applied Psy-
Eack, S. M., Greeno, C. G., Pogue-Geile, M. F., Newhill, C. E., Hogarty, chology, 95, 54 –78. http://dx.doi.org/10.1037/a0017286
G. E., & Keshavan, M. S. (2010). Assessing social-cognitive deficits in Karim, J., & Weisz, R. (2010). Cross-cultural research on the reliability
schizophrenia with the Mayer-Salovey-Caruso Emotional Intelligence and validity of the Mayer-Salovey-Caruso emotional intelligence test
Test. Schizophrenia Bulletin, 36, 370 –380. http://dx.doi.org/10.1093/ (MSCEIT). Cross-Cultural Research: The Journal of Comparative So-
schbul/sbn091 cial Science, 44, 374 – 404. http://dx.doi.org/10.1177/106939711
Eack, S. M., Pogue-Geile, M. F., Greeno, C. G., & Keshavan, M. S. (2009). 0377603
Evidence of factorial variance of the Mayer-Salovey-Caruso Emotional Keele, S. M., & Bell, R. C. (2009). Consensus scoring, correct responses
Intelligence Test across schizophrenia and normative samples. Schizo- and reliability of the MSCEIT V2. Personality and Individual Differ-
phrenia Research, 114, 105–109. http://dx.doi.org/10.1016/j.schres ences, 47, 740 –747. http://dx.doi.org/10.1016/j.paid.2009.06.013
.2009.05.011 Kline, R. B. (2005). Principles and practice of structural equation mod-
Extremera, N., & Fernández-Berrocal, P. (2009). CEIT, Test de Inteligen- eling. New York, NY: Guilford Press.
cia Emocional de Mayer-Salovey-Caruso [●●●]. Madrid, Spain: TEA Kong, D. T. (2014). Mayer-Salovey-Caruso Emotional Intelligence Test
AQ: 26 Ediciones.MS (MSCEIT/MEIS) and overall, verbal, and nonverbal intelligence: Meta-
Extremera, N., Fernández-Berrocal, P., & Salovey, P. (2006). Spanish analytic evidence and critical contingencies. Personality and Individual
version of the Mayer-Salovey-Caruso Emotional Intelligence Test (MS- Differences, 66, 171–175. http://dx.doi.org/10.1016/j.paid.2014.03.028
CEIT). Version 2.0: Reliabilities, age and gender differences. Psico- Koven, N. S., & Max, L. K. (2014). Basal salivary oxytocin level predicts
thema, 18(Suppl.), 42– 48. extra- but not intra-personal dimensions of emotional intelligence. Psy-
 APA NLM
tapraid5/z1t-assess/z1t-assess/z1t00415/z1t3021d15z xppws S⫽1 12/4/15 2:55 Art: 2015-0351

SPANISH MSCEIT: PSYCHOMETRIC PROPERTIES 11

choneuroendocrinology, 44, 20 –29. http://dx.doi.org/10.1016/j tional Intelligence Test (MSCEIT) with established emotions measures.
.psyneuen.2014.02.018 Emotion, 6, 663– 669. http://dx.doi.org/10.1037/1528-3542.6.4.663
MacCann, C., & Roberts, R. D. (2008). New paradigms for assessing Roberts, R. D., Zeidner, M., & Matthews, G. (2001). Does emotional
emotional intelligence: Theory and data. Emotion, 8, 540 –551. http:// intelligence meet traditional standards for an intelligence? Some new
dx.doi.org/10.1037/a0012746 data and conclusions. Emotion, 1, 196 –231. http://dx.doi.org/10.1037/
Mayer, J. D., Caruso, D. R., & Salovey, P. (1999). Emotional intelligence 1528-3542.1.3.196
meets traditional standards for an intelligence. Intelligence, 27, 267–298. Rode, J. C., Mooney, C. H., Arthaud-day, M., Near, J. P., Rubin, R. S.,
http://dx.doi.org/10.1016/S0160-2896(99)00016-1 Baldwin, T. T., & Bommer, W. H. (2008). An examination of the
Mayer, J. D., Roberts, R. D., & Barsade, S. G. (2008). Human abilities: structural, discriminant, nomological, and incremental predictive valid-
Emotional intelligence. Annual Review of Psychology, 59, 507–536. ity of the MSCEIT V2.0. Intelligence, 36, 350 –366. http://dx.doi.org/
http://dx.doi.org/10.1146/annurev.psych.59.103006.093646 10.1016/j.intell.2007.07.002
Mayer, J. D., & Salovey, P. (1997). What is emotional intelligence? In P. Rossen, E., & Kranzler, J. H. (2009). Incremental validity of Mayer-
Salovey & D. Sluyter (Eds.), Emotional development and emotional Salovey-Caruso Emotional Intelligence Test version 2.0 (MSCEIT) after
intelligence: Educational implications (pp. ●●●–●●●). New York, NY: controlling for personality and intelligence. Journal of Research in
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

AQ: 28 Basic Books. Personality, 43, 60 – 65. http://dx.doi.org/10.1016/j.jrp.2008.12.002

Mayer, J. D., Salovey, P., & Caruso, D. R. (2000). Models of emotional
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Rossen, E., Kranzler, J. H., & Algina, J. (2008). Confirmatory factor

intelligence. In R. J. Sternberg (Ed.), Handbook of intelligence (pp. analysis of the Mayer-Salovey-Caruso Emotional Intelligence Test V 2.0
396 – 420). Cambridge, UK: Cambridge University Press. http://dx.doi (MSCEIT). Personality and Individual Differences, 44, 1258 –1269.
.org/10.1017/CBO9780511807947.019 http://dx.doi.org/10.1016/j.paid.2007.11.020
Mayer, J. D., Salovey, P., & Caruso, D. (2002). Mayer–Salovey–Caruso Ryff, C. D. (1989). Happiness is everything, or is it? Explorations on the
Emotional Intelligence Test (MSCEIT© V2.0): User’s manual. Toronto, meaning of psychological well-being. Journal of Personality and Social
ON: Multi-Health Systems. Psychology, 57, 1069 –1081. http://dx.doi.org/10.1037/0022-3514.57.6
Mayer, J. D., Salovey, P., & Caruso, D. R. (2008). Emotional intelligence: .1069
New ability or eclectic traits? American Psychologist, 63, 503–517. Salguero, J. M., Extremera, R., & Fernández-Berrocal, P. (2012). Emo-
http://dx.doi.org/10.1037/0003-066X.63.6.503
tional intelligence and depression: The moderator role of gender. Per-
Mayer, J. D., Salovey, P., & Caruso, D. R. (2012). The validity of the
sonality and Individual Differences, 53, 29 –32. http://dx.doi.org/10
MSCEIT: Additional analyses and evidence. Emotion Review, 4, 403–
.1016/j.paid.2012.02.006
408. http://dx.doi.org/10.1177/1754073912445815
Salovey, P., & Mayer, J. D. (1990). Emotional intelligence. Imagination,
Mayer, J. D., Salovey, P., Caruso, D. R., & Sitarenios, G. (2003). Mea-
Cognition and Personality, 9, 185–211. http://dx.doi.org/10.2190/
suring emotional intelligence with the MSCEIT V2.0. Emotion, 3, 97–
DUGG-P24E-52WK-6CDG
105. http://dx.doi.org/10.1037/1528-3542.3.1.97
Salovey, P., Mayer, J. D., Goldman, S. L., Turvey, C., & Palfai, T. P.
McCrae, R. R. (2000). Emotional intelligence from the perspective of the
(1995). Emotional attention, clarity, and repair: Exploring emotional
five-factor model of personality. In R. Bar-On & J. D. A. Parker (Eds.),
intelligence using the Trait Meta-Mood Scale. In J. W. Pennebaker (Ed.),
The handbook of emotional intelligence (pp. ●●●–●●●). San Francisco,
Emotion, disclosure, and health (pp. 125–151). Washington, DC: Amer-
AQ: 29 CA: Jossey-Bass.
ican Psychological Association.
Meredith, W. (1993). Measurement invariance, factor analysis and factorial
van Dierendonck, D., Díaz, D., Rodríguez-Carvajal, R., Blanco, A., &
invariance. Psychometrika, 58, 525–543. http://dx.doi.org/10.1007/
BF02294825 Moreno-Jiménez, B. (2008). Ryff’s six-factor model of psychological
O’Connor, B. P. (2000). SPSS and SAS programs for determining the well-being, a Spanish exploration. Social Indicators Research, 87, 473–
number of components using parallel analysis and Velicer’s MAP test. 479. http://dx.doi.org/10.1007/s11205-007-9174-7
Behavior Research Methods, Instruments & Computers, 32, 396 – 402. Van Rooy, D. L., Viswesvaran, C., & Pluta, P. (2005). An evaluation of
http://dx.doi.org/10.3758/BF03200807 construct validity: What is this thing called emotional intelligence?
Palmer, B. R., Gignac, G., Manocha, R., & Stough, C. (2005). A psycho- Human Performance, 18, 445– 462. http://dx.doi.org/10.1207/
metric evaluation of the Mayer-Salovey-Caruso emotional intelligence s15327043hup1804_9
test version 2.0. Intelligence, 33, 285–305. http://dx.doi.org/10.1016/j Warwick, J., & Nettelbeck, T. (2004). Emotional intelligence is . . .?
.intell.2004.11.003 Personality and Individual Differences, 37, 1091–1100. http://dx.doi
Petrides, K. V., & Furnham, A. (2001). Trait emotional intelligence: .org/10.1016/j.paid.2003.12.003
Psychometric investigation with reference to established trait taxono- Webb, C. A., Schwab, Z. J., Weber, M., DelDonno, S., Kipman, M.,
mies. European Journal of Personality, 15, 425– 448. http://dx.doi.org/ Weiner, M. R., & Killgore, W. D. S. (2013). Convergent and divergent
10.1002/per.416 validity of integrative versus mixed model measures of emotional intel-
Rast, P., Zimprich, D., Van Boxtel, M., & Jolles, J. (2009). Factor structure ligence. Intelligence, 41, 149 –156. http://dx.doi.org/10.1016/j.intell
and measurement invariance of the cognitive failures questionnaire .2013.01.004
across the adult life span. Assessment, 16, 145–158. http://dx.doi.org/10 Wright, A. G., Lukowitsky, M. R., Pincus, A. L., & Conroy, D. E. (2010).
.1177/1073191108324440 The higher order factor structure and gender invariance of the Patho-
Rey, L., & Extremera, N. (2014). Positive psychological characteristics and logical Narcissism Inventory. Assessment, 17, 467– 483. http://dx.doi
interpersonal forgiveness: Identifying the unique contribution of emo- .org/10.1177/1073191110373227
tional intelligence abilities, Big Five traits, gratitude and optimism.
Personality and Individual Differences, 68, 199 –204. http://dx.doi.org/
10.1016/j.paid.2014.04.030 Received July 8, 2015
Roberts, R. D., Schulze, R., O’Brien, K., MacCann, C., Reid, J., & Maul, Revision received November 16, 2015
A. (2006). Exploring the validity of the Mayer-Salovey-Caruso Emo- Accepted November 17, 2015 䡲