Professional Documents
Culture Documents
Incremental Validity
Nancy C. Sherman
V i l l a Julie College
Joseph E. G. Williams
Note. Correspondence regarding this paper should be directed to Dr. Ralph L. Piedmont,
Department o f Pastoral Counseling, Loyola College i n Maryland, 8890 M c G a w Raod,
Suite 380, Columbia, M D 21045 or via e-mail at: rpiedmont@loyola.edu.
Measuring Emotional Intelligence 2
(1995) seminal work. The definition o f emotional intelligence (EI) has expanded over
own and others' disposition; and mood management or control over emotions. Bar-On
and Parker (2005) have asserted that Emotional intelligence pertains to the emotional,
Empirically, questions have arisen over E l ' s construct validity. Cherniss et al. (2006)
Assessment ( M E I A ; Tett, Fox, & Wang, 2005) is a next generation measure o f E I that
aims to capture its multidimensional aspects while demonstrating its distinctiveness from
personality. The purpose o f this report was to examine the construct and incremental
validity o f the M E I A .
Method
Participants consisted o f a 281 adults (67 men and 216 women), ranging in age 18
years old to 42 years old ( M =19.58, SD = 3.14). Ethnic composition o f the sample
consisted o f 84.2 % Caucasian, 9.2 % African-American, 2.5 % Hispanic, 2.5 Other 0.4
(1965) Self-esteem Scale; the Life Orientation Test -Revised (LOT-R) was used to
measure optimism (Carver & Scheier, 2003); the State Hope Scale (Snyder et al., 1996).
Results
Table 1 provides descriptive statistics for the M E I A and alpha reliabilities for
each scale. Two points o f interest here. First, gender differences are observed on several
scales, effects not examined during the development o f this instrument. Second, alpha
reliabilities are consistent w i t h , although somewhat lower than, those reported by Tett et
al. (2005).
single dimension was extracted. This first unrotated factor explained 3 2 % o f the total
variance, and 9 o f the ten scales loaded significantly on the dimension (e.g., > .30); only
Regulate Emotions in Self did not load (eigenvector = .27). This suggests that the scales
evidence substantial overlap. This raises the question, " D o the M E I A scales have
variance was broken down into four components: reliable variance (denoted b y a), error
variance (1 - a), common variance (which represents the amount o f shared variance each
regressing each M E I A scale on the remaining subscales), and uniquely reliable variance,
w i t h the other subscales between 15% and 48% o f their variance ( M = 35%). Kaufman
(1975) suggested that for a scale to warrant specific interpretation, the uniquely reliable
variance should be greater than the error variance and should account for at least one
quarter o f the total variance. Inspecting Table 2 reveals that all o f the M E I A scales meet
this criterion. Thus, despite a significant amount o f overlap, these scales contain
Construct Validity
Table 3 presents the correlations between the M E I A scales and the F F M domains.
There are numerous associations across the two instruments. The last column summarizes
the extent to which each M E I A scale overlaps w i t h the five personality domains. Overlap
ranges from 4 % shared variance to 53% (Mean = 25%). The last row o f the table presents
the extent to which each o f the five personality domains are involved across the 10 M E I A
scales. Personality shares between 32% (Agreeableness) and 53% (Openness) w i t h these
presented by Tett et al. (2005), 31 o f the 50 correlations (62%) are replicated here.
Incremental Validity
The above analyses indicate that the M E I A scales share considerable variance in
common w i t h personality, which raises the question, " D o the M E I A scales have
incremental validity over and above established measures o f personality to justify their
Measuring Emotional Intelligence 5
were performed, using each o f the outcome variables as the dependent variable. On Step
1 o f the regression, the F F M domains were entered as a block. On Step 2, using forward
entry, the M E I A scales were tested to determine whether any continued to be significant
predictors. The results o f these analyses are presented in Table 5. As can be seen, in each
instance at least one o f the M E I A scales evidence significant incremental validity over
Discussion
Overall, these results provide psychometric support for the new M E I A scales. The
constructs, even after controlling for the predictive influence o f personality. Although the
M E I A scales have considerable overlap among them, each scale contains sufficient
unique variance to warrant their individual interpretation. The existence o f scales that are
independent o f the F F M domains opens the door to the potential identification o f new
References
Bar-On, R., & Parker, J.D.A. (2005). Baron Emotional Quotient Inventory: Youth
Carver, C. S., & Scheier, M . (2003). Optimism. In S J . Lopez & C. R. Snyder (Eds.),
4 1 , 239-245.
Diener, E., Emmons, R. A., Larsen, R. J., & Griffins, S. (1985). The Satisfaction w i t h
Kaufman, A . S. (1975). Factor analysis o f the WISC-R at 11 age levels between 6 lA and
McCrae, R. R., & Costa, P. T., Jr. (1985). Updating Norman's "Adequate Taxonomy":
Measuring Emotional Intelligence 7
University Press.
Snyder, C. R., Harris, C , Anderson, J. R., Holleran, S. A., Irving, L. M . , Sigmon, S. T.,
Yoshinobu, L., Gibb, J., Langelle, C , & Harney, P. (1991). The w i l l and the
Tett, R. P., Fox, K. E., & Wang, A . (2005). Development and validation o f a self-report
Table 1
Men Women
(n = 67) fa = 216)
M E I A SCALE M SD M SD a
Non-Verbal Emotions
3.79 .75 4.21 .72 4.01** .77
Expression
Table 2
Variance Elements
M E I A Scale a R2* a -R2 1-a
Recognize Emotions i n Self .76 .48 .28 .24
Regulate Emotions i n Self .83 .33 .50 .17
Recognize Emotions i n Others .80 .41 .39 .20
Regulate Emotions i n Others .76 .47 .29 .24
Non- verbal Emotion Expression .77 .48 .29 .23
a r>2
R values are based on each MEIA scale's overlap with the remaining M E I A Scales.
Measuring Emotional Intelligence 10
Table 3
Correlations between the Multidimensional Emotional Intelligence Assessment Subscales
and the Domains o f the Five-Factor Model.
MEIA
Scale Neuroticism Extraversion Openness Agreeableness Conscientiousness R2
Kecognize _.07
Cil . 2029**** . 3a0n**** ns
.08 on**
.20*
Emotions in
Others
Mood 04*
I * , * .05 .06 .16** .03 .09
Redirected
Attention
Table 4
Correlations between the MEIA Scales and Measures ofLife Satisfaction, Psychological Maturity, and Positive Psychological Status.
** p < . 0 1 ; ***/?<.001.