Content uploaded by Karma El-Hasan
Author content
All content in this area was uploaded by Karma El-Hasan on Oct 30, 2014
Content may be subject to copyright.
Adapting and Validating the BarOn
EQ–i:YV in the Lebanese Context
Karma El Hassan and Maliha El Sader
Office of Institutional Research & Assessment
American University of Beirut, Lebanon
The purpose of this study was to adapt and validate the BarOn Emotional Quotient
Inventory: Youth Version (BarOn EQ–i:YV), a measure of emotional intelligence, in
the Lebanese context. The scale was translated and adapted into Arabic using a 3-step
process involving judgmental, logical, and empirical methods. It was then adminis-
tered to a representative sample of 433 students in the 10- to 18-year-old age group.
Psychometric properties of the Arabic version in terms of internal consistency,
test–retest reliabilities, factor structure, and scale intercorrelations was investigated.
Gender and age differences were also studied. The reliability of the adapted test, both
internal and over time, was satisfactory. Exploring scale factorial validity identified 2
major factors, whereas exploring it by items revealed findings particular to the Leba-
nese situation. Reported age and gender differences were similar to those on the orig-
inal test except for gender differences on Intrapersonal and Adaptability scales.
Explanations and propose recommendations for future research are provided.
Key words: test adaptation, validation, emotional intelligence
Emotional intelligenceis emerging as an important area of inquiry for research in ed
-
ucation and psychology. The number of scientific journal articles that have recently
been published is evidence of its popularity. The construct is defined as an array of
emotional, personal, and interpersonal abilities thatinfluence one’soverall ability to
cope with environmental demands and pressures (Bar-On & Parker, 2000). Mayor
and Salovey (as cited in Finegan, 1998) defined emotional intelligence as the ability
to perceive, appraise, and express emotions; the ability to access and/or generate
INTERNATIONAL JOURNAL OF TESTING, 5(3), 301–317
Copyright © 2005, Lawrence Erlbaum Associates, Inc.
Correspondence should be addressed to Karma El Hassan, Office of Institutional Research & As
-
sessment, American University of Beirut, Bliss St., P.O. Box 11–0236, Beirut, 1107 2020, Lebanon.
E-mail: kelhasan@aub.edu.lb
Do Not Copy
feelings when they facilitate thought; and the ability to regulate emotions to promote
emotional and intellectual growth. Several authors have argued that emotional intel
-
ligence affects health (Schilling, 1996) and is essential for success in academics as
well as life in general (Finegan, 1998; Gibbs, 1995; Goleman, 1995). In addition,
Goleman (as cited in in Oester & Oester, 1997) argued that emotional competency
can be taught and internalized anytime during ones’ life cycle.
To validate these arguments, it is necessary to have reliable and valid instru
-
ments to measure emotional intelligence. Because of the recency of the concept
and the fact that emotional intelligence deals with abstract abilities, very few mea
-
surement tools exist that can measure it reliably and validly. One such instrument
is the Bar-On Emotional Bar Inventory (BarOn EQ–i) with its Youth Version
(BarOn EQ–i:YV; Bar-On & Parker, 2000). The BarOn EQ–i:YV has many excep
-
tional features that other instruments that measure emotional intelligence do not
have (large normative base, research in different countries, history, age group,
etc.). The inventory is a 60-item self-report designed for use with school-aged
youths, 7 to 18 years old. It is a multidimensional scale consisting of five subscales
that probe an individual’s Interpersonal, Intrapersonal, Stress Management, and
Adaptability abilities in addition to General Mood.
Our purposes in this study were to adapt and validate the BarOn EQ–i:YV in a
Lebanese context. The incidence of emotional disorders and problems among Leb-
anese youth is well confirmed by research and statistics. A study done by the Cen-
ter for Educational Research and Development (CERD; 1999) on a sample of 937
Lebanese students aged 11 to 18 years concluded that Lebanese school children
suffer from aggression, tension, and low self-appreciation. A recent survey (Zein,
2001) conducted in private schools in Lebanon showed that 23.3% of children are
involved in bully/victim problems. A large percentage of children have sleeping
problems because of tension they live under (CERD, 1999). In an article published
in the Revue Medicale Libanaise in 1990, Karam explained that behavior disorders
exhibited by Lebanese youth are the symptoms of posttraumatic stress disorder
(PTSD). Accordingly, The Institute for Development Research and Applied Care,
with the support of United Nations International Children’s Emergency Fund, the
Lebanese Ministry of Education, and the Lebanese National Council for Scientific
Research developed a threefold plan of action with the aim of decreasing the level
of PTSD among students and adolescents. Among the assessed children and ado
-
lescents in the sample, 12% had depression, 18% PTSD, 15% separation anxiety,
and 19% overanxious disorder (Traumatic Stress Points, 1997).
The long period of war and the increase in the percentage of working mothers
from 33% in 1984 to 41% in 1994 (Deeb, 1997) may have contributed to these occur
-
rences. Despitethe fact that the use of emotional training programs with emotionally
troubled children—bullies and children with PTSD—has proven to be effective
(Schilling, 1996), theyhave not yet beenintroduced inLebanon. Therefore, there is a
need for an instrument that measures emotional intelligence among the youth.
302
EL HASSAN AND EL SADER
Do Not Copy
The purposes of this study were to (a) adapt the BarOn EQ–i:YV, an instrument
that measures emotional intelligence, in a Lebanese context; (b) validate the instru
-
ment through obtaining construct validity evidence—factor analysis will provide
evidencefor the construct of emotional intelligence in Lebanese culture andwhether
it produces similar factors to those claimed by the literature and in the Western cul
-
tures; (c) investigate age and gender differences in the emotional quotient (EQ); (d)
provideevidencefor the reliability of results of the BarOnEQ–i:YV witha Lebanese
sample as shown by the stability overtime andthe internalconsistency;and (e)check
for equivalence of long and short forms of the BarOn EQ–i:YV.
Having such a reliable and valid instrument will have the following benefits: (a)
provide cross-cultural validation for the construct of emotional intelligence, (b)
provide educators and administrators with baseline data about level of emotional
intelligence among Lebanese youth, (c) help check the efficacy of emotional train
-
ing intervention strategies, (d) help design or adapt appropriate curriculum compo
-
nents to meet the needs of Lebanese youth, (e) help develop appropriate training
and remedial emotional literacy programs, and (f) help advance the research on
emotional intelligence.
Adapting and validating an already existing instrument is preferable to develop-
ing a new instrument, especially as the construct of emotional intelligence is very
complex. Adapting tests not only saves time and money but also allows for com-
parative studies across cultural and language groups (Hambleton, 1999). However,
the use of any test with a population other than that on which it was validated is un-
supported and inappropriate. A test is validated with respect to a certain population
(Cronbach & Drenth, 1972). It is very important to specify the reliability of the test
in terms of the targeted population to whom it is going to be administered. Simi-
larly, validity must be established with reference to the particular use for which the
test is being considered (Anastasi & Urbina, 1997). Therefore, there is a need for a
reliable and valid adapted measure of emotional intelligence in Lebanon to assess
the extent and type of emotional problems experienced by Lebanese youths.
METHOD
Adaptation and Translation of the BarOn EQ–i:YV
The BarOn EQ–i:YV was translated and adapted using the guidelines of the In
-
ternational Test Commission (2000). These guidelines address concerns about
test context, test development and adaptation, administration procedures, and
score interpretations. They require that test adaptation procedures should use
systematic judgmental evidence at every step. Adhering to standards will reduce
sources of error (Hambleton, 1999). A combination of forward and backward
translation designs was used. An experienced translator translated the BarOn
ADAPTING AND VALIDATING THE BARON 303
Do Not Copy
EQ–i:YV to Arabic and then both versions were given to a group of educators to
judge the equivalence of the translation. Some corrections were made, and the
first draft of the Arabic version was checked by an elementary Arabic language
coordinator, who has a master’s degree in Arabic language with 20 years experi
-
ence, to check for the adequacy of language used to the age group investigated.
Further editing was done before the test was subjected for the counterbalanced
design.
Bilingual participants of equal proficiency in both languages (N = 25) took the
BarOn EQ–i:YV and the Arabic BarOn EQ–i:YV (BarOn EQ–i:YV–A) according
to a counterbalanced design (Arabic–English, English–Arabic). Then we corre
-
lated the ratings for the two versions, and correlations between the English and
adapted Arabic versions are presented in Table 1 under first trial. Table 1 presents
correlations between two versions for all counterbalanced design trials that study
went through.
Because of low correlations between Arabic and English versions on some items
(n = 17 items), the design was repeated in another school (N = 22) with more clarifi-
cations given by Maliha El Sader about the items and the scales. The results of the
Arabic–Englishcorrelations arereported in Table 1under the SecondTrialcolumn.
The correlations were higher and more acceptable, but still some items had low
correlations (n = 8 items). For this reason and to reduce memory effects, another
translator back translated the test. Then, the two versions (The Arabic and the
back-translated English) were given to an elementary/middle level Arabic teacher,
a holder of a bachelor’s degree in Arabic Language and Literature, for evaluation.
There were 12 items that revealed problems such as difficult and unclear terms:
Items 2, 13, 14, 16, 19, 26, 28, 29, 34, 39, 55, and 59.
These items were subjected to further editing before we implemented a third
counterbalanced design (N = 32) on three age groups only this time: 10 tthrough
12, 13 through 15, and 16 through 18 years of age in a third school. The results
of the correlations can be seen in Table 1 under column 4. As evident from Table
1, the last revision improved the correlations and the test was ready for the pilot
study.
304
EL HASSAN AND EL SADER
TABLE 1
Correlations Between Adapted and English Versions on the Three Trials of
the Counter Balanced Design
Correlation Arabic/English First Trial Second Trial Third Trial
≥ .80 0 16 22
≥ .60 < .80 18 18 22
≥ .40 < .60 25 18 16
< .40 17 8 0
Note. Total = 60 items.
Do Not Copy
Pilot Study
The pilot study was done in a school on a sample of 48 students (24 females, 24
males) equally divided into four age groups: 7 through 9, 10 through 12, 13
through 15, and 16 through 18 years. The test was administered and scored as per
standardized administration procedures and the Inconsistency Index was checked
and proved to be acceptable (≤ 10). The Inconsistency Index is a validity measure
designed to assess response consistency. The Bar-On includes several pairs of
highly correlated items that have similar content that gauge response consistency.
The three age groups—10 through 12, 13 through 15, and 16 through 18
years—did not face any difficulty while doing the test, whereas the first age group
(7–9 years) had problems. They took more time filling it out (around 45–50 min),
asked a lot of questions related to words and to the instructions, and were confused
with the rating scales points range (1–4).
We calculated the internal reliability coefficient alpha for the whole scale, for
subscales, and for total EQ, and the coefficients ranged between r = .40
(Intrapersonal) to r = .81 (Adaptability). Test–retest reliabilities done on 20 stu-
dents (10 females, 10 males) over a 3-week period were high, ranging between r =
.93 to .98 for all subscales and for total EQ. In addition, 10% of the students were
given on both times the BarOn EQ–i:YV–A short version.
Moreover, the pilot study revealed first, that the test cannot be given to 7- to
9-year-olds. Classroom teachers felt that the language, concepts, Likert scale
style, and instructions were too sophisticated for 7-year-olds. Also, during the
pilot study, the 8-year-olds took 40 min to complete the test, and they raised a lot
of questions on some of the Arabic terms used. Other age groups did not face
similar problems with the test language. The Arabic language coordinator rec-
ommended not using the test with 7-year-olds and even 8-year-olds. Although
the test seemed long (60 items), almost all students enjoyed filling it in and
asked for more items.
Sample
We administered the final version of the adapted BarOn EQ–i:YV to a sample of
433 students (223 males, 210 females) randomly chosen from 12 schools: 6 public,
5 private, 1 orphanage. We used cluster sampling. We identified three main clus
-
ters: (a) coeducational private schools covering all grade levels, (b) coeducational
public schools covering elementary/middle grades because there are no public
schools that cover all grade levels, and (c) coeducational secondary public schools.
From every area in greater Beirut, 1 private and 1 public school were randomly
chosen by draw. As for the orphanages, there was only one in the Beirut area that
met the characteristics required. A total of 16 tests (11 females, 5 males) had to be
discarded for different reasons. Table 2 summarizes the breakdown of the final
sample by age, gender, and school type.
ADAPTING AND VALIDATING THE BARON 305
Do Not Copy
Administration and Scoring Procedures
The chosen schools were contacted ahead of time. I met with the principals and
gave a brief presentation on emotional intelligence and BarOn EQ–i:YV to the
teachers. Students were randomly selected taking into consideration age and
gender. Filling out the tests took approximately 20 to 30 min. Maliha El Sader
was available in all sessions to answer all questions. A total of 417 students
filled in the long version of the BarOn EQ–i:YV–A, and from each age group, a
male and a female student immediately filled in the short version to check for
equivalence of long and short forms. To determine reliability over time, 3 weeks
later 56 students (approximately 14%) of the sample filled in the BarOn
EQ–i:YV again.
Statistical Analysis
We investigated the factor structure of the BarOn EQ–i YV–A using exploratory
factor analysis (EFA) by performing a principal components analysis with
varimax rotation on subscales and on the items. EFA was used because we were
interested in deriving the factor structure or model of emotional intelligence in
Lebanon to generate a theory regarding its structure, as this had not been at
-
tempted before. In addition, we computed intercorrelations among subscales.
Moreover, we reported means for various subscales and for the total EQ, and we
investigated age and gender differences using analysis of variance (ANOVA).
We also determined the instrument’s internal consistency using coefficient alpha.
We determined the stability over time using Pearson product–moment correla
-
tion coefficient. We did not study content validity because it has already been es
-
tablished in the English version, and it is applicable to the Arabic one, as no ma
-
jor changes in content were done. To check for the validity of the scores, we
computed the Inconsistency Index score in addition to the General Mood and
Positive Impression indexes.
306
EL HASSAN AND EL SADER
TABLE 2
Sample Distribution by Age, Gender, and School Type
Age
10–12 Years 13–15 Years 16–18 Years
Type of School Female Male Female Male Female Male Total
Orphanage 7 7 7 6 3 5 35
Public 28 41 43 39 29 41 221
Private 28 23 30 37 24 19 161
Total 63 71 80 82 56 65 417
Do Not Copy
RESULTS
The final sample on which analysis was conducted consisted of 390 participants
(208 males, 182 females), as a number of tests (N = 27) were discarded due to the
Inconsistency indexes exhibited by some responses.
Reliability
Internal Reliability
Table 3 presents the internal reliability coefficients as measured with
Cronbach’s alpha for the various BarOn EQ–i:YV–A subscales and by gender and
age. Bar-On and Parker’s (2000) internal coefficients are presented between paren
-
theses. With respect to the whole sample, the internal consistency reliabilities
ranged between r = .51 to .80 for the long version and between r = .65 to .81 for the
short version.
Test–Retest Reliability
The stability of the BarOn EQ–i:YV–A was examined on a sample of 56 chil-
dren and adolescents using a 3-week, test–retest interval, and the coefficients
ranged between r = .70 to .80 with the exception of the General Mood and Positive
Impression scales reliabilities, which were r = .60.
ADAPTING AND VALIDATING THE BARON 307
TABLE 3
Internal Reliability Coefficient for BarOn EQ–i:YV–A (by Age and
Gender Groups)
BarOn EQ–i:YV–A Scale 10–12 13–15 16–18
Females
Intrapersonal .60 (.72) .70 (.81) .63 (.87)
Interpersonal .70 (.83) .73 (.83) .70 (.82)
Stress Management .60 (.85) .50 (.87) .60 (.87)
Adaptability .80 (.85) .80 (.87) .74 (.89)
Total EQ .82 (.90) .81 (.90) .69 (.89)
General Mood .70 (.88) .74 (.87) .70 (.90)
Males
Intrapersonal .50 (.72) .62 (.81) .62 (.83)
Interpersonal .61 (.83) .62 (.83) .80 (.82)
Stress Management .60 (.85) .50 (.87) .40 (.87)
Adaptability .80 (.85) .81 (.87) .82 (.90)
Total EQ .80 (.90) .80 (.90) .80 (.89)
General Mood .70 (.88) .73 (.87) .80 (.87)
Note. BarOn EQ–i:YV–A = Arabic Bar-On Emotional Quotient Inventory: Youth Version; EQ =
emotional quotient. Numbers in parenthesis are those of Bar-On.
Do Not Copy
Validity
Validity Checks
Inconsistency Index score.
The BarOn EQ–i:YV is provided with an In
-
consistency Index, a validity measure designed to assess response consistency. The
test includes several pairs of highly correlated items that have similar content that
gauge response consistency. The 10 pairs of the BarOn EQ–i:YV–A items used for
this index (2 pairs for each of the five main scales) and their intercorrelations are
reported in Table 4. Table 4 also provides intercorrelations reported by BarOn and
Parker (2000; between parentheses), and they are similar to those in the Arabic ver
-
sion except for Items 20 and 51, which were lower.
Calculating the absolute value of the difference in response to each pair and
then summing these 10 differences provides the Inconsistency Index score. The
most likely interpretation for a high score on the Inconsistency Index is that the
participant was unmotivated and responded haphazardly or that they were deliber
-
ately trying to distort their results for one reason or another. However, elevated
scores on the BarOn EQ–i:YV Inconsistency Index can also mean that the respon-
dent was experiencing difficulty understanding subtle differences between some
of the items. High scores on this index may also be indicative of poorly developed
insight and lack of self-awareness. Among the test takers (N = 417), 27 (4/35 or-
phans, 18/221 in public school, and 5/161 in private school) had Inconsistency in-
dexes above 10, and these were discarded.
General Mood scale.
Bar-On and Parker (2000) gave importance to the
overall mood level of respondents. Bar-On and Parker believed that General Mood
is another factor that needs to be taken into account when assessing an individual’s
true level of emotional intelligence. For this reason the scale results for individuals
with standard scores below 80 (36 raw score) on General Mood must be interpreted
with caution. Of the 80 test takers (20% of the sample), 7 orphans (20%), 55 public
school students (24%), and 18 (11%) private school students had their General
Mood score below 36.
308
EL HASSAN AND EL SADER
TABLE 4
BarOn EQ–i:YV–A Item Pairs Used in the Inconsistency Index
Items Pairs r Items Pairs R
17 and 43 .62 (.61) 38 and 48 .43 (.46)
7 and 31 .62 (.49) 26 and 35 .70 (.56)
55 and 59 .50 (.48) 3 and 11 .52 (.47)
20 and 51 .24 (.44) 40 and 47 .60 (.59)
30 and 22 .60 (.51) 56 and 60 .70 (.70)
Note. BarOn EQ–i:YV–A = Arabic Bar-On Emotional Quotient Inventory: Youth Version.
Do Not Copy
Positive Impression scale.
This scale aims to check whether the test taker
is attempting to create an overly positive self-impression. This score should be be
-
low 18. Among the test takers, 86 (22%) had a Positive Impression score higher
than 18.
Factor Analyses
Principal components factor analysis for the scales (Intrapersonal, Interper
-
sonal, Stress Management, and Adaptability) with varimax rotation revealed two
main factors explaining 68% of the variance. Items relating to Interpersonal and
Adaptability highly loaded on Factor 1, whereas items relating to Intrapersonal
and Stress Management highly related on Factor 2. Table 5 reports the factor struc
-
ture of the scales.
Factor analysis of responses to the 40 items using a varimax rotation yielded 11
factors explaining 57% of the variance. Table 6 reports the factor structure ob
-
tained. The resulting factor structure generally supported the factors that the
BarOn EQ–i:YV was designed to measure and reported high factor loading for
each variable it is designed to measure. None of the factor loadings were below
(0.30), most were above 0.50, and the loadings were generally low on the other fac-
tors they were not designed to measure.
When we examined the item factor structure closely, we realized that the factors
generally represented subcomponents of the four main factors. Factor 1 was com-
posed of items relating to impulse control, the second subcomponent of Stress
Management. Factor 2 was composed of items relating to assertiveness, the second
subcomponent of Intrapersonal. Factor 3 was composed of items relating to empa-
thy, the first subcomponent of Interpersonal abilities. Factors 4 and 5 were com
-
posed of items relating to problem solving and to persistence, the third and second
subcomponents of Adaptability, respectively. Factors 6 and 7 were composed of
items relating to friends and to social responsibility, the third and second
subcomponents of Interpersonal, respectively. Factor 8 was composed of reality
testing; the first component of Adaptability, whereas Factor 9 was composed of
items relating to emotional self-awareness, the first subcomponent of
Intrapersonal. The other two factors did not have high loadings on any dimension.
ADAPTING AND VALIDATING THE BARON 309
TABLE 5
Factor Structure of the Four Scales
Scale Factor 1 Factor 2 Variance (%)
Interpersonal 0.875 42.22
Adaptability 0.817 25.95
Stress Management 0.801 18.90
Intrapersonal 0.758 12.92
Do Not Copy
310
TABLE 6
Factor Analysis of the 40 BarOn EQ–i:YV–A Items
Factors
Item 1 2 3 4 5 6 7 8 9 10 11
35 .801
26 .773
54 .635
58 .628
39 –.563
3 –.522
31 .819
17 .813
43 .799
7 .792
10 .740
59 .733
55 .712
2 .517
5 .435
30 .786
34 .755
22 .680
12 .545
57 .583
25 .552
11 .537
44 .462
20 .687
51 .650
41 .465
14 .706
21 –.569
36 .441
48 .624
38 .564
45 .457
28 .728
6 .563
53 .508
46 .589
49 .444
16 –.291
15 .700
24 –.510
Note. BarOn EQ–i:YV–A = Arabic Bar-On Emotional Quotient Inventory: Youth Version.
Do Not Copy
Intercorrelation of the BarOn EQ–i:YV–A Scales
The intercorrelations of the BarOn EQ–i:YV–A scales provide additional sup
-
port for the multidimensionality of the measure (see Table 7). Consistent with the
view that the instrument measures relatively distinct aspects of emotional intelli
-
gence, low-to-moderate correlations were found among the scales. Subscale corre
-
lations ranged from .10 to .72, whereas subscale correlations with EQ were moder
-
ate, ranging between.46 and .72. Low intercorrelation confirms the existence of
separate factors. The relatively higher correlations between the Interpersonal and
Adaptability scales (r = .46) and between the Intrapersonal and Stress Manage
-
ment scales (r = .25) confirm the results of the scale factor analysis in which only
two factors were extracted.
Age and Gender Effects
To investigate age and gender differences on various subscales and on total EQ,
a series of two-way ANOVAs (age and gender) were performed with the BarOn
EQ–i:YV–A. Table 8 reports scale means by age and gender. Only students who
had an Inconsistency Index score lower or equal to 10 were included in the analysis
(N = 390). Significant gender differences were identified only on the Interpersonal
scale and at all age levels, F(1, 416) = 26.66, p < .00, η
2
= .06. With respect to age
differences, they were reported on the General Mood scale and on the Positive Im-
pression scale between ages 10 through 12 and the older groups, F(2, 416) = 9.4, p
< .00, η
2
= .04 and F(2, 416) = 14.2, p < . 00, η
2
= .03, respectively.
Correlation Between Long and Short Forms of the
BarOn EQ–i:YV–A
The correlations between the BarOn EQ–i:YV–A long form and the BarOn
EQ–i:YV–A short form scales were good, ranging from r = .78 to r = .86.
ADAPTING AND VALIDATING THE BARON 311
TABLE 7
Intercorrelation of the BarOn EQ–i:YV–A Scales
Scale 1 2 3 4 5 6
Intrapersonal 1.00 0.11 0.25 0.25 0.70 0.22
Interpersonal 0.11 1.00 0.09 0.46 0.60 0.38
Stress Management 0.25 0.09 1.00 0.17 0.55 0.21
Adaptability 0.25 0.46 0.17 1.00 0.72 0.42
Total EQ 0.70 0.60 0.55 0.72 1.00 0.46
General Mood 0.22 0.38 0.21 0.42 0.46 1.00
Note. BarOn EQ–i:YV–A = Arabic Bar-On Emotional Quotient Inventory: Youth Version; EQ =
emotional quotient.
p < .01 level.
Do Not Copy
DISCUSSION
Our purposes in this study were to adapt and validate the BarOn EQ–i:YV in the
Lebanese context to provide insight into the emotional development of intelligence
skills of children and youths.
Adaptation and “Arabization” of the BarOn EQ–i:YV–A
The BarOn EQ–i:YV was translated to Arabic in three stages until a reliable and
acceptable version was developed. This was due to the complexity of emotional in
-
telligence concept, the age of the students, and possibly the use of a Likert-type
scale. All items were subjected to extensive critical review. The goal was to detect
technical problems in the items, such as lack of clarity, as well as to determine the
312
EL HASSAN AND EL SADER
TABLE 8
Means and Standard Deviations for the BarOn EQ–i:YV–A Scales (by Age
and Gender)
Males Females
Scale M SD M SD
10–12 years
Intrapersonal 14.30 3.27 14.13 3.47
Interpersonal 38.19 4.65 38.55 4.74
Stress Management 28.79 5.10 29.50 5.40
Adaptability 31.27 5.23 30.98 4.96
Total EQ 55.47 5.94 55.61 6.99
General Mood 42.74 5.12 44.16 5.21
Positive Impression 17.77 5.78 17.59 4.82
13–15 years
Intrapersonal 13.93 3.48 14.03 3.94
Interpersonal 37.12 4.47 40.14 4.84
Stress Management 28.69 4.70 28.41 4.96
Adaptability 30.81 5.06 30.26 4.99
Total EQ 54.43 6.14 55.38 6.92
General Mood 42.61 5.35 41.47 5.83
Positive Impression 16.23 2.73 16.32 3.00
16–18 years
Intrapersonal 13.45 3.56 13.11 3.74
Interpersonal 37.66 4.97 41.57 4.23
Stress Management 29.13 4.27 27.70 5.05
Adaptability 29.61 4.79 31.70 4.20
Total EQ 53.85 5.60 55.64 5.31
General Mood 40.90 5.37 41.64 4.89
Positive Impression 14.67 2.19 16.92 2.59
Note. BarOn EQ–i:YV–A = Arabic Bar-On Emotional Quotient Inventory: Youth Version.
Do Not Copy
suitability of the items for the age and grade groups for which they were intended.
The pilot study lead to the exclusion of the 7 through 9 years age group, as it was
difficult for this particular age group to respond reliably: They could not under
-
stand instructions, the terms, and the concepts. Results of the pilot study showed
that the instrument has acceptable internal consistency and stability over time and
that the items were understood well by the sample.
Reliability
Internal reliability.
Overall, the reported Cronbach alpha coefficients were
moderate to high across all age groups except for the Stress Management scale.
These reliabilities were similar to those reported by Bar-On and Parker (2000), al
-
though slightly lower. Internal reliability of the short version was higher than that
of the long version. To explain the reason behind the lower reliabilities of the Stress
Management scale dimensions in the long version, we examined the factor analy
-
sis data. Only half of the 12 items of the Stress Management scale clustered around
Factor 1, Impulse Control, whereas the other items loaded on other Factors (10 and
11). This could explain the lower internal consistency of Stress Management in the
long version.
Test–retest reliability.
The obtained test–retest results (see Table 3) were ex-
cellent and comparable to Bar-On and Parker’s (2000) except for the General
Mood reliabilities, which were lower. In Bar-On and Parker’s results, General
Mood and Positive Impression scale correlations over time were not as high as
those of the other scales. The possible explanation could be that both scales are not
stable in nature over time. In addition, the number of students with low General
Mood score (n = 86) in this study may have affected the results.
Validity
Factor analyses.
Principal components factor analysis for scales with
varimax rotation revealed two main factors. Items relating to Interpersonal and
Adaptability highly loaded on Factor 1, whereas items related to Intrapersonal and
Stress Management highly loaded on Factor 2. This can be explained by the fact
that successful relations with others (Interpersonal) require the ability to adapt, and
successful relations with the self and self-understanding (Intrapersonal) help in
stress management. The importance of the two facets of personal intelligence—in
-
terpersonal and interpersonal—has been highlighted by Weinstein (1969),
Gardner (1983), and Salovey and Mayer (1990).
The results of the item factor analysis generally supported the factors that the
BarOn EQ–i:YV was designed to measure. The factors represented the
subcomponents of the scales’ four dimensions. Out of 13 subcomponents stipu
-
ADAPTING AND VALIDATING THE BARON 313
Do Not Copy
lated by Bar-On and Parker (2000), 9 were identified. For example, each of the 3
subcomponents of the Interpersonal scale (empathy, friends, and social responsi
-
bility) was viewed as a separate factor by Lebanese adolescents. This means that
Lebanese look at relationships with friends differently from relationships with oth
-
ers. Similarly, persistence and using different approaches and problem-solving
abilities were viewed as separate entities. Also, Lebanese differentiated between
assertiveness and self-awareness, both clustered under Intrapersonal according to
Bar-On and Parker, but were viewed as separate by the sample in this study.
Intercorrelations of the BarOn EQ–i:YV–A scales.
All subscales had rel
-
atively high correlations with EQ. Intercorrelations of the BarOn EQ–i:YV–A
scales ranged from .10 to .72, and these low-to-moderate correlations provide fur
-
ther evidence for the existence of separate factors. Moreover, the moderately higher
correlation between Intrapersonal and Stress Management on one side and Interper
-
sonal and Adaptation on the other provides further support for the scale factor analy-
sis in which only two factors were extracted. Evidence for the two factors has a rich
theoretical base from Freud, to James, to social and personal intelligences, and mov-
ing on to Gardner’s (1983) Intrapersonal and Interpersonal intelligences and to
Bar-On and Parker’s (2000) definition of emotional intelligence.
Age and gender effects.
Bar-On and Parker (2000) did not report age dif-
ferences on the EQ or on any of the scales. Similarly, in this research, we found no
significant age differences on EQ and on most of the scales except for the General
Mood and Positive Impression scales. On both scales, the younger age group
(10–12 years) had significantly higher means. Developmental theory expects
young adolescents to go through a period of turmoil and “stress and storm.” Feel
-
ings of confusion, low levels of self-confidence, and extreme mood swings are felt
to be typical of this age group (Snowman & Biehler, 2003). As for gender differ
-
ences, Bar-On and Parker reported significant gender differences in favor of fe
-
males on the Intrapersonal, Interpersonal, and total EQ scales and in favor of males
on the Adaptability scale. The higher female means were not confirmed in this
study except on the Interpersonal scale, and no significant gender differences were
noted on EQ and all other scales.
Correlation of the long and the short versions of BarOn EQ–i:YV–A.
Correlations between the long and the short versions were high and quite similar to
Bar-On and Parker’s (2000), an indication that the short version can be used in situ
-
ations in which the long version cannot be used. Moreover; the short version, in
general, had higher internal consistency than the long one. The reason could be
that participants may have found it easier to fill out, as it takes less time.
314
EL HASSAN AND EL SADER
Do Not Copy
Arab and Lebanese Emotional State
Emotional disorders and problems exist in Arab or Lebanese societies. When com
-
paring the means obtained in this research with those reported by Bar-On and
Parker (2000), a generally lower EQ for the Lebanese sample at all age levels with
a lower variability of scores was identified. Some of the subscale (Stress Manage
-
ment and Intrapersonal) means in both samples were close, but large differences
were noted in the Adaptability and the General Mood scales for which the Leba
-
nese means were lower. This is a confirmation of Yaakoub’s (1980) findings that
22% and 80% of students have adaptability problems and are dissatisfied, respec
-
tively. The Positive Impression score was higher, and this is very common for
young ages (Bar-On & Parker, 2000). The lower EQ obtained can be explained by
the fact that Arabs, in general, and Lebanese, in particular, are very emotional and
impulsive. Moreover, neither at home nor at school were children trained in these
specific skills. Similarly, the higher Positive Impression and lower General Mood
can be explained by the number of respondents (22% and 20%, respectively) who
had unacceptable scores on each of the two scales. The Lebanese sample had lower
mood and overly positive self-impression. These results support the research done
by Yaakoub and Yaakoub (1992) about the confusion of youth in Lebanese society.
In this research, we revealed a secondary finding in terms of differences in re-
sponding between orphanage and public school children and those coming from
private schools. A higher percentage of the former (6% to 24%) had unacceptable
subscale scores versus 2% to 11% of the private. Their deprived situations and the
total absence of EQ programs in their environment may provide an explanation.
SUMMARY AND CONCLUSION
Adapting the BarOn EQ–i:YV to Arabic has been a very difficult task due to the
complexity of the construct. The translation and adaptation had to be repeated
three times before an acceptable version was adopted. The pilot study resulted in
further revisions and in excluding the 7 to 9 years age group, as it was difficult for
them to respond to the questions reliably. The collected data revealed some prob
-
lems with participants from orphanages and public schools. Despite all of this, this
study on emotional intelligence on youth produced encouraging and interesting
findings. The reliability of the adapted test, both internal and over time, was satis
-
factory. When exploring factorial validity of the test by scales, two major factors
were identified. While exploring it by items, the resulting factorial structure re
-
vealed some peculiarities of the Lebanese situation. The culture where the BarOn
EQ–i:YV was normed deals with some items as one dimension, whereas in the
Lebanese context, these same items are viewed as different factors. This is ex
-
pected especially when dealing with a construct like emotions.
ADAPTING AND VALIDATING THE BARON 315
Do Not Copy
Reported age and gender differences were in general similar to those on the
original test except for Intrapersonal and Adaptability scales, where Bar-On and
Parker’s (2000) gender differences were not confirmed. We provided explanations
in relation to the Lebanese culture.
RECOMMENDATIONS FOR FUTURE RESEARCH
Based on the results of this study, we would recommend (a) using the BarOn
EQ–i:YV–A with ages 10 and above, as this research revealed the difficulty of
younger age groups with the instrument; (b) using the short version, as it proved to
have higher reliability and was easier to use; (c) reinvestigating the factorial struc
-
ture using confirmatory factor analysis; (d) developing norms based on the con
-
firmed factorial structure; (e) developing or adapting the instrument for the 7
through 9 years age group; and (f) adapting emotional intelligent programs and
conducting experimental research to investigate their validity.
REFERENCES
Anastasi, A., & Urbina, S. (1997). Psychological testing (7th ed.). Upper Saddle River, NJ: Prentice
Hall.
Bar-On, R., & Parker. J. (2000). Bar-On Emotional Inventory: Youth Version. Toronto, Ontario, Can-
ada: Multi-Health Systems.
Center for Educational Research and Development (CERD). (1999, May). Al Tarbiyah al Sohia Wa al
Bi’aa [Environmental and health education]. Beirut, Lebanon: Ministry of Health.
Cronbach, L. S., & Drenth, P. J. D. (Eds.). (1972). Mental tests and cultural adaptation. Hague, The
Netherlands: Mouton.
Deeb, M. E. (1997). Beirut: A health profile 1984–1994. Lebanon: American University of Beirut.
Finegan, J. E. (1998). Measuring emotional intelligence: Where we are today. Paper presented at the
annual meeting of the Mid South Education Research Association, New Orleans, LA. (ERIC Docu
-
ment Reproduction Service No. ED426087)
Gardner, H. (1983). Frames of mind: The theory of multiple intelligence. New York: Basic Books.
Gibbs, N. (1995, October 16). The EQ factor. Time, 146, 54–73.
Goleman, D. (1995). Emotional intelligence: Why it can matter more than IQ for character, health and
lifelong achievement. New York: Bantam Books.
Hambleton, R. (1999). Issues, designs, and technical guidelines for adapting test in multiple languages
and cultures (Laboratory of Psychometric and Evaluative Research Rep. No. 353). Washington, DC:
Georgetown University.
International Test Commission. (2000, April 21). ITC test adaptation guidelines. Retrieved October 15,
2003, from http://www.intestcom.org/test_adaptation.htm
Karam, E. G. (1990). Children war. Revue Medicale Libanaise, 2, 5–6.
Oester, T. K., & Oester, D. F. (1997). Life long learning to be productive. Paper presented at the Confer
-
ence on Education, Beijing, People’s Republic of China. (ERIC Document Reproduction Service
No. ED425409)
316 EL HASSAN AND EL SADER
Do Not Copy
Salovey, P., & Mayer, J. D. (1990). Emotional intelligence. Imagination, Cognition, and Personality, 9,
185–211.
Schilling, D. (1996). 50 activities for teaching emotional intelligence. Level one: Elementary. Torrance,
CA: Innerchoice Publishing.
Snowman, J., & Biehler, R. (2003). Psychology applied to teaching (10th ed.). Boston: Houghton
Mifflin.
Traumatic Stress Points. (1997, Fall). News for the Society for Traumatic Stress Studies, 11, 7–8.
Weinsten, E. A. (1969). The development of interpersonal competence. In D. A. Gosline (Ed.), Hand
-
book of socialization theory and research (pp. 753–775). Chicago: Rand McNally.
Yaakoub, G. (1980). Al tifl al lobnanimin 6–7 sanawat [Lebanese child 6–7 years]. Beirut, Lebanon:
Educational Center for Development and Research.
Yaakoub, G., & Yaakoub, L. D. (1992, January). Ilm Nafss al Mourahek: Binaa al Hawia Ind al
Mourahiq wa al Shabab [The psychology of youth: development of identity in teenagers and adults].
Al Thakafa al Nafssia, 3, 45–49.
Zein, B. (2001). A survey of the extent of bully/victim problems in private schools in Beirut. Unpub
-
lished master’s thesis, American University of Beirut, Lebanon.
ADAPTING AND VALIDATING THE B
ARON 317
Do Not Copy
