Psychometric properties of the Arabic version of the Trait Emotional Intelligence Questionnaire Short Form (TEIQue-SF)

Abstract

The Trait Emotional Intelligence Questionnaire Short Form (TEIQue-SF), a brief and effective tool used for assessing emotional intelligence, originated in the Western culture and needs to be adapted to other cultures. The current study aimed to evaluate the psychometric properties of an Arabic-adapted version of the TEIQue-SF. A cross-sectional study was conducted on a sample of 487 Arabic-speaking participants, including students and laypeople. Confirmatory factor analysis (CFA) was used to assess the conceptual factorial structure. Measurement invariance analyses were performed across genders. The big five were employed as external validation measures. The CFA findings validated the proposed TEIQue-SF structure, which included one latent factor (global trait EI) and four indicator factors. Analysis of measurement invariance revealed that the TEIQue-SF Arabic version was factorially equivalent across genders. Furthermore, TEIQue-SF demonstrated good internal consistency. No significant differences are observed between women and men in terms of trait emotional intelligence. The correlations between TEIQue-SF and the big five dimensions demonstrated the theoretically predicted relationships with these variables. We concluded that the Arabic version of the TEIQue-SF is psychologically sound and appropriate for research and practice in Jordan.

Keywords:

• trait Emotional Intelligence
• TEIQue-SF
• Arabic Version of TEIQue
• psychometric properties
• measurement Invariance

1. Introduction

The Emotional Intelligence Questionnaire (TEIQue) was created to reflect the theoretical and conceptual frameworks of Petrides’ trait emotional intelligence theory (2001, 2009, and 2010) and comprehensively cover the sampling domain of this psychological construct (Petrides, 2001). In contrast to ability models, which use maximum performance tests to assess emotional intelligence (EI), TEIQue belongs to the trait model, which utilizes typical performance tests, such as questionnaires, to assess EI (Mayer et al., 2003; Petrides, 2011; Petrides et al., 2007; Szczygieł et al., 2015). Petrides (2009a) defines EI as a constellation of emotional perceptions at the lower level of the human personality hierarchy. Previous research has demonstrated the significance of EI and its relationship with many aspects of life, including professional success, mental health, and education (Hjalmarsson & Dåderman, 2020; Siegling et al., 2015). EI is positively correlated to life satisfaction (Freudenthaler et al., 2008; Petrides et al., 2007), happiness (Petrides et al., 2007), optimism (Mikolajczak et al., 2007), performance in stressful situations and emotion-focused coping (O’Connor et al., 2017), and career decisions (Farnia et al., 2018). However, EI is also negatively correlated with mental disorders (Mikolajczak et al., 2007) and depression (Rudenstine & Espinosa, 2018). The family of TEIQue scales is widely used in various fields, including commercial, medical, and educational fields (Perazzo et al., 2021; Petrides, 2009a). The full version of TEIQue includes 153 items distributed across 15 facets and four factors, with two of the 15 facets directly related to the global trait EI. While the TEIQue-SF short version includes 30 items dispersed across four factors, it is designed to yield a global score, even though scores on the four factors can be derived with low reliability (Petrides, 2009b).

Although the TEIQue-SF has undergone extensive research to support its psychometric properties in various languages and cultures, far fewer studies have been conducted on this subject compared to the full version. In Europe, versions have been made in German (Jacobs et al., 2015), Polish (Szczygieł et al., 2015), Spanish (Laborde et al., 2016), and Swedish (Hjalmarsson & Dåderman, 2020), in addition to the English version (Cooper & Petrides, 2010). It has also been translated into Portuguese (Carvalho et al., 2010), Greek (Stamatopoulou et al., 2016), and Australian English (Perera, 2015). The Chinese version has been adapted in Asia (Feher et al., 2019), as has the Pakistani version (Shahzad et al., 2014). In Latin America, research has been conducted in Mexico (Neri-Uribe & Juárez-García, 2016), Chile (Pérez-Díaz & Petrides, 2021), and Brazil (Perazzo et al., 2021). Only one study was conducted in the Arab world, with a sample of Lebanese American University students (Sanchez-Ruiz et al., 2021) for the full version of the TEIQue, which was administered in English and not translated into Arabic because the language of study used in the university was English. To the researcher’s knowledge, no TEIQue-SF adaptation study has been conducted in the Arab world, in general, or Jordan, in particular.

1.1. Psychometric properties of TEIQue-SF

In their UK study, Cooper and Petrides (2010) reported that the original questionnaire was reliable for the global trait EI (α: women = .87; men = .88). Jacobs et al. (2015) validated the German TEIQue-SF in a sample of occupational therapists in Germany, reporting a similar reliability value for global trait EI (α = .88). Most factor-level alpha coefficients were less than 0.70 (α = .58- .67), with the exception of well-being (α = .85). Thereafter, an adequate model was obtained using high-order confirmatory factor analysis (CFA) with TEIQue-SF item parcels. Szczygieł et al. (2015) validated the Polish version of the TEIQue-SF and reported high reliability of the global trait EI (α = .89). Similarly, Laborde et al. (2016) found that the global trait EI (α = .88) was highly reliable in their Spanish sample, even though the reliability values for self-control, emotionality, and social were all around 0.7, with the exception of the well-being factor (α = .83). Based on four factors and a global EI, the model assessed Spanish TEIQue-SF, and the researchers obtained excellent model fit indices using CFA analysis. Consequently, the TEIQue-SF was validated in two independent studies, the first in Australia (Perera, 2015) and the second in Portugal (Carvalho et al., 2010). Both studies reported the same high-reliability values for the global trait EI mentioned in the original English study and other European studies. Stamatopoulou et al. (2016) validated the TEIQue-SF in the general population of Greece at various educational levels. Researchers reported similar reliability scores at the global trait and factor levels of the original English study, with the global trait EI (α = .89) and reliability at the factor level ranging between (α = .60) for self-control and (α = .78) for the well-being factor. In China, Feher et al. (2019) reported similar values for the global trait EI (α = .88), and Cronbach’s alpha at the factor level ranged from 0.47 (sociability) to .82 (well-being). On the other hand, the model fit was tested using CFA with four facets loaded on the global trait EI factor. Although most model fit indicators in the Chinese TEIQue-SF were deemed acceptable, researchers have found a lack of measurement invariance between the two culturally compared samples, Chinese and Canadian. The Urdu TEIQue-SF validation study (Shahzad et al., 2014), which was conducted with normal participants from different regions of Karachi, Pakistan, revealed the same reliability values for global trait EI (α = .89) and retest correlation 0.82. In Latin America, Neri-Uribe and Juárez-García (2016) reported adequate overall score reliability (Ʊ = .61, α = .83) in their Mexican sample (155 undergraduate students), but the factor reliability was rather low (α = .32–.82). A first-order CFA with items directly designed on their principal factors confirmed the instrument’s factor structure. Consequently, the model fit was less than ideal. The Chilean TEIQue-SF, on the other hand, provided significantly better model-fit indicators in the general population and clinical patients. The instrument’s endogenous structure was confirmed through the Exploratory Structured Equation Model (ESEM), where the model-fit statistics were marginally better in the general population than in the case of clinical patients (Pérez-Díaz & Petrides, 2021). In a more recent study, Perazzo et al. (2021) adapted the Brazilian TEIQue-SF. A model was tested using ESEM. Thereafter, measurement invariances were tested against the UK and Chilean measurement datasets. The researchers obtained the adequate goodness of fit for the bi-factor ESEM model. Despite the low Cronbach alpha values at the factor level (α = .60–.85), the internal consistency and test-retest reliability for the global trait EI (α = .88, .89) are high. In summary, substantial evidence supports the finding that the global trait of EI, as measured by TEIQue-SF, is highly reliable and that factor-level reliability scores range from satisfactory to high (Pérez-Díaz & Petrides, 2021). The researchers used exploratory and confirmatory factor analysis and ESEM to validate the tool’s factorial structure. The majority of these findings confirmed the theoretical robustness and validity of the scale construction. Given the importance of the TEIQue-SF scale and the power of its theoretical basis, as well as its high effectiveness in estimating the trait of EI, regardless of its implementation in research or practice, its validation in Jordan is an urgent and worthwhile task. This is because the tool is psychologically powerful, easy to administer, and supported by a large number of cross-cultural studies (Perazzo et al., 2021; Pérez-Díaz & Petrides, 2021).

1.2. Gender differences in trait EI

Previous research studies have revealed significant gender differences in global trait EI, favoring men over women (Perazzo et al., 2021; Petrides, 2009b; Stamatopoulou et al., 2016). Controversial research reports that women are higher values than men in terms of global trait EI (Tsaousis & Kazi, 2013; Cooper & Petrides, 2010). Other research studies have found no gender differences in the global trait of EI (Pérez-Díaz & Petrides, 2021; Petrides & Furnham, 2000, 2006). With respect to the level of EI factors, significant gender differences in favor of women were found on emotionality (Mikolajczak et al., 2007; Perera, 2015; Petrides, 2009b; Petrides & Furnham, 2000), whereas gender differences in favor of men were found on sociability and self-control with effect sizes varying from small to large (Mikolajczak et al., 2007; Perazzo et al., 2021; Perera, 2015; Petrides, 2009b). However, Sanchez-Ruiz et al. (2021) reported in their Lebanese sample that women scored higher than men on well-being, whereas men scored higher than women on self-control. Other studies reported no gender differences in terms of the well-being factor (Mikolajczak et al., 2007; Perera, 2015; Petrides, 2009b).

1.3. Present study

The current study’s main goals are to translate, adapt, and validate TEIQue-SF for Arabic speakers in Jordan. Practically, we aimed to identify the psychometric properties, internal consistency, and factor structure of the translated Arabic version of TEIQue-SF. We were also interested in a) assessed the measurement invariance between males and females; b) confirming TEIQue-SF’s relationships with the big five dimensions; c) collecting information on possible gender differences regarding global trait EI and the four-factor levels. This is the first study to validate the TEIQue-SF as a measure of trait EI in the Arab world and, specifically, in Jordan.

2. Method

2.1. Participants and procedures

The data were collected from Al-Balqa Applied University students and laypeople. This study included 487 participants, 30 of whom participated in the pilot study. In total, 457 participants (313 women, 68.5 percent; 144 men, 31.5 percent) (M age = 32.6, SD = 9.8, range = 18–59 years) were included in the main study. Eligible participants had to be aged 18 years or older. A total of 264 students were recruited on campus, and 193 laypeople were recruited using a snowball sample. Details of the sample characteristics are listed in Table 1. The Arabic TEIQue-SF was administered alongside the Big Five Inventory (Arabic version of the NEO-FFI, Costa and McCrae’s Inventory, Al-Ansari, 1997). Students completed the survey on paper and pencil, whereas the laypeople responded online. Regarding the students’ sample, a non-probability sample was chosen from various al-Balqa University colleges. Participants completed the questionnaires in their classrooms while being supervised by the researcher. The administration took about 20 minutes on average. In terms of laypeople, the study was promoted online via social media (Facebook and WhatsApp), where a link to the Google Forms questionnaire was sent to a group of people, and each of them was asked to distribute the questionnaire to his acquaintances and friends across Jordan. Participants read and signed an informed consent form before data collection and were reassured of their right to freely participate in or withdraw from the study at any time. The participants did not receive any incentives or rewards for their participation. The data were collected anonymously between October 2021 and April 2022. The study was conducted following the ethical standards of the Jordanian Scientific Research Ethics Unit and the Helsinki Declaration guidelines.

Table 1. Participants’ characteristics

Variables		students	laypeople	Total
gender	Male	75	69	144 (31.5%)
	Female	189	124	313 (68.5%)
	total	264	193	457 (100%)
Age(years) M. (SD)		26.3(6.2)	41.3 (6.6)	32.6(9.8)
Educational level	Elementary	-	2	2(.4%)
	Middle School	-	3	3(.7%)
	Secondary	-	22	22(4.8%)
	Technical Diploma	-	41	41(9%)
	Bachelor’s	173	115	288(63%)
	Master	91	10	101(22.1%)
professions	student	264	-	264(57.8)
	Public sectors	-	149	149(32.6%)
	private sectors	-	8	8(1.8%)
	unemployed	-	14	14(3.1%)
	retired	-	15	15(3.3%)
	homemakers	-	7	7(1.5%)

M= mean, SD = standard deviation

2.2. Procedure of translation

TEIQue-SF was translated and cross-culturally adapted following the standard-specific guidelines (Beaton et al., 2000; Hambleton, 2001; Wild et al., 2005). Two bilingual English and Arabic translators independently translated the TEIQue-SF into Arabic. After discussion, the translations were standardized into a single version, and the preliminary Arabic version of the TEIQue-SF was formed. A bilingual, psychology-proficient professor performed the back translation, and the two versions were compared in terms of the items’ conceptual equivalence. A panel of specialized arbitrators later produced the second Arabic version by comparing the Arabic translation with the back translation and the original English version. Subsequently, the questionnaire was administered to a pilot sample of 30 students enrolled in the author’s counseling and career guidance course who did not participate in the main sample. The students were asked to specify the difficulties they encountered while responding, and they made some minor observations that had no bearing either on the essence of the questionnaire or its items (presenting or delaying a word and substituting eloquent words with simple common words), which have been taken into account. Following this testing, we made minor adjustments to enhance the instrument’s legibility. Finally, a language expert used the Arabic version to create a TEIQue-SF version in Standard Arabic that can be shared across all Arab countries.

3. Measures

3.1. Trait Emotional Intelligence Questionnaire (TEIQue-SF)

The Trait Emotional Intelligence Scale (Petrides, 2009b) includes 30 items that assess the following four factors: well-being, self-control, sociability (each with six items), and emotionality (eight items), as well as four additional items 3, 14, 18, and 29, which do not belong to any factor and directly contribute to the evaluation of the overall degree of the trait emotional intelligence, as indicated by the scoring key. The questionnaire is graded on a seven-point Likert scale, with responses ranging from 1 (strongly disagree) to 7 (strongly agree). This TEIQue-SF scale has been used in numerous studies (Cooper & Petrides, 2010; Perera, 2015) owing to its good psychometric properties.

3.2. Big five

The Big Five Inventory (Arabic version prepared by Al-Ansari, 1997) was used for measuring the five major personality factors, which comprise 60 items and five domains (neuroticism, extraversion, openness, agreeableness, and conscientiousness), 12 items per domain, which were rated on a five-point Likert-type scale from 1 = strongly disagree to 5 = strongly agree. Many studies (for example, Abu Hashem, 2007; Al-Ansari, 1997) were conducted on the big five inventory in the Arabic environment, and they were found to have good psychometric properties. The internal consistency coefficients for the five sub-factors ranged between 0.53 and 0.82 in the current study.

3.3. Data analysis strategy

We conducted statistical analyses in SPSS 23 and Amos 28. Confirmatory factor analysis (CFI) with maximum likelihood (ML) estimation was used to examine the Arabic TEIQue-SF structure. A comparative fit index (CFI), Incremental Fit Index (IFI), Tucker Lewis Index (TLI), root mean square error of approximation (RMSEA), and standard root mean residual (SRMR) were used to evaluate the model’s fit. To judge the fit of a good model, the following thresholds were used: CFI, IFI, TLI > .95, RMSEA < .06, and SRMR < .08 for acceptable model fit; CFI, IFI, TLI > .90, RMSEA < .08, and SRMR < .10 for adequate fit, we also presented the Chi-square value (χ²; Hu & Bentler, 1999). To test the measurement invariance of the Arabic version of the TEIQue-SF across gender, a set of CFA models were tested in a hierarchical order using maximum likelihood estimation using Amos-28, the first step was to determine whether the factor structure was similar in males and females using configural invariance. Following that, metric invariance is used to determine whether the two groups’ factor loadings are equivalent. Finally, the scalar invariance is used to determine whether the intercepts are equivalent; latent mean differences among groups should be compared at this level of invariance (Feher et al., 2019; Milfont & Fischer, 2010; Putnick & Bornstein, 2016). Difference tests χ2, CFI, RMSEA, and SRMR were used to compare the configural, metric, and scalar measurement invariance models. The following criteria were applied to these tests: ∆CFI values less than or equal to 0.01 units, and ∆RMSEA values less than 0.015 units. ∆SRMR values less than 0.03 units indicate metric invariance, while changes within 0.01 indicate scalar invariance. (Chen, 2007; Cheung & Rensvold, 2002). To compare males and females on the global trait EI and the four structural factors, we used independent t-test and Cohen’s d effect sizes in descriptive and inferential analyses and an estimate of Cronbach’s alpha for global and the four trait factors. Pearson’s correlations were also computed to test the convergent and divergent validity of trait EI and the big five inventory.

4. Results

The mean, standard deviation, skewness, and kurtosis of the global TEIQue-SF score, as well as the four factors separately for men and women and the whole sample, are shown in Tables 2 and 4. According to Curran et al. (1996), absolute skewness and kurtosis values ranging from 0 to 2 and from 0 to 7, respectively, indicate sufficient univariate normality. Tables 2 and 4 show that a common negative skewness is more stressful for the well-being factor, where participants score higher (Pérez-Díaz & Petrides, 2021). The correlations between the factor and global trait scores were positive and significant. KMO and Bartlett’s tests were used to determine whether the data was suitable for factor analysis. The Kaiser-Meyer-Olkin value was .92, which was higher than the recommended value of .60 (Kaiser, 1974). The factorability of the correlation matrix was supported by Bartlett’s test of sphericity (Bartlett, 1954), χ² (435, N = 457) = 7624.1, p = .001.

Table 2. Descriptive, α-coefficients and gender differences in the Arabic TEIQue-SF (N = 457)

	Well-being		Self-control		Emotionality		Sociability		Global Trait
variables	men	women	men	women	men	women	men	women	men	women
mean	5.02	4.92	4.49	4.36	4.44	4.41	4.47	4.35	4.61	4.52
SD	1.32	1.17	.98	.87	1.06	.98	1.02	1.00	.97	.84
range	6	6	4.5	5	4.50	5.25	4.67	6	4.60	5.23
S	−1.16	−.57	−.19	.41	−.12	.29	−.15	−.23	−.52	−.04
K	1.2	.40	−.52	.54	−.79	−.24	−.09	.96	.13	.88
α	.81	.73	.53	.38	.68	.61	.54	.58	.91	.87
t-test (p-value)	.41		.18		.74		.23		.30
Effect size	.08		.14		.03		.12		.11

S = Skewness, K = Kurtosis, SD = Standard Deviation

4.1. Factor structure

The internal structure of the Arabic TEIQue-SF comprises one latent factor (global trait EI) and four indicator factors (well-being, self-control, emotionality, and sociability). The model yielded the following good fit indices: χ² (2) = 5.58, p = .061, SRMR = .0131, CFI = .996, IFI = .996, TLI = .988, RMSEA = .063 [90 percent CI: .127 to .282]. All factor loadings were high and ranged between .77 and .84, indicating that the Arabic version of TEIQue-SF achieved a good fit, providing support for the construct validity of the instrument (Figure 1).

Figure 1. Confirmatory factor analysis of the Arabic TEIQue-SF (N = 457).

4.2. Measurement invariance by gender

To compare men and women in the trait of emotional intelligence, the measurement invariance of the factorial structure of the TEIQue-SF must be tested (Reise et al., 2000). To achieve this goal, nested CFA models were compared to determine whether there was gender equivalence. As shown in Table 3, the baseline model’s fit indices supported the hypothesis of configural measurement invariance [χ2 (4) = 5.49; p < 0.24, RMSEA = 0.019 (90% CI = 0.081–0.685), CFI = 0.998]. According to Byrne (2016), the ratio between the chi-square value and the degrees of freedom of the model (χ²/ df) must be less than three in order for the model to be accepted. As shown in Table 3, the values of χ²/ df of configural, metric, and scalar invariance in the current study were less than 3. Thus, the Arabic version of TEIQue-SF achieved configural invariance of the one latent factor and four indicators’ scale structure across gender. Although changes in the RMSEA were outside the threshold values for metric invariance, the CFI and SRMR changes were acceptable (∆CFI = −0.008, ∆SRMR = 0.015). For scalar invariance, changes in the CFI, RMSEA, and SRMR were acceptable (∆CFI = −0.009, ∆RMSEA = 0.007, ∆SRMR = 0.004).

Table 3. Measurement invariance across gender

Models	χ^2 (df)	CFI	∆ χ^2	χ^2/ df	RMSEA (90% C.I.)	SRMR	∆CFI	∆RMSEA	∆SRMR
Configural	5.49(4)	.998	-	1.37	.029 (.081-.685)	.014	-	-	-
Metric	16.43(8)	.990	10.94	2.05	.048 (.081-.487)	.030	−.008	.019	.015
Scalar	28.35(12)	.981	11.92	2.36	.055 (.081-.346)	.036	−.009	.007	.004

Note. χ² = Chi Square, df = degrees of freedom, CFI = Comparative Fit Index, ∆ χ² = χ2 difference, RMSEA = Root Mean Square Error of Approximation, SRMR = Standardized Root Mean Square Residual, ∆CFI = CFI difference, ∆RMSEA = RMSEA difference, ∆SRMR = SRMR difference.

4.3. Reliability and gender differences

Tables 2 and 4 show that Cronbach’s internal consistency coefficient for the overall sample (α = .89) was excellent for the global score, good for well-being (α = .77), and poor for the three factors, namely self-control, sociability, and emotionality (.44, .57, .63), respectively. While men’s alpha values exceeded .90 on the global trait score, women’s alpha values were .87. With the exception of the factor of self-control for women being too small (α = .38), the alpha at the factor level ranged between .53 and .81, which is considered to range from medium to high. No gender differences were observed in terms of trait EI between women and men, whether on a global trait or four-factor level of TEIQue-SF, as found by the independent sample t-test: t (global trait) = 1.03, p = .30. Table 2 shows the t values for the four-factor level.

Table 4. Descriptive, α-coefficients and correlation for TEIQue-SF (N = 457)

Variable	Mean	SD	α	S	K	1	2	3	4	5
1. Well-being	4.95	1.22	0.73	−.79	.69	1
2. Self-control	4.40	0.90	0.44	.19	.05	.61**	1
3. Emotionality	4. 42	1.00	0.56	.15	−.46	.66**	.64**	1
4. Sociability	4. 38	1.00	0.63	−.20	.60	.57**	.61**	.65**	1
5. Global Trait	4.5 5	0.88	0.89	−.22	.53	.85**	.80**	.89**	.82**	1

**. Correlation is significant at the 0.01 level (2-tailed). S = Skewness, K = Kurtosis, SD = Standard Deviation.

4.4. Relationships between TEIQue-SF and the big five personality traits

The four Arabic TEIQue-SF factors were shown to have significant correlations with most of the big five personality traits (p < .01). The only non-significant results pertained to self-control and agreeableness, as well as sociability and conscientiousness. The global trait score EI revealed a median correlation with all the big five dimensions. As expected, neuroticism had the highest negative coefficient (−.57, p < .01), whereas the other four dimensions had positive correlations ranging from .44 to .51 (p < .01; Perazzo et al., 2021; Petrides, 2010). These results are populated in Table 5.

Table 5. Correlations between the TEIQue-SF (factors and global trait scores) and the Big five (N = 457)

Variables	Trait EI	Well-being	Self-control	Emotionality	Sociability
Neuroticism	−.57**	−.52**	−.33**	−.46**	−.54**
Extraversion	.49**	.43**	.27*	.51**	.32**
Openness	.51**	.41**	.37**	.49**	.42**
Agreeableness	.44**	.42**	.21	.51**	.27*
Conscientiousness	.49**	.46**	.31**	.56**	.21

** p ≤ 0.01; * p ≤ 0.05.

5. Discussion

The major strength of the current study is that it is the first study to test the robustness of the TEIQue-SF (Petrides, 2009b) factor structure in a sample of Arabic speakers in Jordan. We need to validate the construct of emotional intelligence in non-Western samples to confirm its applicability to Eastern cultures, particularly Arab culture. The findings revealed that the TEIQue-SF has robust psychometric properties, which was consistent with previous studies (for example, Feher et al., 2019; Laborde et al., 2016; Stamatopoulou et al., 2016). The CFA analyses also showed that the Arabic translation of TEIQue-SF was reasonably appropriate, implying that the Arabic version is factorially equivalent to the original version. We investigated the measurement invariance across groups to confirm the scale’s validity when used to compare males and females. The lack of invariance indicates that the meaning of the degrees observed on the scale varies across groups, making direct comparison impossible (Vandenberg & Lance, 2016). The measurement invariance analysis assumes that the meanings of the emotional intelligence trait are the same for men and women on the Arabic TEIQue-SF version. The current study findings showed that the TEIQue-SF had strong onfigural, metric, and scalar invariance across gender. The conclusions about measurement invariance are based on a thorough examination of all indicators, not just one; the current study relied on several indicators, with the vast majority of change indices falling within acceptable ranges (Putnick & Bornstein, 2016). It established the foundation for claiming gender equality in the trait of emotional intelligence, and this result means that the Arabic version of TEIQue-SF can be interpreted as estimates of the same trait and can be compared directly across gender. According to the researcher’s knowledge, the current study is one of the first to confirm the measurement invariance of TEIQue-SF between men and women. The Arabic internal consistency analysis of TEIQue-SF revealed high alpha values on the global trait EI; however, the values were lower when evaluated at the TEIQue-SF factor level. While the alpha values of the factors in the short model are expected to be slightly lower (Petrides, 2009a), the sociability, emotionality, and self-control factors in the current study had lower alpha values than expected (< .70). Overall, the results presented here support the reliability of the TEIQue-SF. In published papers, self-control, emotionality, and sociability reliability seem lower than in well-being, a finding that aligns with our findings. In their German version of the TEIQue-SF, Jacobs et al. (2015) reported a low value for sociability (α = .58). Indeed, Cronbach’s alpha values for men and women at the factor level are not reported in all studies. Perazzo et al. (2021) reported that men have low emotionality and sociability reliability values (α = .55). Hjalmarsson and Dåderman (2020) and Feher et al. (2019) also found a low value for sociability in the Swedish and Chinese versions (α = .55, .47), respectively, which reflects the tool’s design to measure the overall trait of emotional intelligence rather than at the factor level. In contrast to previous studies that found a significant difference in global trait EI between men and women (Mikolajczak et al., 2007; Perazzo et al., 2021; Petrides, 2009a; Stamatopoulou et al., 2016), we found no significant differences in trait EI means between women and men at the global or factor levels, a finding that is partially consistent with previous research (Pérez-Díaz & Petrides, 2021; Petrides & Furnham, 2000, 2006).

Pattern correlation with the big five provided evidence for the validity of the TEIQue-SF (EI is defined as a set of self-perceptions located at lower levels of a person’s hierarchy; Petrides et al., 2007). Consequently, the EI trait is not distinct from personality constructs but is a component of them. Therefore, we found low to moderate correlations between the Arabic version of TEIQue-SF and the big five, confirming their interrelationships and supporting the convergent and divergent validity of TEIQue-SF. Such effects were expected because the EI trait conceptually overlaps with personality dimensions. Previous studies have yielded nearly identical results (Perazzo et al., 2021; Petrides, 2010).

6. Limitations and future studies

The current study reports on the psychometric properties of TEIQue SF. TEIQue-SF appears to be a very useful EI trait research tool due to its brevity and good psychometric properties, as demonstrated in the context of many countries. According to our findings, the TEIQue-SF is a valid and reliable measure of EI traits. The current study, however, has limitations in that it only used one personality measure as a criterion. Future research will need to expand on these metrics. As the instrument’s incremental validity is considered, the scale presented in this article should be regarded as a preliminary version of the TEIQue-SF. Furthermore, the current study used mixed-strategy data collection, which may exhibit measurement bias (AERA et al., 2014). In subsequent studies, it may be preferable to use the bifactor exploratory structural equation modeling (ESEM) approach to verify the internal structure of TEIQue-SF. It may also be useful to investigate cross-cultural measurement invariance to support the tool’s validity.

7. Conclusion

This first TEIQue-SF psychological study in Arab samples adds a theoretically sound tool to the relevant psychological literature and enriches the debate in this field. In summary, Arabic TEIQue-SF has been proven reliable and valid in the Jordanian population. In terms of factor validity, it was found that a general factor model with a four-factor structure verifies the goodness of fit. The measurement invariance across gender was achieved in the Arabic version, so the scale can be used to study the differences between men and women in the trait of emotional intelligence, but more research is needed to confirm the possibility of replicating the current study’s findings. Finally, the assessment feature, which includes a validated and reliable brief scale in Arabic, allows practitioners to conduct accurate psychological assessments in various clinical, educational, and practical settings (Pérez-Díaz & Petrides, 2021).

Ethical statement

The procedures are done in line with the ethical standards of the committee in charge of human experimentation and the Helsinki Declaration. https://doi.org/10.6084/m9.figshare.21378474

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/23311908.2023.2171184.

Additional information

Funding

The author received no direct funding for this research.