You are on page 1of 32

This article was downloaded by: [University of California, San Diego

On: 09 June 2015, At: 08:27
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,

Cognition and Emotion
Publication details, including instructions for authors
and subscription information:

On the criterion and
incremental validity of trait
emotional intelligence
a b
K. V. Petrides , Juan Carlos Pérez-González &
Adrian Furnham
Institute of Education, University of London ,
London, UK
Universidad Nacional de Educación a Distancia
(UNED) , Madrid, Spain
University College London , London, UK
Published online: 03 Feb 2007.

To cite this article: K. V. Petrides , Juan Carlos Pérez-González & Adrian Furnham
(2007) On the criterion and incremental validity of trait emotional intelligence,
Cognition and Emotion, 21:1, 26-55, DOI: 10.1080/02699930601038912

To link to this article:


Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or
suitability for any purpose of the Content. Any opinions and views expressed
in this publication are the opinions and views of the authors, and are not the
views of or endorsed by Taylor & Francis. The accuracy of the Content should
not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions,
claims, proceedings, demands, costs, expenses, damages, and other liabilities
whatsoever or howsoever caused arising directly or indirectly in connection
with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan, sub-
licensing, systematic supply, or distribution in any form to anyone is expressly
forbidden. Terms & Conditions of access and use can be found at http://
Downloaded by [University of California, San Diego] at 08:27 09 June 2015

2007, 21 (1), 26 55

On the criterion and incremental validity of trait
emotional intelligence
K. V. Petrides
Institute of Education, University of London, London, UK
Downloaded by [University of California, San Diego] at 08:27 09 June 2015

Juan Carlos Pérez-González
Universidad Nacional de Educación a Distancia (UNED), Madrid, Spain

Adrian Furnham
University College London, London, UK

This paper presents a comprehensive investigation of the criterion and incremental
validity of trait emotional intelligence (trait EI or trait emotional self-efficacy),
which is defined as a constellation of emotion-related self-perceptions and
dispositions located at the lower levels of personality hierarchies (Petrides &
Furnham, 2001). In Studies 1 and 2 (N /166 and 354, respectively) trait EI is
shown to be related to measures of rumination, life satisfaction, depression,
dysfunctional attitudes, and coping. Most relationships remained statistically
significant even after controlling for Big Five variance. In Study 3 (N /212) trait
EI is shown to be related to depression and nine distinct personality disorders. Most
relationships remained significant, even after controlling for positive and negative
affectivity (mood). It is concluded that trait EI has a role to play in personality,
clinical, and social psychology, often with effects that are incremental over the basic
dimensions of personality and mood.

The construct of trait emotional intelligence (trait EI or trait emotional self-
efficacy) refers to a constellation of emotion-related self-perceptions and
dispositions located at the lower levels of personality hierarchies (Petrides,

Correspondence should be addressed to: K. V. Petrides, Institute of Education, University of
London, 25 Woburn Square, London WC1H 0AA, UK. E-mail:
KVP was supported by a Postdoctoral Fellowship from the Economic and Social Research
Council (ESRC) and research grant SGS/01075/G from the Nuffield Foundation. JCP-G was
supported by a Predoctoral Fellowship from the Universidad Nacional de Educación a
Distancia (UNED).
We would like to thank Universidad de Huelva, Escuela Universitaria Cardenal Cisneros de
Alcalá de Henares and Universidad Complutense de Madrid for facilitating the data collection
for Study 3.
# 2007 Psychology Press, an imprint of the Taylor & Francis Group, an informa business DOI: 10.1080/02699930601038912

Furnham. the conception of the construct as a constellation of emotion-related self-perceptions and dispositions suggests that it must have statistically significant effects on many different variables in many different contexts. see also Austin. Schulte. social. competencies. when properly conducted. Petrides. whereas Study 3 examined incremental validity vis-à-vis the basic dimensions of mood. On the other hand. trait EI and ability EI (Petrides & Furnham. MacCann. Its roots lie in the distinction between two EI constructs. and skills it would be rather problematic if these questionnaires did not consistently predict substantial amounts of variance over and above the basic personality dimensions. With this in mind. San Diego] at 08:27 09 June 2015 The three studies in this paper report results from the trait EI research programme. 2001. The conceptualisation of EI as a personality trait leads to a construct that lies wholly outside the taxonomy of human cognitive ability (Carroll. 2003. Fox. Studies 1 and 2 were specifically designed to address this question. & Carroll. The first aim of the paper was to establish the criterion validity of trait EI.g. All were chosen for their theoretical relevance to trait EI (hypotheses are advanced in the study introductions) and some have also been incorporated in the cross-cultural arm of our programme. The criteria in this paper span the domains of personality. O’Connor & Little. This was essential both for the systematic develop- ment of a nomological network and for the empirical testing of the nature of trait EI. & Mavroveli. 2005. Ree. 2006. requires a painstaking validation process along the lines described by Cronbach and Meehl (1955). This paper seeks to make three specific contributions in relation to this process. from the perspective of trait EI theory. 2000. Warwick & Nettelbeck. & Zeidner. 1993). The question of incremental validity is not equally relevant to the various EI models that are operationalised through self-report. hitherto supposedly undiscov- ered. The operationa- lisation of a new construct. 2004. Some researchers (e. viz. this question is of little consequence. TRAIT EMOTIONAL INTELLIGENCE 27 2001. 2005. More specifically. self-report measures of EI can predict over and above the basic personality dimensions. abilities. Therrien. Day. 2003. and clinical psychology. if anything. Theoretical motivation Downloaded by [University of California. which seeks to develop a comprehensive operationalisation of emotion-related aspects of personality across the life span. Roberts. & Carretta.. Matthews. in press-a). 2004) feel it is essential to demonstrate what. The second aim of the paper was to explore the incremental validity of trait EI. 2004. we selected 23 dependent variables to cover a broad spectrum of constructs. . 2004).. Tett. & Wang. For models based on questionnaires and theorising about novel.

Demonstrations of discriminant and incremental validity are considerably more difficult in a Big Five than a Giant Three context. because the former is a broader taxonomy (Draycott & Kline. but focus on a comparison between participants with high versus low residualised trait EI scores. & Eysenck. we maintain that trait EI theory can provide a scientific context for the meaningful interpretation of data from these measures (Petrides et al. offer of an accurate label for EI measured through self-report. Costa & McCrae.. 1999). Schulze. Pita. & Roberts. Shani-Zinovich. San Diego] at 08:27 09 June 2015 personality dimensions (Psychoticism. Studies 1 and 2 examine the incremental validity of trait EI vis-à-vis the Big Five (see Petrides. 2001. Frederickson.. which posits five basic personality dimensions (Neuroticism. Openness to Experience. John & Srivastava. Petrides & Furnham. & Furnham. PÉREZ-GONZÁLEZ. Petrides. Matthews. Big Five). and Conscientiousness. 2001).g. Therefore. 2005). & Matthews. 1995. in press-a). 2005. which posits three basic Downloaded by [University of California. Saklofske. it is expected to show strong correlations with the higher-order dimensions that define its factor space (Petrides. Moreover. 28 PETRIDES. Petrides. Study 1 utilises a modified version of the EQ-i (Bar-On. Austin. trait EI theory predicts that the construct will exhibit incremental predictive validity over the basic personality dimensions only in relation to criteria that are sufficiently affect laden. Petrides. Petrides & Furnham. Giant Three) and the five-factor (e. FURNHAM Because trait EI is explicitly conceptualised as a lower-order personality construct. competencies. whose author claims that it measures capabilities. 1998. The two established trait hierarchies are the Eysenckian (Barrett. 1990). Zeidner. 1997). & Furnham. 1992. The third aim of the paper was to highlight the generality of trait EI theory. Eysenck. & Kokkinaki. Eysenck. which extends far beyond a simple. Study 3 investigates the incremental validity of trait EI in relation to mood (dispositional affect). Although their authors invariably insist they assess abilities (Zeidner. 2004. which comprises only 60 items. while Studies 2 and 3 utilise the Trait Emotional Intelligence Questionnaire . since mood variance has also been seen as a baseline that the construct must exceed in criterion prediction. 2005). and Minski (2003) present an incremental validity study with the NEO FFI. see also De Raad. and Neuroticism. Pérez. and skills. Studies 1 and 2 are among the few in the literature to examine the incremental validity of trait EI over the NEO PI-R operationalisation of the Big Five. Agreeableness. which comprises 240 items assessing 30 different personality facets. Extra- version. 2005). Self-report measures of EI have proliferated to an extent that has led to requests for a moratorium (Roberts.. while Petrides and Furnham (2003) partial out NEO PI-R variance. for analyses with the Giant Three). Extraversion. in press-b.g. we view these questionnaires as measures of trait EI (e. 2000). albeit necessary.

over the Big Five personality dimensions. & Stough. it was hypothesised that trait EI would be positively correlated to life satisfaction (H2a. Compared to their low trait EI counterparts. If the findings are incompatible. San Diego] at 08:27 09 June 2015 questionnaires. Coping styles can be adaptive or maladaptive. then the theory would have to be amended. Life satisfaction judgements involve a cognitive evaluation of one’s circumstances against a set of subjective criteria (Diener. Like various EI models (e. that all self-report measures of EI (and cognate constructs) are measures of trait EI*irrespective of whether they were meant to operationalise abilities or competencies *then the two instruments should yield compatible findings. 1991). 2001. & Griffin. see also Palmer. particularly in individualist Western societies (Suh. 1985). trait EI would be reliably associated with lower scores on rumination (H1b). 1997. it was hypothesised that after all of the variance accounted for by the five dimensions of personality has been partialled out. based on data from two different countries. it was hypothesised that they would be less likely to ruminate (H1a). 1995) trait EI theory views the construct as central to the development and implementa- tion of successful coping mechanisms. viz. Because emotional states have a direct impact on this evaluation. & Larson. Larsen.. Petrides & Furnham. in relation to the following theoretically relevant criteria: Rumination is defined as ‘‘passively and repetitively focusing on one’s symptoms of distress and the circumstances surrounding these symptoms’’ (Nolen-Hoeksema.g. Oishi. 855). To summarise. . p. Saklofske et al. Donaldson. Goleman. 2003). McBride. high trait EI individuals should be more likely to employ adaptive coping styles (H3a) and less likely to employ maladaptive coping styles (H4a) when dealing with stress. Thus. although this depends on contextual factors too (Lazarus. 2002. Because high trait EI individuals believe they are aware of their feelings and able to regulate them.. & Triandis. If the tenet of the theory is correct.. The hypotheses concerning incremental validity are a reformulation of those presented above. Emmons. 1998). Coping is the process by which people try to manage stress. which is specifically developed as the measurement vehicle for trait EI. Petrides. TRAIT EMOTIONAL INTELLIGENCE 29 (TEIQue. 2003). STUDY 1 Study 1 examined the criterion and incremental validity of trait EI. the paper extends the current literature by presenting one of the largest investigations of the criterion and incremental validity of EI Downloaded by [University of California. Diener. and with specific emphasis on the interpretation of the findings from the perspective of trait EI theory.

and Conscientiousness.83. we do not report such analyses in this paper. 2000.27 years). the internal consistencies of the five factors were . and . Mills.81. unlike short forms. 2000). 2001). Hemmati. measuring six facets for each of the five basic personality dimensions. Although we used the long form of the inventory. 30 PETRIDES. Third. Bar-On. respectively. 1992). PÉREZ-GONZÁLEZ. however. Gignac. On this sample. Taylor. . frequent use of adaptive coping styles (H3b). 2004. Method Participants One hundred sixty-six individuals from a British university participated in the study (54 males and 110 females.. two unreported).90. however. we would be happy to release the relevant data to interested researchers upon request. 1997). Neuroticism. . Dawda & Hart. viz. of limited relevance to the present paper. are based on global scores. Downloaded by [University of California. Its psychometric properties were scrutinised in Petrides and Furnham (2001) and Palmer. as in the case of trait EI. & Kroner.. Second. Manocha. . It consists of 240 items. McCarthy. NEO PI-R (Costa & McCrae. our statistical analyses. Parker. This is the most widely used inventory for assessing normal adult personality on the dimensions of the five factor model. As in previous studies. For reasons of space. Most participants were single (:/87%) and did not have an undergraduate degree (:/77%). we incorporated an additional 15-item scale (‘‘emotion mastery’’) in an effort to cover some of the salient aspects of the trait EI domain that the EQ-i does not. Openness to Experience. Extraversion.80. & Bagby.83. the fundamental problem with this instrument lies not so much in its psychometric properties as in the conceptually flawed model that underpins it. and as briefly explained in the general discussion. There are three advantages in using the long form of an instrument. The internal consistency of the global score on this sample was . the use of the long form of the NEO PI-R allows data analyses at the facet level. long forms have more desirable internal consistency properties.81. & Anderson. and infrequent use of maladaptive coping styles (H4b). First.g. FURNHAM higher levels of satisfaction with life (H2b). and Stough (2003). The mean age for the sample was 22.17 years (SD/4. who reported factor structure problems that are. Similar to most other EI questionnaires. The EQ-i comprises 133 self-report items and has been used in many studies in the literature (e. San Diego] at 08:27 09 June 2015 Measures: Predictors BarOn Emotional Quotient inventory (EQ-i. Agreeableness. . it guarantees adequate coverage of the intended sampling domain (Smith.

and adding the Big Five personality dimensions at step 2. On this sample.5j standard deviations away from the mean were removed. ‘‘Feel that time will sort things out’’). .. Satisfaction with Life Scale (Diener et al. The 14-item rehearsal scale from the ECQ was used to measure rumination (e. Special attention was paid to potential outliers and all cases with residuals greater than j3. respectively. Participants responded on a 4-point Likert scale.g.. This cut-off value is conservative. 2001). the internal consistency was . The questionnaires took approximately 90 minutes to complete. ‘‘In most Downloaded by [University of California.. Results Due to the large amount of data. to investigate criterion validity.g.. given the large size of the sample (Stevens. ‘‘Feel worthless and unimpor- tant’’) and ‘‘avoidance coping’’ (e. 1985). It measures four factorially distinct coping strategies. On this sample. . entering trait EI on its own at step 1...82.. viz. Analyses involved two-step hierarchical regressions.70 for the ‘‘rational’’.g. and . Procedure Participants were given a battery of questionnaires.80. This questionnaire consists of five items and measures global life satisfaction (e. viz. ‘‘Take action to change things’’) and ‘‘detached coping’’ (e.79. TRAIT EMOTIONAL INTELLIGENCE 31 Measures: Criteria Emotion Control Questionnaire (ECQ)*Rehearsal Scale (Roger & Najarian. ‘‘rational coping’’ (e. which they completed in class or in their own time. These results support .82.. Coping Styles Questionnaire (Roger. On this sample. to investigate incremental validity.g. Trait EI was a statistically significant negative predictor of ‘‘rehearsal’’ (rumination) at both steps of the regression equation. San Diego] at 08:27 09 June 2015 ways my life is close to my ideal’’). ‘‘emotional’’. two of which are adaptive. ranging from ‘‘always’’ to ‘‘never’’. the internal consistency was . Instructions were presented directly on the questionnaires and participation was voluntary. Jarvis.85. the internal consistencies were . ‘‘detached’’. the results will be presented succinctly.. This questionnaire comprises 60 items assessing how one typically reacts to stress. & Najarian. 1989). Participants were asked to respond on a 6-point Likert scale. ‘‘emotional coping’’ (e. All relevant statistical details are given in Table 1.g. and ‘‘avoidance’’ coping styles. Participants were asked to respond on a 7-point Likert scale.g. ‘‘Just take nothing personally’’) and two are maladaptive. 1993). ‘‘I remember things that upset me or make me angry for a long time afterwards’’).

San Diego] at 08:27 09 June 2015 PETRIDES.23 /. R2adj /.41 .63** .88**.71 /.277 2.389 R2adj /.55**.50**.476 5.053 0.11**.79** /.04 .568 8.222 2. 159)/7. R2adj /. F (6.30** .120 1. F (1.343 4.98**. 164)/39.40**.57** . F (1.318 R2adj /.060 0. 159)/17.33 C .418 4.110 1. 164)/74.182 2.242 3.86**.159 1.042 0.192 b t b t b t b t b t b t Trait EI /. F (1.457 R2adj /.59**.54**.68 /.594 9.00* /. 159)/16.161 1. 164)/78. 159)/18.44** .382 R2adj /. F (6.84 /. 164)/62.67 /.21 A /.141 1.218 3. 164)/89.57** .53**. aOne outlier was removed from step 2 of the regression.388 3.36** Trait EI /. 32 Downloaded by [University of California.57** /.017 0.93** .33 (step 2) Note : *p B/.189 R2adj /.109 1.082 1.042 0.24** . F (6.25* /.016 0.74 .61 . F (6.058 0. PÉREZ-GONZÁLEZ. .68** (step 1) N .315 R2adj /.66**.308 R2adj /.368 3.123 1.017 0.88 .042 0.43 /.003 0.362 R2adj /.69 /.77 .045 0.31** .59 .024 0.05.28** /.87 /.45** .84** .01.112 Step 2 F (6.83 O .04**.440 6.057 0.40 E .53 /.374 4. F (1. 159)/24.24 .29 /.273 R2adj /.518 4.586 6. F (1. F (6.111 1.65** /.096 1.349 R2adj /.559 8.180 2. FURNHAM TABLE 1 Study 1: Hierarchical regressions with trait EI entered at step 1 and the Big Five entered at step 2 Rational coping Detached coping Emotional copinga Avoidance coping Rumination Satisfaction with life (Adaptive) (Adaptive) (Maladaptive) (Maladaptive) Step 1 F (1. **p B/. 164)/21.46** /. 159)/13.526 7.

especially in relation to incremental validity. Similar results were obtained with ‘‘satisfaction with life’’. we incorporated the four distinct coping styles from Study 1. but the more interesting question in this case was whether it does so incrementally over the Big Five. Spence. 2005. The hypotheses were the same.. except ‘‘avoidance coping’’. & Egan. 2001. H1a) and a negative predictor of the two . 2002). & Caputi. Tett et al. It was also a statistically significant negative predictor at step 1 of the two regressions with maladaptive coping styles (‘‘emotional’’ and ‘‘avoidance’’). as hypothesised. but hitherto unexamined. partially. For purposes of replication and comparison. San Diego] at 08:27 09 June 2015 step 2. The second aim was to investigate the validity of a different trait EI measure that has been specifically designed to cover the sampling domain of the construct comprehensively. The third aim was to expand the nomological network of trait EI by exploring its relationship to theoretically relevant. 2004. Subsequently.. trait EI was expected to be a positive predictor of the two adaptive coping styles (‘‘rational’’ and ‘‘detached’’. In short. variables. With respect to criterion validity. Discussion Trait EI was a reliable predictor of all criteria in the study. i. Oades. at Downloaded by [University of California.e. hypotheses H1a to H4a were fully supported. a set of four hierarchical regressions was carried out with each of the coping styles (two adaptive and two maladaptive) as dependent variables. These results support hypotheses H3a and H3b. Saklofske. TRAIT EMOTIONAL INTELLIGENCE 33 hypotheses H1a and H1b. These results support hypotheses H1b to H4b. Moreover. There is plenty of evidence that trait EI relates meaningfully to other variables (Austin. Partialling out all five personality dimensions did not nullify the construct’s associations with any criterion. most relationships were incrementally valid over the Big Five personality dimensions. it reached significance only in the equation with ‘‘emotional coping’’ as the criterion. These results support hypotheses H4a and. as hypothesised.. Trait EI was a statistically significant positive predictor of the two adaptive coping styles (‘‘rational’’ and ‘‘detached’’) at both steps of the hierarchical regressions. H4b. It is clear from the results of this study that it does. STUDY 2 The first aim of this study was to replicate the main findings of Study 1. thus supporting hypotheses H2a and H2b. Schutte et al. trait EI was incrementally associated with the criteria. However. Trait EI was positively associated with life satisfaction and the two adaptive coping styles and negatively associated with rumination and the two maladaptive styles. 2005. Wong & Law.

86 years (SD /6. H6b.17 years). Study 2 also looks at constructs from the clinical. FURNHAM maladaptive coping styles (‘‘emotional’’ and ‘‘avoidance’’. The last criterion in the study was aggression. Because high trait EI individuals believe they can observe and control their emotional reactions.. social. It was further hypothesised that the negative associations would persist after partialling out Big Five variance (H3b and H4b). see also Schutte et al. respectively). In both cases. Sample 1 comprised 200 individuals (75 males and 125 females) with a mean age of 22. we hypothesised a positive relationship with all three self-monitoring criteria (H5a. In this study. PÉREZ-GONZÁLEZ. Furthermore.03 years). 1984). Most participants were single (:/79%) and did not have an undergraduate degree (:/78%). Because high trait EI individuals believe they can regulate emotions and their expression. it was hypothesised that these relationships would remain statistically significant after partialling out Big Five variance (H1b and H2b. ‘‘anger’’ H10a. and personality domains. Method Participants The study employed two samples examining different criteria. San Diego] at 08:27 09 June 2015 Beck. we expected the negative associations with ‘‘anger’’ and ‘‘hostility’’ to remain statistically significant after partialling out Big Five variance (H10b and H11b. we examine the two distinct subcomponents of self-monitoring. which covers four distinct components. As in Study 1. H7b). Depression and dysfunctional attitudes are two straightforward criteria to be used in the validation of trait EI.99 years (SD /6. 34 PETRIDES. ‘‘ability to modify self-presentation’’ and ‘‘sensitivity to emotional expression. Most participants were single (:/84%) and did not have an undergraduate degree (:/87%). ‘‘verbal’’ H9a. and ‘‘hostility’’ H11a). respectively).’’ as well as the global construct itself (Lennox & Wolfe. we would expect negative associations (H3a and H4a) because high trait EI individuals believe they can regulate their emotions to stave off depressogenic cognitions that may trigger disorders when combined with stressful life events (Clark & Downloaded by [University of California. 1999). H2a). We further hypothesised that these relationships would persist after controlling for Big Five variance (H5b. Sample 2 comprised 154 individuals (30 males and 124 females) with a mean age of 21. as operationalised by Buss and Perry’s (1992) aggression questionnaire. 2001). H6a. The construct of self-monitoring was introduced by Snyder (1974) to account for individual differences in self-presentation and expressive behaviour. viz. . we hypothesised that they would score lower on all four facets of aggression (‘‘physical’’ H8a. H7a.. We did not advance any incremental validity hypotheses for the instrumental components of aggression (‘‘physical’’ and ‘‘verbal’’) because they concern purposeful and context-specific behaviour.

Measures: Criteria *sample 1 Coping Styles Questionnaire (Roger et al. and Conscientiousness (. Petrides & Furnham. Dysfunctional Attitudes Scale (DAS.79). Participants were asked to indicate how frequently they experienced certain depressive symptoms during the ‘‘past week’’ (e. Center for Epidemiologic Studies Depression Scale (CES-D. . 1.84). .82 (‘‘rational coping’’). The DAS was developed from a cognitive theory perspective and consists of two 40-item parallel forms. Extraversion (. ranging from ‘‘completely disagree’’ to ‘‘completely agree’’. free of charge.g. On this sample. The TEIQue consists of 144 items and 15 subscales. The internal consistency of the scale on this sample was . For the purposes of the present study. 1977).64).89 on samples 1 and 2.83 and . Luminet. . and .77).84 (‘‘detached coping’’).84 and . versions.83 (‘‘emotional coping’’).87). and translations are available from the first author.00)1. A detailed psychometric analysis of the inventory is presented in Mikolajczak. for research purposes. we selected the 20 odd-numbered items from Form A.68 (‘‘avoidance coping’’).. Radloff. 1993). ranging from ‘‘Rarely or none of the time (less than 1 day)’’ to ‘‘Most or all of the time (5 to 7 days)’’. 1978). Downloaded by [University of California. and Roy (in press). the NEO PI-R was used to assess five-factor personality and we focused exclusively on global scores. were: Neuroticism (. predicated on trait EI theory and covering the sampling domain of trait EI comprehensively (Petrides. measuring depressogenic attitudes and beliefs.92. Participants responded on a 7-point Likert scale.80 and .87 and . The internal consistencies of the five factors on samples 1 and 2. Weissman & Beck. The CES-D uses a 4-point Likert scale response format. San Diego] at 08:27 09 June 2015 NEO PI-R (Costa & McCrae. Leroy. Agreeableness (. respectively. TRAIT EMOTIONAL INTELLIGENCE 35 Measures: Predictors Trait Emotional Intelligence Questionnaire (TEIQue v. Openness (.79 and . The DAS has been found to differentiate depressed patients from nondepressed psychiatric controls and nondepressed normal controls 1 All TEIQue forms. the internal consistencies were . 1992). respectively. The internal consistency of the full scale was . 20-item measure of depressive sympto- matology designed specifically for use in non-clinical settings.. ‘‘I was bothered by things that usually don’t bother me’’).91 and . As in Study 1. 2001. 2003). This is a well-established. This questionnaire was described in Study 1.

79. The RSMS is one of the most widely used measures of self-monitoring.80. Hollon. 1986). ‘‘physical aggression’’.87. the internal consistencies of the four subscales were . p/.85 for ‘‘ability to modify self-presentation’’. Aggression Questionnaire (AQ. viz.69. and six measuring ‘‘sensitivity to emotional expression’’. ‘‘anger’’. It consists of seven items measuring ‘‘ability to modify self-presentation’’. entering trait EI on its own at step 1. respectively.81. Sample 1 All relevant statistical details are given in Table 2. but did not attain. 36 PETRIDES. Downloaded by [University of California. Trait EI was a reliable positive predictor at both steps2 of the hierarchical regressions with the two adaptive coping styles (‘‘rational’’ and ‘‘detached’’). & Lumry. Lennox & Wolfe. and the global scale score. 1984). . Instructions were presented directly on the questionnaires and participation was on a voluntary basis. PÉREZ-GONZÁLEZ. thus supporting hypotheses H1a and H1b.69.. to investigate criterion validity. Results Due to the large amount of data. It is responded to on a 5-point Likert scale. measuring four distinct aggression subscales. The questionnaires took approximately 90 minutes to complete. 1992).181. t /1. the internal consistencies were . All analyses involved two-step hierarchical regressions. On this sample. which they completed either in class or in their own time. Buss & Perry. Procedure Participants at three British universities were given a battery of questionnaires. San Diego] at 08:27 09 June 2015 . ‘‘sensitivity to emotional expression’’. Outliers were treated as in Study 1. and ‘‘hostility’’. On this sample. . The internal consistency of the scale on this sample was . Kendal. respectively. the results will be presented succinctly.80. statistical significance (bTEI /. 1983. It was also a reliable negative predictor at both steps of the hierarchical regressions with the maladaptive coping styles 2 In the regression with ‘‘detached coping’’ as the criterion.09). and adding the Big Five personality dimensions at step 2. The AQ comprises 29 items. ‘‘verbal’’. to investigate incremental validity. and . Measures: Criteria *sample 2 Revised Self-Monitoring Scale (RSMS. the partial coefficient for trait EI at step 2 approached. and . FURNHAM (Hamilton & Abramson.82. Participants responded on a 6-point Likert scale. .

bp/.010 0. 194)/58.081 1.260 2.032 0.090 1.39**.11** .052 0.40 /.097 1.181 2.024 0.42* .72** .31 /.456 7.17 .89** N .68 .45**.66 E .039 0.019 0. F (6.152 b t b t b t b t b t b t TRAIT EMOTIONAL INTELLIGENCE Trait EI (step 1) /.119 2.213 3.02**.015 0. R2adj /. F (6. F (1.81**.66 /.65** . 188)/6. 195)/144.598 R2adj.269 R2adj /.422 R2adj /.488 R2adj /.010 0.56 /.92** /.04** . F (6.032 0.37* /.27** /.19**.05**. F (1.23 .211 2.613 9.12 /.24* Trait EI (step 2) /.Downloaded by [University of California.038 0.54** .01. 37 . F (6. F (1.011 0.198 2. **p B/.17** .652 12.665 12.584 6. 188)/31.09 /.07* Note : *p B/.49* .36** .228 R2adj /.10 O .106 1. R2adj /.83**.05.173 2.181 1. 188) /49. 193) /50.23 /.58 /. 188)/12. F (1.33** /.070 1.69b /.618 10.23 A /.09.74**.96** .16* /.012 0.382 R2adj /.451 R2adj /. 193)/152.062 0.76 C .199 2. 193)/34.009 0.75**.390 5.439 R2adj /.148 Step 2 F (6.379 R2adj /.02** /.86 .21 /.378 3. San Diego] at 08:27 09 June 2015 TABLE 2 Study 2 *sample 1: Hierarchical regressions with trait EI entered at step 1 and the Big Five entered at step 2 Dysfunctional Rational coping Detached coping Emotional coping Avoidance coping Depression attitudesa (Adaptive) (Adaptive) (Maladaptive) (Maladaptive) Step 1 F (1.38 . 189)/27.481 7. 193)/119.66**.33 .199 2.204 R2adj /.559 5. aOne outlier was removed from both steps of the regression. 188) /21.503 5./. F (1.27 .065 0. F (6.30* /.11** /.90**.282 3.59**.

.. These results support hypothesis H11b. H4a. but not ‘‘anger’’. With respect to aggression. It could well be the case that some measures have more desirable psychometric properties than others. in press-b).. in press-a) does not have to be synonymous with a particular measurement instrument. Discussion The findings have three important implications for the construct validity of trait EI. 2003. but not H9a. These results provide support for hypotheses H5a to H7a and H5b to H7b. thus supporting hypotheses H2a and H2b. At step 2. 2001. and despite the fact that they were not tied to any hypotheses. where trait EI did not reach significance levels. 2002). and H11a. 38 PETRIDES. thereby supporting H8a. but we must reiterate that our conceptualisation (Petrides et al. 2002. with the Big Five added in the equations. but not ‘‘verbal aggression’’). Petrides et al. San Diego] at 08:27 09 June 2015 regressions with the three self-monitoring variables (‘‘ability to modify self- presentation’’. Schutte et al. H3b. Schakel. Second. they show that extant trait EI measures tend to produce convergent results (cf. & Thijs. where trait EI was a reliable negative predictor at both steps. these results strongly support the incremental validity of the construct over higher-order traits (see also Saklofske et al. but not H10b. and ‘‘global self- monitoring’’). First. Trait EI was a statistically significant positive predictor at both steps in the hierarchical Downloaded by [University of California. and ‘‘hostility’’. H10a. we also report in Table 3 the second steps of the hierarchical regressions with ‘‘verbal’’ and ‘‘physical’’ aggression. FURNHAM (‘‘emotional’’ and ‘‘avoidance’’). thus supporting hypotheses H3a. Van der Zee. 2001. 2003. Wong & Law. trait EI was a significant negative predictor of ‘‘hostility’’. in line with empirical demonstrations of discriminant validity. Trait EI was related to the clinical variables (depression and dysfunctional attitudes) as well as to the personality and social variables (self-monitoring and aggression). PÉREZ-GONZÁLEZ. as hypothesised. Sample 2 All relevant statistical details are given in Table 3. Similar results were obtained in the hierarchical regressions with ‘‘depres- sion’’ and ‘‘dysfunctional attitudes’’. trait EI was a negative predictor of three of the four subscales at step 1 (‘‘physical’’. they confirm previous research showing that the effects of trait EI span several different basic and applied domains (Petrides & Furnham. The only two hypotheses that were not . 2004).. Van der Zee & Wabeke. coping styles findings in Studies 1 and 2). ‘‘sensitivity to emotional expression’’. Third. For purposes of completeness. and H4b. ‘‘anger’’. involving the identification of a distinct trait EI factor in personality factor space (Petrides & Furnham.

320 2.123 1.042 0.08* /.51** /.58** . F (1.Downloaded by [University of California.150 2.048 0.39 .03 /.56 /.062 0.53* /.109 1. 148)/37.063 0.59**. aOne outlier was removed from both steps of the regression.55 /.132 1.56* /.081 0.05. F (6.119 1. 149)/1. 141)/5. F (1.533 7.06. F (6. R2adj /. 142)/10.54**. 149)/25.449 6.381 5.104 1.074 1.505 TRAIT EMOTIONAL INTELLIGENCE b t b t b t b t b t b t Trait EI (step 1) .473 5.449 4.030 R2adj /.28 /.59 .69**.206 2. F (1. F (1.03 /.67*.139 R2adj /.287 R2adj /.150 1.014 0.69** /. 149)/12.20* Trait EI (step 2) .230 2.603 9. 141)/6.196 R2adj /. 142)/16.211 2. F (6.164 R2adj /.113 1.40** E .210 R2adj /.66** .27**.93**.186 R2adj /.38* /. 39 .60 .394 2.040 0.55 /.370 4.54 O /. 142)/7.46 .61**. F (6.27**. 141)/26.04**.473 3.062 0.228 2.05 /.22** /.084 1.01.097 1.20** N .84** .65 .57 . 148)/84. F (6.003 0.309 4.94** .004 0.04 .387 R2adj /.03** /.225 2. F (1. San Diego] at 08:27 09 June 2015 TABLE 3 Study 2 *sample 2: Hierarchical regressions with trait EI entered at step 1 and the Big Five entered at step 2 Ability to modify Sensitivity to self-presentation emotional expression Physical aggression Verbal aggression Anger Hostilitya Step 1 F (1.109 1.32 /.070 0.11** .63 .15 .14 A /.58**.072 R2adj /. 148)/5.360 Step 2 F (6.78** /. **p B/.192 2.51** C .279 3.55** /.80**.216 3.48 .05* Note : *p B/.000 R2adj /. R2adj /.17 /.88 /.

& Mack. we hypothesised that trait EI would be a reliable negative predictor of depression (H1a) and that it would remain such in the presence of the two dimensions of affectivity (H1b). Studies exploring this question have revealed significant relationships (Ciarrochi. Matthews. probably because the conceptual correspondences between them have been repeatedly highlighted (Davies. McKenley. Haas.. 40 PETRIDES. 2000. therefore. Although most research has focused on social and personality variables. although the relationship was not significant after partialling out Big Five variance. 2000). Neuroticism and Agreeableness were the only significant predictors in the second step of this regression (positive and negative. Zeidner.g. we believe that trait EI is at least as likely to play an important role in clinical contexts (Petrides. Given the links between personality and mood (Canli. rather than qualitative. Chan. 2003. 2004. which characterises high trait EI individuals. 2002. STUDY 3 The aim of this study was to examine the criterion and incremental validity of trait EI in relation to both new variables and a new baseline. Davison. & Constable. The results on depression and dysfunctional attitudes suggest that very low trait Downloaded by [University of California. One example is the strong negative relationship with depression (e. Bing.. the latter is related to trait EI. McCrae. see also Study 2). PÉREZ-GONZÁLEZ. Nestadt. Omura. 1998. as expected. Davies et al. FURNHAM borne out by the data concerned step 1 of ‘‘verbal’’ aggression (H11a) and step 2 of ‘‘anger’’ (H10b). The fact that many mental disorders represent quantitative. a proposition we investigate in the last study of this paper. substituting the Big Five with the two basic dimensions of mood (positive and negative affectivity). it is meaningful to ask how. 2000). Watson. and to what extent. Schutte. & Hollander. San Diego] at 08:27 09 June 2015 EI may have psychopathological implications. respectively). arises as to whether trait EI has incremental predictive validity over the basic dimensions of mood. Simunek. Saklofske et al. ‘‘Anger’’ was negatively related to trait EI. 2003). Amin. which may have adaptive as well as maladaptive implications (Petrides & Furnham. & Roberts. 2002). & Bajgar. 2001. 2001). Watson. The lack of association between trait EI and the ‘‘verbal’’ AQ factor is likely due to the fact that the latter is conceptually confounded with assertiveness. Therefore. The question.. Incremental validity studies have focused primarily on whether trait EI explains criterion variance over and above personality. & Roberts. The criteria in this study exclusively comprise variables of clinical relevance. Ghorbani. 1998. abnormalities on personality dimensions (Bienvenu. Malouff. 2002. Petrides & Furnham. which we sought to replicate while controlling for positive and negative affectivity. Stankov. .

Chorot. Partici- pants were students in three Spanish universities.. & Eaton.00). Lostao. the experts were asked to use a 5-point Likert scale to rate each of the inventory’s 144 items for ‘‘clarity and comprehen- sibility’’. Items were translated into Spanish. ‘‘dependent’’ (H9a). Clark. Consequently. Unclear items were reviewed and translated again. Santed. Subsequently.33 years). 2001. 2001). Study 3 examines this possibility with reference to the personality disorders (PDs) incorporated in the Tenth Revision of the International Classification of Diseases (ICD-10. TRAIT EMOTIONAL INTELLIGENCE 41 Samuels. Howard.e. with median ratings of 5. 1970. The questionnaire was adapted into Spanish under the direction of the second author (see Pérez. 1989) was used to measure individual . We used the Spanish adaptation of the TEIQue (see Study 2). ‘‘obsessive-compulsive’’ (H8a). Finally. 1999) of the 20-item PANAS (Watson. Costa. 1992). The average age was 23. we hypothesised that trait EI would be negatively related to the ICD-10 PDs. three bilingual individuals (1 philologist and 2 educationists) evaluated and confirmed the linguistic equivalence of the English and Spanish versions. and ‘‘avoidant’’ (H10a). & Valiente. Although high trait EI scores are not always adaptive or functional (Petrides et al. 1. San Diego] at 08:27 09 June 2015 typal’’ (H4a). ‘‘histrionic’’ (H5a). Measures Trait Emotional Intelligence Questionnaire (TEIQue v. then back-translated into English and compared. including ‘‘paranoid’’ (H2a). i. ‘‘schizo- Downloaded by [University of California. Most items (133) were rated ‘‘very clear’’. Widiger. Fifteen experts (university lecturers in education or psychology) were asked to rate each of the inventory’s 15 subscales for content validity. We further hypothesised that trait EI would remain a statistically significant predictor after controlling for positive and negative affectivity (H2b to H10b). The Spanish version (Sandı́n.90.. the degree to which they believed the subscales are relevant to the construct. Joiner. On this sample. 2003) in line with current test adaptation guidelines (Hambleton. they do indicate an overall healthy mental state in normal adults. ‘‘antisocial’’ (H6a) ‘‘borderline’’ (H7a). 1992) suggests that very low trait EI may have psychopathological consequences. Method Participants Two hundred twelve individuals participated in the study (37 males and 175 females). Most subscales (12) were rated as either ‘‘relevant’’ or ‘‘essential’’ to the construct. Eysenck. ‘‘schizoid’’ (H3a). & Tellegen.07 years (SD /3. the internal consistency of the global score was . World Health Organization. Positive and Negative Affect Schedule (PANAS). in press-a).

The results for ‘‘depression’’ were consistent with those of the British sample.89 and . Beck Depression Inventory (2nd ed. Procedure Participants from three Spanish universities completed a battery of questionnaires in class. on a 5-point Likert scale. Navarro. The IPDE is a semi-structured interview schedule. 1996) of the IPDE questionnaire. The BDI-II consists of 21 items that are responded to on a 4-point scale. BDI-II). Steer. entering trait EI on its own at step 1. Outliers were treated as in the previous studies. The internal consistencies of the scales. PÉREZ-GONZÁLEZ. to investigate incremental validity. The materials took approximately 90 minutes to complete.87. how they feel ‘‘usually’’. We employed the Spanish version (Pérez & Rubio.85. & Sartorius. Participants were asked to indicate. At step 2. which reduces the likelihood of obtaining statistically significant effects. Results Due to the large amount of data. 2001) of the second edition of the BDI (Beck. trait EI was a statistically significant negative predictor at step 1 in all nine hierarchical regressions. which is typically used as a screening instrument. San Diego] at 08:27 09 June 2015 of the scale on this sample was . The Spanish version (Sanz. the results will be presented succinctly. ‘‘narcissis- tic’’ was excluded from further analysis. Janca. designed to produce diagnoses consis- tent with the ICD-10 and DSM-IV classifications (Loranger. . Analyses involved two-step hierarchical regressions. respectively. the internal consistencies of positive and negative affectivity were . Due to its particularly low alpha. followed by the two mood dimensions (positive and negative affectivity) at step 2. to investigate criterion validity. along with a brief description of each. ‘‘schizotypal’’. are given in Table 4. thus supporting hypotheses H2a to H10a. International Personality Disorder Examination (IPDE). statistically significant results were obtained with the following criteria: ‘‘paranoid’’. All relevant statistical details are presented in Tables 5 and 6. & Brown.. 42 PETRIDES. ‘‘dependent’’. 1996) was used to measure depression. thus supporting hypotheses H1a and H1b. & Vázquez. It is worth noting that the alphas were uniformly low on this sample. FURNHAM differences in positive and negative affectivity. On this sample. ‘‘schizoid’’. It comprises 77 dichotomous (true/false) items and yields dimensional scores on 10 distinct PDs. The internal consistency Downloaded by [University of California. Instructions were shown on the questionnaires and participation was voluntary. With respect to the personality disorders. ‘‘borderline’’. 1997).

preoccupation with introspection.868 /.42 Unconcern for the feelings of others.32 Indifference to praise or criticism. Obsessive-compulsive (8) .58 Fear of being left to care for oneself. perfectionism.44 Self-dramatisation. Dependent (8) . Avoidant (8) . inappropriate seductiveness. threats or acts of self-harm.125 .570 .67 Persistent feelings of tension and apprehension. Schizoid (7) .113 .361 . TRAIT EMOTIONAL INTELLIGENCE Antisocial (7) .446  . inferiority complex. recurrent suspicions. Note : Numbers in parentheses indicate the number of items in each scale.572 . exaggerated expression of emotions.541 . Loadings greater than j.375 .642 . San Diego] at 08:27 09 June 2015 TABLE 4 Factor pattern matrix and descriptive information for the IPDE ICD-10 scales IPDE scales Factor 1 Factor 2 Alpha Symptomatology Paranoid (7) .60 Unusual perceptual experiences.34 Feelings of excessive doubt and caution. Loadings less than j. excessive pedantry. tendency to bear grudges persistently. suspiciousness of others. consistent choice of solitary activities. Histrionic (8)  . Borderline (9) . odd behaviour.580 . 43 . disregard for social norms.374 . incapacity to maintain enduring relationships.10j are suppressed.114 . rigidity and stubbornness. avoidance of activities due to fear of criticism.Downloaded by [University of California. disturbances in and uncertainty about self-image.145 . Schizotypal (9) . inappropriate affect.230 .30j are in bold. quasi-psychotic episodes. shallow and labile affectivity.573 /.63 Excessive efforts to avoid abandonment.53 Excessive preoccupations. undue compliance with others’ wishes. inability to take decisions on one’s own.

68. The internal consistencies of the two factors were .12. respectively. but also broader.06. San Diego] at 08:27 09 June 2015 grouped into a small number of more reliable factors. where it can be seen that the first factor mainly concerns psychoses (‘‘psychosis’’). t/1.73. and H10b.01. for relevant results with other trait EI measures).453. factor mainly concerns neuroses (‘‘neurosis’’). The scree plot and Kaiser eigenvalue criterion converged on a two-factor solution. which are known to underlie psychopathology (Watson. we performed a principal axis factor analysis of the nine PDs to determine whether they can be Downloaded by [University of California.01. F(1. contributing to the aetiology of mental disorders that are only partially related to emotional malfunctioning (e. t/ 2. as well as at step 2.75 and . 2004. whereas the second.01. but did not attain. Low trait EI can be seen as a global susceptibility factor.198. pB/.186. PÉREZ-GONZÁLEZ. ‘‘Neurosis’’: R2adj /. 210) / 147. p /. ‘‘Psychosis’’: R2adj /. predisposing individuals to a range of mental abnormalities. Its effects are not only stronger than those of affectivity. As hypothesised. p B/. accounting for 53. The oblimin-rotated factor pattern matrix for this solution is given in Table 4. 210) /42. respectively). Due to the low alphas of many IPDE scales. F(3. 207) /63.642. the partial regression coefficients for trait EI in the equations with ‘‘antisocial’’ and ‘‘obsessive-compulsive’’ approached. these results provide full support for hypotheses H2b.40% of the variance.44. 207) /17. t /1. p B/. the negative associations held up after partialling out individual differences in dispositional mood. see also Leible & Snell.01.90. at step 1. p B/. bTEI / /.410. pB/. 44 PETRIDES. bTEI / /. More important.01. pB/.01. p B/. H3b. ‘‘Psychosis’’: R2adj /. t/12. F(3.409. 2000). antisocial personality). H9b.164. FURNHAM and ‘‘avoidant’’.01. bTEI / /. In addition. H7b.g.471.57. smaller.73. Discussion The findings suggest that trait EI may have an important diagnostic role to play in relation to virtually all PDs included in the two major classification systems (ICD-10 and DSM-IV. t /5. trait EI scores were negatively related to all PDs in the IPDE.50.01. Taken together.01. Trait EI was a statistically significant negative predictor of both.269. F(1. This should be expected because the construct extends beyond core emotional self-perceptions to . ‘‘Neurosis’’: R2adj /. H4b. t /6. p B/. We subsequently regressed the two factors on trait EI and the affectivity dimensions using two-step hierarchical regressions as above. bTEI / /. but no support for hypothesis H5b (‘‘histrionic’’).91. p /.08 and bTEI / /.51. It bears repeating that these results were obtained in spite of the low internal consistencies of the IPDE scales and were replicated when PDs were grouped into two factors with improved Cronbach alphas. significance (bTEI / /.. partial support for hypotheses H6b and H8b.195.

288 4. San Diego] at 08:27 09 June 2015 TABLE 5 Study 3 *part I: Hierarchical regressions with trait EI entered at step 1 and the PANAS affectivity dimensions entered at step 2 Depressiona IPDE paranoid IPDE schizoidb IPDE schizotypala IPDE histrionicb Step 1 F (1.38 .579 10.218 R2adj /.86** PAN-N . 206)/3. PAN-N / PANAS negative affectivity.040 0.038 0.202 2.041 0. R2adj /.42 . F (1. 209)/16.28** /.88**. F (3.36 Note : *p B/.14* /.27** /.37 .033 Step 2 F (3.96**.85** /.250 3. 207)/35.066 R2adj /.82**.10** .492 8.81**.274 4. **p B/.114 1.06** .332 R2adj /. aOne outlier was removed from step 1 of the regression.039 b t b t b t b t b t TRAIT EMOTIONAL INTELLIGENCE Trait EI (step 1) /.63**.05.208 2. F (1.80*.194 2.42* /.30** .64** /.331 R2adj /.92** /.431 6. F (3. 207)/20.182 R2adj /.133 1.49 . R2adj /. F (3.Downloaded by [University of California. 206)/50.17** /.201 2. bOne outlier was removed from both steps of the regression.239 R2adj /.029 0.147 1.12** /. F (1. 210)/47. 209)/105. F (1.18**.071 R2adj /.417 R2adj /.394 6. 206)/5. F (3.43**. PAN-P /PANAS positive affectivity.58 PAN-P /.384 4.51**.281 2.44 Trait EI (step 2) /. 209)/8.94**. 45 .01. 209)/66.

72**. F (3.245 R2adj /. **p B/. F (1.117 1. 207)/4.88** /. 210)/16.206 R2adj /. 209)/114.91c /.01 .57**.190 2.331 Step 2 F (3.73c /.594 10.485 5. R2adj /. PAN-P /PANAS positive affectivity. FURNHAM TABLE 6 Study 3 *part II: Hierarchical regressions with trait EI entered at step 1 and the PANAS affectivity dimensions entered at step 2 IPDE antisociala IPDE borderlineb IPDE OCD IPDE dependent IPDE avoidant Step 1 F (1.40**.06** /.05.11**.69** /.31**.44** /.60 .03* /. PÉREZ-GONZÁLEZ.36** .98 /. PAN-N /PANAS negative affectivity. bOne outlier was removed from both steps of the regression.244 3. F (3.46** /.068 R2adj /.458 7. San Diego] at 08:27 09 June 2015 PETRIDES. 207)/9. F (3.046 R2adj /.350 R2adj /.578 10. F (1.370 3. aThree outliers were removed from both steps of the regression.270 4.386 4. .104 R2adj /. 210)/105.78**.01.06.228 3.66 Trait EI (step 2) /.41** Note : *p B/.146 1.102 1.386 R2adj /.27** .139 2.051 0. F (1.015 R2adj /.333 b t b t b t b t b t Trait EI (step 1) /. 204)/4.080 0. 207)/35.74 . R2adj /. F (3.198 1.230 2.06* /.91**. cp/. 210)/55.11*.186 1.69**.50**.090 1.84** .28** PAN-N .45 PAN-P . 206)/44. 46 Downloaded by [University of California. F (1. 207)/23.

In contrast to criterion validity.. models operationalised through self-report questionnaires and theorising about abilities.. therefore. 2001). Therefore. but at the various models underpinning them. not least because they can account for the co-occurrence (comorbidity) of PDs and thus contribute to the identification of common aetiologies. it would be interesting to . Global susceptibility factors are clinically useful.. ‘‘assertiveness’’. which limits their applicability in treatment contexts (Matthews et al. the investigation of the incremental validity of self-report measures of EI has led to considerable debate in the literature. such as self-esteem (Petrides & Furnham. The downside is that such factors are insufficient to explain the wide range of disorder-specific symptomatology. 2002. 1998).g. in press-a). are flawed in ways that have been discussed exhaustively in the scientific literature. the TEIQue is predicated on a theory that uniquely operationalises EI in a manner that resolves major criticisms about construct validity (Petrides et al. While this issue has broader implications for trait EI.. ‘‘adaptability’’. There can be little doubt that the constellation of emotion-related self-perceptions and dispositions that trait EI encompasses is implicated in numerous areas and will.. social. predictive value over and above the basic personality dimensions (e. The emphasis on questions of discriminant and incremental validity may be due to the fact that certain trait EI facets are already included in the established trait taxonomies (e. San Diego] at 08:27 09 June 2015 scope for research in clinical settings. Zinbarg & Barlow. A persistent criticism is that these measures add little. 2002. 1996). The tendency to lump self-report measures together and criticise them irrespec- tive of their theoretical basis is problematic because it fails to differentiate between theories and measurement vehicles (Jensen. As far as simple prediction is concerned. in what follows. or hitherto allegedly unknown dimensions of individual differences. Matthews et al. Setting aside its psychometric advantages. competencies. we do not believe they should be directed at self-report measures in general. In contrast. 2004). ‘‘empa- thy’’). and clinical psychology. While we generally second these reservations. be of interest to researchers and practitioners in many different fields. Schulte et al. TRAIT EMOTIONAL INTELLIGENCE 47 encompass a number of outcome self-evaluations.. the data bore out 20 out of 21 hypotheses involving variables from the domains of personality. we address incremental validity issues specifically from the perspective of trait EI theory. if any.g. GENERAL DISCUSSION As far as criterion validity is concerned. the findings of this study highlight the Downloaded by [University of California.

Direct comparisons pitching a single trait EI variable against three (or five) personality variables are inevitably inequitable. 2001). 2002). what do we stand to gain by attempting to reconceptualise every construct at the lower levels of personality hierarchies as a mixture of the higher-order traits? Another issue to consider is that incremental validity analyses in this and much other personality research are skewed. rather than an artefact of facet duplication. Indeed. Saklofske et al. Alas. PÉREZ-GONZÁLEZ. 1998). The question critics must address is. most constructs can be expressed as mixtures of the basic personality dimensions (Paunonen. Few personality and social psychology variables would retain much predictive validity following the removal of all of the variance they have in common with the Giant Three or the Big Five. it is far more likely to produce statistically significant associations with external criteria. Bearing the foregoing in mind. ‘‘real-life’’ (e. While trait EI carries only one degree of freedom. Removing overlapping variance stemming from meaningful associations. When we ask whether trait EI. it is precisely why these factors are deemed fundamental. However. However. there is no generally accepted method of quantifying and removing overlapping variance due to duplication of facets. strips trait EI of much. ‘‘personality’’ carries three (or five) and. we note that from a total of 17 hypotheses concerning incremental validity over the Big Five. 15 were borne out by the data. predicts ‘‘over and above personality’’ we are posing an inherently biased question. and experimental (Petrides & . such reconcep- tualisations fail to capture the essence of these constructs (Funder.. in addition to the spuriously overlapping variance stemming from duplica- tion or common method assessment. For example. the correlation between ‘‘emotion expression’’ and Neuroti- Downloaded by [University of California. 48 PETRIDES. but not all *as this paper demonstrates *of its predictive power. San Diego] at 08:27 09 June 2015 cism is the result of a meaningful relationship between those two variables. 9 out of 10 hypotheses concerning incremental validity over mood were also borne out by the data. once the variance of the facets that it shares with the basic personality dimensions has been removed. They complement results based on self-report (e.. It should come as no surprise that some weighted linear combination of the Giant Three or the Big Five personality factors accounts for a large amount of variance in most personality constructs (see O’Connor. These results were obtained with two different trait EI measures and data collected in two different countries. such as those employed in this paper. 2004).g. 2003).. The crux lies in recognising that ‘‘personality’’ is not a unitary construct. FURNHAM establish whether trait EI can account for statistically significant portions of criterion variance. Overlapping variance can be removed in its entirety through statistical procedures. Petrides et al.. consequently. or any other variable. In addition.g. this drastic approach is problematic because it also removes variance due to valid associations between conceptually distinct constructs.

as we have argued elsewhere (Petrides & Furnham. trait EI predicts criterion variance over and above the basic dimensions of personality and mood. we urge researchers and practitioners to abandon the mushrooming number of models emanating from commercial test user manuals. in this sense. We can. such that higher absolute values will be obtained with criteria that are more affectively relevant. However. Niven. like academic perfor- mance. In the context of so many statistically significant associations with such a broad range of criteria. and skills. The paper’s findings support a basic premise of trait EI theory. questions may arise in relation to the construct’s boundaries and discriminant validity. obviously not all variables are affect-laden. viz. rather than whether it can be discriminated from the major personality Downloaded by [University of California. with findings that are consistent across different measures.. Much trait EI research has been based on global . The similarity of the results obtained through the modified EQ-i (Study 1) and the TEIQue (Studies 2 and 3) corroborates the generality of trait EI theory. So what does this paper tell us about the ‘‘importance’’ of trait EI as an individual differences variable? The evidence shows that trait EI is potentially important. for comparable findings based on another scale). of variable quality and adequacy. the influence of emotions on most aspects of everyday life means that trait EI will be related to many different variables in many different contexts. in press-b). the notion of discriminant validity concerns what the construct is not expected to predict. Luminet. given that a distinct trait EI factor can be isolated in Giant Three and Big Five factor space (Petrides et al. inasmuch as it is implicated in many different domains. these questionnaires provide coverage.. 2004) and strong correlations with heavily affective variables. 2005). of a constellation of emotion-related self-perceptions and dispositions that is located at the lower levels of personality hierarchies. As expected. Consequently. that self-report questionnaires of EI and cognate variables operationalise a construct that is unrelated to capabilities. in press-b). competencies. at one end (see Petrides et al. 2006). at the other end (see Study 2 and Mikolajczak. which provides a platform for the interpretation of data from any EI questionnaire (see Saklofske et al. we must emphasise that EI-related questionnaires are measures of trait EI only in so far as their results are interpreted through the lens of trait EI theory.. As pointed out elsewhere (Petrides.. 2001. like depression and stress. languages. Note that. Rather. We should witness near-zero correlations with IQ and heavily cognitive variables. and countries. therefore. & Mouskounti. San Diego] at 08:27 09 June 2015 dimensions (a question we address in Petrides et al. Pérez et al. 2003) criteria.. & Menil. At the same time. 2006). hypothesise that correlation strength will vary as a function of the affective load of a criterion. in clear favour of trait EI theory. and as this series of studies shows.. 2003. TRAIT EMOTIONAL INTELLIGENCE 49 Furnham.

This limitation can be overcome through the Downloaded by [University of California. Second. Deane. Manuscript received 18 June 2003 Revised manuscript accepted 29 June 2006 REFERENCES Austin. Ciarrochi. E.. & Anderson. Ultimately. Austin. 2000). Personality and Individual Differences. which tend to be based on flawed conceptions. 36. foremost among which is a relative loss in explanatory power. First. J. it has helped retain research focus on the development of the nomological network of the construct. . the TEIQue can be used for detailed analyses at the factor and subscale levels to address the aforementioned limitations of global scores. as opposed to factor or subscale scores. We expect that ongoing research with the TEIQue (Studies 2 and 3) will soon lead to substantial progress in this direction. 1855 1864. & McKenney. with the initial stage of construct operationalisation virtually complete and questions concerning measurement and validity conclusively addressed (Petrides et al. their use has helped develop and sustain a common research database that goes beyond specific models.. because global scores are much less sensitive than factor scores to sampling domain variability across instruments. PÉREZ-GONZÁLEZ. especially the notion that abilities can somehow be measured via self-report. it is the latter of the two that will determine the importance of trait EI as an individual differences variable. Although relevant factor analytic data exist in the literature (e. they preceded the content analysis that gave rise to the sampling domain of trait EI (Petrides & Furnham. 2002. Saklofske. Such analyses would also go some way towards balancing the inherently biased comparisons in incre- mental validity studies that pitch a single degree of freedom for trait EI against multiple degrees of freedom for personality. in press-a). However. The emphasis on global scores has been beneficial for two reasons. 2001) and hence are tied to early specific models. As a multifactorial instrument. New research questions should now be formulated such that they force a shift of emphasis from applied predictive utility to theoretical explanatory power. Petrides & Furnham. Much progress has been achieved in trait EI research. specifically designed to provide a measurement vehicle for trait EI theory. An investigation of the relationship between trait emotional intelligence and emotional task performance. instead of on introspective debates about factor structure. FURNHAM scores. San Diego] at 08:27 09 June 2015 identification of robust and homogeneous clusters of trait EI facets. Global scores are not without drawbacks. little is currently known about the factor structure of trait EI. 50 PETRIDES.g. (2004). 2004. Trait EI theory is general and enables the meaningful interpretation of results from all EI questionnaires. Huang.

D. A. Personality and Individual Differences. & McCrae.. 989 1015.. A.. L. (1999). T. Buss. E. TX: The Psychological Corporation. 31. & Eysenck. J.. Personality. Canli. 25. P. J. T. E. Stankov.. Scientific foundations of cognitive theory and therapy of depression. S... T. (2005). and major depressive disorders and the five-factor model of personality.). . Davies. D. New York: Cambridge University Press. Y. S. S. F. P. V. New York: Wiley.. Toronto. 71 75. De Raad. E. 75.. Barrett. Eysenck. Ciarrochi... (2005). Cronbach. replication and elaboration. P. Bar-On. Steer. Emotional intelligence: In search of an elusive construct. Costa. Beck. type A behaviour. M. 118. The aggression questionnaire. K.. V. New approaches to personality classification (pp. T. & Kline. (1985). 189. Chan. Huang. (1998). Larsen. 63. Behavioral Neu- roscience. (1955). Costa. and L across 34 countries. D. K. 519 536. Clark.. San Diego] at 08:27 09 June 2015 ed. H. E. P. FL: Psychological Assessment Resources. Journal of Nervous and Mental Disease. B. (1995). Journal of Personality and Social Psychology. H. 38.. Z. & Meehl. J. T. Measurement of trait emotional intelligence: Testing and cross-validating a modified version of Schutte et al. G. Personality and Individual Differences. Human cognitive abilities: A survey of factor-analytic studies. 36. Personality and Individual Differences.. 32. & Perry. & Constable. (1992). (2005). Emotional intelligence moderates the relationship between stress and mental health. Deane. Carroll. BarOn Emotional Quotient Inventory: Technical manual. H. & McKenney. 897 904. (1998)... The Big Three or the Big Five *the EPQ-R vs. (1996). T. 452 459. O. Amin. & Eaton.’s (1998) measure. The satisfaction with life scale. Assessing emotional intelligence: Reliability and validity of the Bar-On Emotional Quotient-Inventory (EQ-i) in university students. J. Omura. 197 209. L. E... and daily hassles.. (2001). TRAIT EMOTIONAL INTELLIGENCE 51 Austin. P. D. H. W. S. 38. Construct validity in psychological tests. Diener. 19. W. Odessa. V. Emmons.. J. Personality and Individual Differences. 805 819. & Bajgar. K. (1992). Haas. Samuels. Howard. C. Canada: Multi-Health Systems Inc. A. D. Phobic. Personality and Individual Differences. S. Predicting psychological health: Assessing the incremental validity of emotional intelligence beyond personality. N.. New York: Columbia University Press. & Roberts. A. V.. A dimensional system of psychodiagnostics. A. 555 562. 801  804.. J. L. Therrien. (2002). H. R. S. H. Jr. J. A. F. D.. & Beck. (2004). The Eysenck Personality Questionnaire: An examination of the factorial similarity of P. R. J. Journal of Personality and Social Psychology. (1970). well-being and health correlates of trait emotional intelligence. Draycott. Saklofske. Personality and Individual Differences. Manual for the Beck Depression Inventory (2nd Downloaded by [University of California. 673 687. In A. Jr. (2004). Personality and Individual Differences. B. Psychological Bulletin. R. Day. G.. 154 161.. R. W... 49. 797 812. Journal of Personality Assessment... San Antonio. L. 169 208). & Brown.. 281 302. Austin. panic. T. Dawda. B. & Carroll.). Eysenck. B. R. A double dissociation between mood states and personality traits in the anterior cingulate. 1105 1119... Personality and Individual Differences. J. P. (1997). S. & Egan. Measuring emotional intelligence in adolescents. 18. 28. European Journal of Personality. D.. A. the NEO-PI: A research note. 547 558. The trait-coverage of emotional intelligence. J. Nestadt. D. S. M. The Revised NEO Personality Inventory (NEO PI-R) and NEO Five-Factor Inventory (NEO-FFI) professional manual. T. & Anderson.. Petrides. A. Saklofske. (2000). (1993). Ciarrochi. R. G. 52.. & Griffin. G. Bienvenu. A. (2001). R. Mahrer (Ed. R. R. J. & Hart.. J.

. 197 221. Watson. The validity of the Bar-On Emotional Intelligence Quotient in an offender population. T. 164 172. Emotion and adaptation. (2003). Psicothema. Journal of Personality. & Wolfe. Specificity of depressotypic cognitions in clinical depression. & Abramson. & Sartorius. M. MacCann. M. Zeidner. Revision of the Self-Monitoring Scale. N. P. J. S. Jensen. A. I. D. In R. 72. Goleman. Luminet. Personality and Individual Differences. E... (1998). (2004). (in press). Westport. (2006).. 58. G. Revisiting the predictive validity of emotional intelligence: Self-report versus ability-based measures. R. Journal of Abnormal Psychology. 95. FURNHAM Eysenck. 1349 1364. (1997). The Big Five trait taxonomy: History.). N. Cognitive patterns and major depressive disorders: Downloaded by [University of California. M. Bing. Self-reported emotional intelligence: Construct similarity and functional dissimilarity of higher-order processing in Iran and the United States. R. Leroy. Consensus scoring and empirical option weighting of performance-based emotional intelligence (EI) tests. Leible. Journal of Personality and Social Psychology. C.. European Journal of Psychological Assessment. Handbook of emotional intelligence (pp. Lazarus. and theoretical perspectives. (2000). Psychometric properties of the Trait Emotional Intelligence Questionnaire (TEIQue. & Zeidner. 36. Mikolajczak. E. L. Emotional intelligence: Why it can matter more than IQ. D. 855 862. A quantitative review of the comprehensiveness of the five-factor model in relation to popular personality inventories... A. Matthews. A. Parker (Eds. Mills. Cambridge. In L. C. K. W. Matthews. & Little. B. D. (2001). San Diego] at 08:27 09 June 2015 A longitudinal study in a hospital setting. W. McCrae. The g factor.. John (Eds.. 37. G. O’Connor. (1999). & Roberts. Funder. (1986). R. A. J.. 9. W. 17. & Srivastava. K. New York: Guildford Press. Handbook of personality: Theory and research (2nd ed.. (1997). M. PÉREZ-GONZÁLEZ. R.. J. D. 52. (2002). McBride. 52 PETRIDES. 37.. & Larson. (1995). L. International Journal of Psychology. The next generation of the ITC Test Translation and Adaptation Guidelines. Emotional intelligence: Science and myth. H.. D.. 79 88.. E. Petrides & Furnham. Genetic and environmental contributions to individual differences: The three major dimensions of personality. 92. A. Personality and Individual Differences. 188 203. P. Luminet.. (1991). (2002). 645 662. (2001). Hollon. D. 173 184. John. J. 52 59. Assessment.. 102 138). M. R. 37. 695 706. Personality. A. S. Roberts. E. H. D. . C. Mikolajczak. Janca. & Lumry. (Eds. D. Hambleton... London: Bloomsbury. O. Emotional intelligence from the perspective of the five-factor model of personality. R. & Snell. Borderline personality disorder and multiple aspects of emotional intelligence. Loranger. R. Annual Review of Psychology. Hamilton. 297 308. 263 276).. (1984). Journal of Personality and Social Psychology. 35. C. Nolen-Hoeksema. O’Connor. (2004). & Menil. CT: Praeger. S. measurement.).) (pp. New York: Oxford University Press. N. Kendall. Personality and Individual Differences. MA: MIT Press. Journal of Personality Assessment. A. G. R. P. R. O. Rumination and psychological distress among bereaved partners. P. M. S. & Roy. Assessment and diagnoses of personality disorders: The ICD-10 International Personality Disorder Examination (IPDE). & Kroner. P. Predicting resistance to stress: Incremental validity of trait emotional intelligence over alexithymia and optimism. R. (1983). A. N. T. O. Journal of Abnormal Psychology. Ghorbani. 18...). F. 245 261. S. (2004). 46. UK: Cambridge University Press. Bar-On & J.. Pervin & O. Personality and Individual Differences. Y. Lennox. 393 404. Cambridge... (2002). & Mack. New York: Jossey-Bass. D. Davison. 1893 1902. 2003).. Hemmati. (1990).

& Furnham. B. TRAIT EMOTIONAL INTELLIGENCE 53 Palmer. Pérez. (2004). Personality and Individual Differences. European Journal of Personality. Hierarchical organization of personality and prediction of behavior. D. 1. L. 29. Petrides. Paunonen. V. & Furnham. 845 853. 538 556. V. Measuring trait emotional intelligence. R.. K. Trait emotional intelligence: Moving forward in the field of EI. & Mouskounti. (2006). B. G. V. (2000). Pérez. 15. Applied Psychological Measurement. 313 320.). A. 107 115. OMS (WHO).. European Journal of Personality. Parker.. Emotional intelligence: An international handbook. G. & Najarian. . R. Furnham. & Furnham. A. Doctoral dissertation.. Niven. Personality and Individual Differences. A. San Diego] at 08:27 09 June 2015 Madrid: Ediciones Meditor. Taylor. The location of trait emotional intelligence in personality factor space. (2003).. & Mavroveli. Pita.. Personality and Individual Differences.. & Kokkinaki. Petrides. 35. C. Módulo DSM-IV. 74. The relationship between emotional intelligence and alexythimia. D.. 18.. Frederickson. Petrides. B. & Furnham. V. 619 626. Personality and Individual Differences. Detachment and coping: The construction and validation of a new scale for measuring coping strategies. MA: Hogrefe & Huber. Matthews. & Najarian. The construction and validation of a new scale for measuring emotion control. Downloaded by [University of California. & Stough. Jarvis. Gignac. (2002). Palmer. (2005). (1993). K. The role of trait emotional intelligence in academic performance and deviant behavior at school. C.). Schulze. 277 293. Roberts. (2006). 425 448. Donaldson. On the dimensional structure of emotional intelligence. V. M. The trait emotional intelligence of ballet dancers and musicians. & Furnham. Zeidner. measuring. British Journal of Psychology. (in press-a)... V. 36. and applying emotional intelligence: What have we learned? What have we missed? In R. 552 569.. K. D. 1091 1100. S. R. Schulze & R. Understanding. & Furnham.. Petrides. Petrides. B. G. (1977).. D. N. Emotional intelligence and life satisfaction.. Petrides. & Bagby. C. Manocha. Petrides. (1989).. V... 10. J. D. K. Programa IPDE 100. R.. UK: Oxford University Press.. Roger. In R. L. Pérez. Trait emotional intelligence: Psychometric investigation with reference to established trait taxonomies. & Stough. 1191 1210. Journal of Applied Social Psychology. Personality and Individual Differences.. R.. M. K. V. Schulze & R. A. International handbook of emotional intelligence. K. D. (2001).... V. Examining the factor structure of the Bar-On Emotional Quotient Inventory with an Australian general population sample. (2001). K. J. E. A psychometric investigation into the construct of emotional intelligence. Módulo CIE-10. T. C. Adaptación y validación española del Trait Emotional Intelligence Questionnaire (TEIQue) en población universitaria. MA: Hogrefe & Huber. R. A. & R.). V. K. M. In G. 278 283. 36. D. (2003). S. 39 57. J. Trait emotional intelligence: Behavioural validation in two studies of emotion recognition and reactivity to mood induction. Radloff. 17. (2005). Encuentros en psicologı́a social.. (in press-b). 15. Petrides.. 33. The CES-D Scale: A self-report depression scale for research in the general population. V. A. Roberts (Eds. K. 1. (2003). 30. & Rubio. Petrides. V. I. Journal of Personality and Social Psychology. Zeidner. A. G. C. A. Petrides. Oxford. J. 385 401. Psicothema. K. Emotional intelligence: Knowns and unknowns (Series in Affective Science). P. S. (1996). 101 107. The role of trait emotional intelligence in a gender- specific model of organizational variables. Roberts (Eds. F. University College London. Roger. R. Personality and Individual Differences. A. (2001). (1998). Cambridge. Examen Internacional de los Trastornos de la Personalidad. Roberts (Eds.. Personality and Individual Differences. Cambridge. & Matthews.

E. D. Jedlicka.. . International classification of diseases and related health problems (10th revision). M.. (1992). World Health Organization. T. J.). M. 18. The effects of leader and follower emotional intelligence on performance and attitude: An exploratory study... D. N. The relationship of emotional intelligence with academic intelligence and the Big Five. M.. H. San Diego] at 08:27 09 June 2015 Emotional intelligence and interpersonal relations. Applied multivariate statistics for the social sciences (4th ed. Watson. (2002). Journal of Social Psychology. Oishi... J. T. Van der Zee. N. 37... E. 37. J. 13. T. M. Development and validation of a self-report measure of emotional intelligence as a multidimensional trait domain.. T. Schakel. & Beck. S. (1998). Austin. L. & Thijs. K. P. N. S. McCarthy.. (2000). Warwick... Is trait emotional intelligence simply or more than just a trait? European Journal of Personality. Canada. G. 37 51. 31. Greeson. Geneva. Journal of Personality Disorders.. Factor structure and validity of a trait emotional intelligence measure. P. Navarro. Paper presented at the meeting of the American Educational Research Association. Facultad de Psicologı́a de la Universidad Complutense de Madrid. Personality and Social Psychology Bulletin.. 6. Smith. K. M. Personality and Individual Differences. (2003). G. T. Switzerland: World Health Organization. E.. Escalas PANAS de Afecto Positivo y Negativo: Validación factorial y convergencia transcultural. K. Personality and Individual Differences. E. (2000). Coston. FURNHAM Saklofske. et al. 37. 141. A. & Anderson. & Caputi. Wong. E. Watson. (2004). 1091 1100. Schutte.. C. Fox. M. 54 PETRIDES. Adaptación española de la segunda edición del inventario de depresión de Beck en población universitaria. (2001).. 287 300. A.. & Wabeke. Categorical versus dimensional classification: Implications from and for research. & Tellegen. M. J. Weissman. (1978... 16.. Schutte. & Carretta. The shifting basis of life satisfaction judgments across cultures: Emotions versus norms.. Characteristic emotional intelligence and emotional well-being. Stevens. C. Van der Zee.. P. 103 125. Journal of Personality and Social Psychology. (2002). L. 74. H. Toronto. & Hollander. T. 769 785. (2001). S. M. Journal of Personality and Social Psychology. Mahwah. (1974). Downloaded by [University of California. 34. A.. M.. D. B. (2002).. K. A. A. (2004). R. C. 102 111. 859 888. R. (2001). On the sins of short-form development. J. C. Emotional intelligence: Not much more than g and personality. Malouff.. S.. Schulte. Snyder. Clark. G. Psychological Assessment. 54. November). Lostao. R. European Journal of Personality. M.. 449 461. Inc. PÉREZ-GONZÁLEZ. Emotional intelligence is? Personality and Individual Differences. G. Spence.. & Triandis. Santed. & Valiente. (2005). J. Tett. Joiner. 707 721. & Minski.. D. 1059 1068. 30. 1063 1070. 526 537. Bobik. Ree. E. Suh. NJ: Lawrence Erlbaum Associates. (1989). 523  536.. C.. D. McKenley. & Nettelbeck. J. Diener. Sanz. 16. Journal of Personality and Social Psychology. K. Chorot. & Wang. (1999). A. S. L. C. 12. 11. 243 263. Widiger. S. The self-monitoring of expressive behavior. 243 274. A. S. Malouff. L. Oades. Mood and temperament. & Vázquez. Trait emotional intelligence and goal self- integration: Important predictors of emotional well-being? Personality and Individual Differences. J. Sandı́n.. T. Development and validation of the Dysfunctional Attitudes Scale: A preliminary investigation. R. Leadership Quarterly. J. (2004). P. 482 493. & Law.. Cognition & Emotion.. Simunek. (2004).. Development and validation of brief measures of positive and negative affect: The PANAS scales. (1992). New York: Guilford Press. M. Psicothema.

. R.. & Barlow. 105. M. Structure of anxiety and the anxiety disorders: A hierarchical model. E. D. Intelligence. (2005). 369 391. I. R. D. (1996). Shani-Zinovich. 181 193. Downloaded by [University of California. Assessing emotional intelligence in gifted and non-gifted high school students: Outcomes depend on the measure. Journal of Abnormal Psychology.. TRAIT EMOTIONAL INTELLIGENCE 55 Zeidner. 33. H. San Diego] at 08:27 09 June 2015 . G. & Roberts.. Zinbarg. Matthews.