Trait Emotional Intelligence: Behavioural Validation in Two Studies of Emotion Recognition and Reactivity to Mood Induction

This paper presents two experiments concerning trait emotional intelligence (‘trait EI’). In study 1, ten high and ten low trait EI individuals were selected from a sample of 85 persons to participate in a computerized experiment involving the recognition of morphed emotional expressions. As hypothesized, high trait EI participants were faster at identifying the expressions than their low trait EI counterparts. In study 2, trait EI scores from 102 persons were residualized on the Big Five and subsequently 15 high and 15 low trait EI individuals were selected to participate in a mood induction experiment. As hypothesized, high trait EI participants exhibited greater sensitivity to the mood induction procedure than their low trait EI counterparts. The findings are discussed in terms of the construct validity of trait EI, with particular emphasis on the issue of incremental validity vis-à-vis broad personality traits. Copyright # 2002 John Wiley & Sons, Ltd. In the last few years, emotional intelligence (EI) has been the object of intense and intensive theoretical examination. The number of empirical studies is also steadily increasing, although the quantity of the emerging evidence has certainly not kept pace with theoretical propositions. As a result, core parts of several salient EI conceptualizations remain untested. The relative dearth of empirical evidence is more pronounced for controlled experiments than for correlational research. The two experimental studies in this paper aim to contribute towards addressing this imbalance in the EI literature. The theoretical antecedents of EI have been presented in many papers (e.g. Ciarrochi, Chan, & Caputi, 2000; Davies, Stankov, & Roberts, 1998; Salovey & Mayer, 1990). Broadly speaking, the construct of EI posits that individuals differ in the extent to which they attend to, process, and utilize affect-laden information of an intrapersonal (e.g. managing one’s own emotions) or interpersonal (e.g. managing others’ emotions) nature. There are several salient conceptualizations, differing primarily in their scope, with some being relatively narrow (see e.g. Mayer & Salovey, 1997) and others relatively broad (see Received 26 March 2002 Copyright # 2002 John Wiley & Sons, Ltd. Accepted 28 August 2002 *Correspondence to: K. V. Petrides, Institute of Education, University of London, 25 Woburn Square, London WC1H 0AA, UK. E-mail: [email protected] e.g. Bar-On, 1997; Goleman, 1995; Petrides & Furnham, 2001). With respect to the theoretical domain of the construct, these conceptualizations overlap, sometimes substantially. Thus, for example, ‘emotional self-awareness’ is a facet common to all salient conceptualizations of EI. Nevertheless, there is a crucial way in which the various approaches to the study of EI differ. As further explained below, this difference concerns the measurement of the construct. An important theoretical shortcoming in the area of EI concerns the neglect of the distinction between typical and maximal performance (Cronbach, 1949; Hofstee, 2001). The latter has cogently argued that the empirically demonstrable independence of personality traits and cognitive abilities (Eysenck, 1994) should be taken into account when developing new theories or hypothesizing new constructs. This distinction between typical and maximal performance has been largely overlooked in EI research, which has led to some degree of theoretical confusion and contradictory results (cf. Petrides & Furnham, 2001). For some time, we have been trying to advance the idea that the type of measurement (self-report versus maximum performance) used to operationalize EI bears directly on the conceptualization of the construct, on the hypotheses that are advanced and tested during the validation process, and, ultimately, on the results and conclusions of this process (Petrides & Furnham, 2000a, 2000b, 2001). The operationalization of EI through maximum-performance tests cannot, and will not, produce the same findings as its operationalization through self-report inventories. Based on the considerations above, we proposed a theoretical distinction between trait EI (or ‘emotional self-efficacy’) and ability EI (or ‘cognitive–emotional ability’; see Petrides & Furnham, 2000a, 2001). The former is measured through self-report questionnaires, whereas the latter is measured through maximum-performance tests, i.e. tests that are based on items that have correct and incorrect answers. It is important to realize that these are two different constructs because the procedures used in their operational definition are fundamentally different, even though their theoretical domains might overlap. Thus, the primary basis for discriminating between trait EI and ability EI is to be found in the type of measurement approach one chooses to employ and not in the theoretical domains of the various EI conceptualizations. Along with other researchers, we have argued that it seems very difficult to devise EI items that are amenable to objective scoring and cover the domain of the construct comprehensively (Roberts, Zeidner, & Matthews, 2001; Zeidner, Matthews, & Roberts, 2001; but also see Mayer, Salovey, Caruso, & Sitarenios, 2001, for the opposing viewpoint). For example, the entire intrapersonal component of EI seems to be impervious to maximum-performance measurement because the information required to score as correct or incorrect responses to items such as ‘I am aware of my emotions as I experience them’ is available only to the individual who is providing those responses. Our research, like that of the majority of researchers in this area (e.g. Ciarrochi, Chan, & Bajgar, 2001; Parker, Taylor, & Bagby, 2001; Reiff, Hatzes, Bramel, & Gibbon, 2001; Saklofske, Austin, & Minski, in press; Schutte et al., 2001; van der Zee, Schakel, & Thijs, 2002), has focused on what we have labelled ‘trait EI’ or ‘emotional self-efficacy’. Trait EI refers to a constellation of emotion-related self-perceptions and dispositions, assessed through self-report. The precise composition of these self-perceptions and dispositions varies across different conceptualizations, with some (see e.g. Bar-On, 1997) being broader than others (see e.g. Schutte et al., 1998). Petrides and Furnham (2001; see Table 1) presented a sampling domain of trait EI derived through a content analysis of several existing models. As noted, the construct has received much attention in the 40 K. V. Petrides and A. Furnham Copyright # 2002 John Wiley & Sons, Ltd. Eur. J. Pers. 17: 39–57 (2003) literature, although most researchers have examined it under the general label of ‘emotional intelligence’. Thus, it has been found that trait EI is related to goal orientation and life satisfaction (Martinez-Pons, 1997), depression and affect intensity (Dawda & Hart, 2000), and marital satisfaction (Schutte et al., 2001). In one of the few experimental studies in the area, Ciarrochi et al. (2001) showed that a specific facet of trait EI (emotion self-regulation) is related to mood management behaviour. In this paper, we focus on global trait EI scores and attempt experimentally to address several issues, including construct validity, incremental validity in a controlled context, and the extent to which the emotion-related self-perceptions that the construct encompasses are related to specific actual abilities.