The dimensional structure of the emotional self‐efficacy scale (ESES)

Abstract

This study aims to investigate the underlying dimensionality of the emotional self‐efficacy scale (ESES) and determine its relationship with measures of ability emotional intelligence (EI) (Mayer–Salovey–Caruso EI Test), trait EI (Trait EI Questionnaire), personality, and cognitive ability. Participants included 822 undergraduate students and 263 graduates already in the workplace. Analyses of the data suggested a multidimensional factor structure for the ESES. The measure was found to correlate with trait EI and showed expected correlations with personality. It did not correlate with ability EI or cognitive ability. These findings are discussed and are interpreted as offering support for the use of the ESES as a reliable measure of emotional self‐efficacy.

Key words:

• emotional intelligence
• emotional self‐efficacy
• self‐efficacy
• validation

Emotional intelligence (EI) has been conceptualised as an emotion‐related cognitive ability involving the ability to perceive, use, understand, and regulate emotion (Mayer & Salovey, 1997; Mayer, Salovey, & Caruso, 2004). Others have defined EI as a constellation of emotion‐related self‐perceptions at the lower levels of personality hierarchies (Petrides, Furnham, & Mavroveli, 2007; Petrides, Pita, & Kokkinaki, 2007). These two perspectives have been termed ability EI and trait EI, respectively. More recently, the notion of emotional self‐efficacy (ESE), as distinct from the trait EI approach, has been discussed (Kirk, Schutte, & Hine, 2008; Qualter, Barlow, & Stylianou, 2011) and is defined as beliefs in one's emotional functioning capabilities. ESE and trait EI differ inasmuch as ESE is solely concerned with self‐perceptions related to emotional functioning, whereas trait EI includes other aspects of self‐perception and dispositions not encompassed by ESE (Kirk et al.). This study is designed to investigate the underlying dimensionality of the emotional self‐efficacy scale (ESES) developed by Kirk et al. (2008) and determine its relationship with an ability EI measure (Mayer–Salovey–Caruso EI Test (MSCEIT)), a trait EI measure (trait EI questionnaire (TEIQue)), and measures of personality and cognitive ability.

ABILITY EI, TRAIT EI, AND ESE

Ability EI is defined as a cognitive ability involving four hierarchical skills: perceiving, facilitating, understanding, and managing emotion (Mayer & Salovey, 1997). Ability EI is distinct from trait EI, which is conceptualised as a constellation of emotion‐related self‐perceptions (e.g., trait‐based emotion perception and emotion regulation) located at the lower levels of personality hierarchies and operationalised via self‐report measure (Petrides et al., 2007). Meta‐analytic research finds a weak correlation of 0.14 between the two types of EI (Van Rooy, Viswesvaran, & Pluta, 2005).

Although ability EI has been the subject of critical debate in the literature for some time (e.g., Brody, 2004; Locke, 2005), particularly in relation to measurement issues (e.g., Fiori & Antonakis, 2011; Maul, 2011), the four‐branch model proposed by Mayer and Salovey (1997) is widely accepted by many researchers as having the potential to explain and predict important outcomes (e.g., Côté, Lopes, Salovey, & Miners, 2010; Mayer, Salovey, & Caruso, 2008).

With regard to measurement, there currently exists only one measure of ability EI that covers all domains described in the Mayer and Salovey (1997) model: the MSCEIT Version 2.0 (MSCEIT V2.0: Mayer, Salovey, & Caruso, 2002). The MSCEIT is scored using consensus‐ or expert‐scoring methods, with the two methods found to correlate highly (Mayer, Salovey, Caruso, & Sitarenios, 2003).

Factor analysis of the MSCEIT has found that a four‐factor model fits data well (Day & Carroll, 2004; Mayer et al., 2003; Roberts, Zeidner, & Matthews, 2001). However, other researchers have found that different solutions are a better fit (Fan, Jackson, Yang, Tang, & Zhang, 2010; Palmer, Gignac, Manocha, & Stough, 2005; Roberts et al., 2006; Rode et al., 2007; Rossen, Kranzler, & Algina, 2008). Despite these inconsistencies, the four branches are most often used in exploratory research, which includes exploration of the relationships between ability EI and theoretically related constructs, such as cognitive ability and personality. Such studies reveal typically only modest relationships with cognitive ability (around 0.25: Joseph & Newman, 2010). These significant correlations occur more frequently or tend to be stronger for the Understanding Emotions branch of the MSCEIT (Bastian, Burns, & Nettelbeck, 2005; Livingston & Day, 2005; Lopes et al., 2004; O'Connor & Little, 2003). Only weak correlations are found between ability EI and personality (significant rs range from 0.11 to 0.33: Day & Carroll, 2004; Lopes, Salovey & Straus, 2003). Data from a recent meta‐analysis (Joseph & Newman) reported the following correlations between total ability EI and personality traits: agreeableness = 0.29; conscientiousness = 0.13; emotional stability = 0.20; extraversion = 0.18; and openness = 0.21.¹

Trait EI is defined as behavioural dispositions and self‐perceptions of one's ability to recognise and understand emotions (Petrides & Furnham, 2000, 2001), which are assessed through self‐report questionnaires (Gardner & Qualter, 2010). Such questionnaires require an individual to reflect on their own perceived ability to recognise and understand emotions in themselves and others, but they also tap into qualities of emotional functioning that cannot be measured via cognitive performance, including self‐motivation and adaptability (Petrides & Furnham, 2000, 2003).

Empirical work typically shows that trait EI is related to personality (Joseph & Newman, 2010; Newsome, Day, & Catano, 2000; O'Connor & Little, 2003; Van der Zee, Thijs, & Schakel, 2002) but not cognitive ability (e.g., Bastian et al., 2005; Fox, Tett, & Palmer, 2003; Newsome et al., 2000; O'Connor & Little; Zeng & Miller, 2003).

THE TRAIT EI, ABILITY EI, AND ESE DISTINCTION

It has been argued that ESE, more recently referred to as trait ESE (Sánchez‐Ruiz, Pérez‐González, & Petrides, 2010), is an appropriate alternative label for trait EI (Petrides & Furnham, 2001; Petrides, Pérez‐González, & Furnham, 2007). However, Kirk et al. (2008) argue that although ESE may be an aspect of trait EI, the two are not identical; other aspects and dispositions are encompassed within the trait EI concept. ESE is concerned with confidence in one's emotional functioning capabilities as operationalised by the four‐branch model of EI. This does not include elements such as self‐perceptions of adaptability or self‐motivation, which are included in trait EI models (e.g., see Sánchez‐Ruiz et al.).

The argument for the ability EI and ESE distinction builds on previous work suggesting an association between beliefs about ability to perform a behaviour and actually performing that behaviour (e.g., Bandura, 1986; Wigfield & Eccles, 1992). Self‐efficacy is even more important than actual task‐related abilities and skills in explaining individual differences in performance (Gundlach, Martinko, & Douglas, 2003), so self‐efficacy in relation to emotional capability is likely to be important; the suggestion being that a person higher in ESE is more likely to use the ability they have (Kirk et al., 2008). Bandura, Caprara, Barbaranelli, Gerbino, and Pastorelli (2003) propose that it is one thing to have the ability to self‐regulate emotion but another to actually use this ability in challenging situations; this is more likely to happen if a person has a strong sense of efficacy. It seems, then, that it is one thing to possess emotional knowledge but another to believe that you have this ability and use it accordingly (Qualter et al., 2011).

The ESES was developed and validated by Kirk et al. (2008). It is based on the four‐branch model of ability EI and contains questions that pertain to self‐efficacy in relation to the ability to perceive, use, understand, and manage emotion. Principal components analysis of the ESES found a one‐component solution with high internal reliability of 0.96 (Cronbach's alpha) for this solution. It was also found to significantly correlate with the overall MSCEIT score (0.34) and with the Understanding and Managing subscales (0.30 and 0.35, respectively). The authors propose the ESES to be a viable measure, which could be useful in future studies aimed at furthering understanding of the processes involved in adaptive emotional functioning.

The overall aim of the current study is to investigate the underlying dimensionality of the ESES. Assessment of the associations among the ESES, MSCEIT, a trait EI measure (TEIQue), personality, and cognitive ability was also of interest.

METHOD

Participants

Eight hundred twenty‐two undergraduate students and 263 participants from the wider university community participated in the study. This gave a total sample of 1,085 participants (M = 403, F = 682). The age range was 18–59-years, and the mean age of the sample was 23-years (SD = 5 years and 10 months).

Measures

ESE

The ESES developed by Kirk et al. (2008) comprises 32 items, with eight items representing each of the four branches of the Mayer et al. (2004) model. For each item, participants rate their confidence at performing this function on a 5‐point scale in which a ‘1’ indicated ‘not at all’ and a ‘5’ indicated ‘very’. Kirk et al. showed that the measure had good internal consistency (α (total scale) = 0.96); 2‐week test–retest reliability was also good (r(26) = 0.85, p < .0001).

Ability EI

The pen‐and‐paper version of the MSCEIT V2.0 (Mayer et al., 2002) was used in this study. It includes 141 items covering all four branches. The tests were scored by the test publisher Multi‐Health Systems using consensus scoring. Internal consistency for the MSCEIT has been reported, with Cronbach's alpha of 0.91 (Mayer et al., 2002); 3‐week test–retest reliability is good at 0.86 (Brackett & Mayer, 2003).

Trait EI

The TEIQue‐short form (Petrides & Furnham, 2006) was used to measure trait EI. This is a self‐report measure of trait EI completed as a pen‐and‐paper task; it has 30 items. Internal consistency has been reported as satisfactory for both males and females (α = 0.84 and 0.89, respectively: Petrides & Furnham).

Cognitive ability

Raven's advanced progressive matrices (Set 1) (Raven, Raven, & Court, 1994) was used as a test of cognitive ability. This is a non‐verbal measure of ability to form perceptual relations and to reason by analogy. It consists of a set of 12 items and is appropriate for use with young adults of above‐average intelligence. All items were presented in black ink on a white background, and participants were asked to identify the missing item that completes a pattern. Participants were given a time limit of 5-min to complete the tasks.

Personality

The International Personality Item Pool (IPIP, Goldberg, 1999) is a 50‐item personality scale that measures the big five factors of personality (N, E, O, A, and C). There are 10 items for each factor, and the 5‐point scale is rated from 1 = very inaccurate to 5 = very accurate. Good internal consistency has been reported (α = 0.77–0.86 for the five factors: Goldberg).

Procedure

Participants from two universities in the North West of England completed the measures in lectures. In addition, recent graduates from one of the universities completed the ESES online in response to an email request.

Analyses plan

Our sample was split into two halves by allocating odd participant numbers to a calibration sample for exploratory factor analysis and even numbers to a validation sample to replicate the model using confirmatory factor analysis. The calibration sample comprised 543 participants (M = 194, F = 349; mean age 23 years), and the validation sample included 542 participants (M = 209, F = 333; mean age 22 years 10 months). Once factor structure was established, we investigated associations with the MSCEIT subscales, TEIQue subscales, IPIP dimensions, and the Raven's advanced matrices score.

RESULTS

The suitability of the data for factor analysis for both the calibration and validation samples was examined. Inspection of the correlation matrix for each sample revealed the presence of many coefficients of 0.30. Also, the Kaiser–Meyer–Oklin values were above 0.90 (0.92 and 0.89, respectively), exceeding the recommended value of 0.60 (Kaiser, 1970, 1974); Bartlett's test of sphericity (Bartlett, 1954) reached statistical significance, supporting the factorability of the correlation matrix.

Principal axis factoring of data in the calibration sample revealed the presence of six components with eigenvalues exceeding >1 (11.36, 2.71, 1.93, 1.34, 1.20, 1.08). The scree plot suggested four components (see Cattell, 1978). Parallel analysis (Lance, Butts, & Michels, 2006; Velicer, Eaton, & Fava, 2000) also showed four components with eigenvalues exceeding the corresponding criterion values for a randomly generated data matrix of the same size (32 × 543 participants).

The four factors were correlated 0.62 or above. Thus, oblique rotation was appropriate (Tabachnick & Fidell, 2007) to calculate common variance. Following this oblique rotation, these four factors accounted for 54.18% of the common variance and were labelled (1) ‘Using and Managing Your Own Emotions’, (2) ‘Identifying and Understanding Your Own Emotions’, (3) ‘Dealing with Emotions in Others’, and (4) ‘Perceiving Emotion through Facial Expressions and Body Language’. Table-1 details the items that loaded onto each factor. Questions 28, 10, 29, 5, and 16 all loaded below 0.45 and were not retained (Tabachnick & Fidell). The internal reliabilities of the four subscales were 0.88 (subscale 1), 0.86 (subscale 2), 0.85 (subscale 3), and 0.80 (subscale 4).

Table 1 Standardised factor loadings for the EFA and CFA models

Factor and items	Factor loading EFA: PME^a	Factor loading CFA
Factor 1. Using and managing your own emotions
8. Change your negative emotion to a positive emotion	0.88	0.76
3. Create a positive emotion when feeling a negative emotion	0.82	0.71
18. Use positive emotions to generate good ideas	0.62	0.76
30. Generate the right emotion so that creative ideas can unfold	0.57	0.76
14. Get into a mood that best suits the occasion	0.55	0.61
12. Regulate your own emotions when under pressure	0.55	0.68
22. Create emotions to enhance cognitive performance	0.54	0.65
20. Calm down when feeling angry	0.53	0.55
6. Use positive emotions to generate novel solutions to old problems	0.49	0.67
26. Create emotions to enhance physical performance	0.48	0.63
Factor 2. Identifying and understanding your own emotions
27. Figure out what causes you to feel differing emotions	0.79	0.82
11. Know what causes you to feel a negative emotion	0.72	0.81
1. Correctly identify your own negative emotions	0.68	0.67
9. Correctly identify your own positive emotions	0.68	0.64
19. Understand what causes your emotions to change	0.61	0.61
4. Know what causes you to feel a positive emotion	0.60	0.65
Factor 3. Dealing with emotions in others
7. Realise what causes another person to feel a positive emotion	0.79	0.74
24. Help another person calm down when he or she is feeling angry	0.63	0.63
31. Figure out what causes another person's differing emotions	0.62	0.69
32. Help another person regulate emotions after he or she has suffered a loss	0.61	0.62
15. Realise what causes another person to feel a negative emotion	0.59	0.65
2. Help another person change a negative emotion to a positive emotion	0.55	0.57
23. Understand what causes another person's emotions to change	0.54	0.61
13. Correctly identify when another person is feeling a positive emotion	0.50	0.58
Factor 4. Perceiving emotion through facial expressions and body language
25. Recognise what emotion you are communicating through your facial expression	0.67	0.79
21. Notice the emotion another person's body language is portraying	0.63	0.68
17. Notice the emotion your body language is portraying	0.57	0.77

Notes. CFA model used maximum likelihood estimation. In the final model, the four factors were allowed to correlate with one another: Factors 1 and 2, r = 0.65; Factors 1–3, r = 0.62; Factors 1 and 4, r = 0.66; Factors 2 and 3, r = 0.69; Factors 2 and 4, r = 0.63; and Factors 3 and 4, r = 0.72. PME = pattern matrix element.

a Questions 28, 10, 29, 5, and 16 all loaded below 0.45 and are not retained in the final four‐factor solution.

Using Amos 18 (Arbuckle, 2009), confirmatory factor analysis was conducted using the validation sample. First, a confirmatory factor analysis using all items was conducted to test for a one‐factor solution, which was found in earlier research by Kirk et al. (2008). Multiple fit indices were consulted to assess model fit: the root mean square error of approximation (RMSEA) and the comparative fit index (CFI). As suggested by Hu and Bentler (1999), cut‐off criteria indicative of good fit are RMSEA < 0.06, and CFI and TLI > 0.95. However, given the exploratory nature of the study, we also deemed it important to consider less conservative criteria indicative of moderate levels of model fit: <0.08 (RMSEA) and >0.90 (CFI) (Marsh, Hau, & Wen, 2004). In addition, two parsimony adjusted fit indices were used to compare models: the parsimonious normed fit index (PNFI) and parsimonious comparative fit index (PCFI). When two or more competing models fit the data equally well, quality can be assessed by examining parsimony (Mulaik et al., 1989). Parsimonious fit indices in the region of 0.50 are not inconceivable even with high goodness‐of‐fit indices (e.g., CFI = 0.90/1), but the larger the value, the more parsimonious the model (Mulaik et al.). Using these rules of thumb, fit indices revealed that a one‐factor model failed to fit the observed data (χ²(464) = 3,830.11 (p = .001), RMSEA = 0.12 ((confidence interval) CI_.95 = 0.112, 0.20), CFI = 0.63, PNFI = 0.53, PCFI = 0.56). This indicated that a unidimensional model was not a good fit to these data.

Second, we tested the four‐factor solution on the validation sample. In this model, the four factors were allowed to correlate with one another. CFA revealed an adequate fit to the data (χ² = 1,793.89, RMSEA = 0.07 (CI_.95 = 0.068, 0.077), CFI = 0.91, PNFI = 0.75, PCFI = 0.78). Factor loadings were reasonably sized, ranging from 0.55 to 0.82 (see Table-1 for exact details). The internal reliabilities of the four subscales used in the CFA were 0.89 (subscale 1), 0.86 (subscale 2), 0.86 (subscale 3), and 0.79 (subscale 4).

Two hundred sixty‐four participants (M = 61, F = 203) completed, in addition to the ESES, the MSCEIT, the TEIQue, IPIP, and Raven's advanced matrixes. Correlations between the respondents' scores on the ESES subscales and these other variables can be found in Table-2. The ESES total showed weak correlations with the ‘Using’ and ‘Managing’ branches of the MSCEIT, and stronger correlations with personality including extraversion, agreeableness, conscientiousness, emotional stability, and openness. It also correlated well with TEIQue (total score) and all TEIQue subscales. As found in previous studies (Joseph & Newman, 2010), cognitive ability (Raven's) was significantly correlated with total ability EI (MSCEIT) and all four branches. There was no significant association between cognitive ability and ESE. The ESES total and all four subscales showed significant positive correlations with age.

Table 2 Correlations among emotional self‐efficacy, ability and trait EI, personality cognitive ability, and age

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22
1. ESES (Total)	(0.93)	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
2. ESES Us/Ma/Own	0.88**	(0.88)	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
3. ESES Id/Un/Own	0.80**	0.58**	(0.86)	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
4. ESES Emo/Others	0.81**	0.53**	0.57**	(0.85)	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
5. ESES Per/Fac/Bod	0.74**	0.53**	0.52**	0.61**	(0.80)	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
6. MSCEIT (Total)	0.09	0.01	0.14*	0.10	0.04	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
7. MSCEIT (Perc)	0.00	−0.02	0.04	0.03	−0.07	0.75**	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
8. MSCEIT (Use)	0.13*	0.07	0.15*	0.13*	0.09	0.78**	0.49**	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
9. MSCEIT (Und)	0.03	0.00	0.04	0.02	0.01	0.58**	0.20**	0.32**	–	–	–	–	–	–	–	–	–	–	–	–	–	–
10. MSCEIT (Man)	0.13*	0.08	0.13*	0.11	0.10	0.68**	0.28**	0.44**	0.38**	–	–	–	–	–	–	–	–	–	–	–	–	–
11. TEIQue (Total)	0.68**	0.67**	0.58**	0.48**	0.38**	0.21**	0.06	0.23**	0.07	0.24**	(0.86)	–	–	–	–	–	–	–	–	–	–	–
12. TEIQue (WB)	0.45**	0.56**	0.40**	0.18**	0.15*	0.10	0.08	0.14*	−0.06	0.13*	0.75**	(0.85)	–	–	–	–	–	–	–	–	–	–
13. TEIQue (SC)	0.56**	0.62**	0.45**	0.32**	0.29**	0.16**	0.08	0.20**	0.11	0.12	0.74**	0.50**	(0.67)	–	–	–	–	–	–	–	–	–
14. TEIQue (Em)	0.44**	0.32**	0.43**	0.43**	0.27**	0.18**	−0.06	0.18**	0.11	0.23**	0.67**	0.29**	0.32**	(0.66)	–	–	–	–	–	–	–	–
15. TEIQue (Soc)	0.44**	0.34**	0.31**	0.43**	0.40**	0.18**	0.07	0.13*	0.11	0.22**	0.59**	0.26**	0.26**	0.26**	(0.70)	–	–	–	–	–	–	–
16. IPIP (Ext)	0.32**	0.29**	0.21**	0.30**	0.26**	0.02	−0.01	0.05	−0.05	0.08	0.46**	0.40**	0.14*	0.25**	0.49**	(0.89)	–	–	–	–	–	–
17. IPIP (Agr)	0.28**	0.19**	0.22**	0.38**	0.09	0.15*	−0.02	0.17**	0.02	0.14*	0.43**	0.27**	0.22**	0.58**	0.08	0.23**	(0.78)	–	–	–	–	–
18. IPIP (Cons)	0.29**	0.20**	0.31**	0.25**	0.22**	0.05	−0.05	0.07	−0.01	0.07	0.43**	0.21**	0.28**	0.30**	0.26**	0.09	0.19**	(0.79)	–	–	–	–
19. IPIP (Neu)	0.44**	0.54**	0.34**	0.21**	0.19**	0.17**	0.12	0.21**	0.05	0.12	0.58**	0.45**	0.74**	0.20**	0.21**	0.19**	0.14*	0.13*	(0.85)	–	–	–
20. IPIP (Open)	0.35**	0.30**	0.23**	0.31**	0.30**	0.20**	0.15*	0.13*	0.21**	0.14*	0.31**	0.05	0.20**	0.21**	0.41**	0.25**	0.09	0.17**	0.09	(0.83)	–	–
21. Raven's	0.01	−0.06	0.04	0.06	0.00	0.28**	0.20**	0.24**	0.19**	0.16*	−0.07	−0.12	0.00	0.01	0.02	−0.09	0.07	−0.13*	0.02	0.15*	–	–
22. Age	0.24**	0.24**	0.17**	0.16**	0.17**	−0.16**	−0.05	−0.06	−0.30**	−0.21**	0.19**	0.10	0.21**	0.12	0.07	0.04	0.18**	0.16*	0.13*	0.07	−0.02	–

Note. Available internal reliability coefficients (alpha) are listed along the diagonal. Correlations among ESES total, ESES subscales, and age run for full sample (N = 1,085) to maximise variability. Perc = perceiving emotions; Use = using emotions; Und = understanding emotions; Man = managing emotions; ESES Us/Ma/Own = using and managing own emotions; ESES Id/Un/Own: identifying and understanding own emotions; ESES Emo/Others = dealing with emotions in others; ESES Per/Fac/Bod = perceiving emotion through facial expressions and body language; WB = well‐being; SC = self‐control; Em = emotionality; Soc = sociability; Ext = extraversion; Agr = agreeableness; Cons = conscientiousness; Neu = neuroticism (items reverse coded); Open = openness; Ravens = Raven's advanced progressive matrices (Set 1).

* p < .05, **p < .01.

DISCUSSION

This study explored the factor structure of the ESES. Exploratory and confirmatory factor analyses suggested that the ESES is multidimensional. Further, the ESES did not correlate with the MSCEIT, but it correlated well with a trait EI measure; it also showed similar patterns of association with personality to the trait EI measure.

In contrast to the original validation study (Kirk et al., 2008) that suggested a unidimensional structure of the ESES, we find the ESES to be multidimensional. Using exploratory and confirmatory factor analyses on two different large samples of participants, we found a consistent four‐factor structure. This structure does not map clearly onto the four‐branch model of EI but instead primarily shows distinctions between confidence in emotional functioning related to oneself and to others. These are important aspects of the ability model of EI; as such, the ESES has face‐validity as a measure of ESE.

In line with the empirical findings of Brackett, Rivers, Shiffman, Lerner, and Salovey (2006) and theoretical models of ESE (Bandura, 1995, 1997; Saarni, 1999), we find clear differences between people's actual emotional skills (as measured by the MSCEIT) and their judgements of these abilities (assessed using the ESES). Previous research with children also shows little association between actual EI skills and beliefs about using these emotional skills in social relationships (Qualter et al., 2011), and we have shown this same association among young adults. However, although the concepts of ability EI and ESE are distinct, both may be important in terms of the behaviour they predict, and future research will want to address this.

Our findings also contribute to the literature on self‐efficacy development. Self‐efficacy beliefs are evidenced to develop from mastery experiences and social modelling (Bandura et al., 2003), so they change over the lifespan. The significant correlations between age and ESE in the current study support this assertion, and higher correlations between ability EI and ESE may be evident among middle or older adult populations.

We find that the ESES correlates well with the TEIQue. This supports the original validation of the measure, where the ESES was found to correlate 0.70 with the ‘Assessing Emotions’ measure of EI (Schutte et al., 1998), another measure of trait EI. Further, we find that the ESES and TEIQue showed similar patterns of association with personality dimensions. Given that the ESES is based exclusively on the four‐branch ability model of EI and does not include measurement of other emotion‐related dispositions, its association with the TEIQue and personality is likely due to shared method variance and semantic overlap between the questionnaire items. Unlike the TEIQue, which measures the broader concept of trait EI, the ESES is directly focused on self‐efficacy in relation to emotional functioning, as defined by the four‐branch model of EI; it does not measure additional individual differences, which may be related to but are not the same as EI. In this sense, the ESES could be seen as a more appropriate measure to use in studies, where the researcher aims to investigate confidence in emotional functioning ability (as operationalised by the original EI ability framework) or is solely interested in the ESE concept.

The results of the current study suggest that ESE can be reliably measured using the ESES. The measure produces four subscale scores that detail how able a person feels at (1) perceiving and managing their own emotions, (2) identifying and understanding their own emotions, (3) managing the emotions of others, and (4) perceiving emotions through facial expressions and body language. The ESES correlates with another trait EI measure (TEIQue) and shows expected associations with personality; it does not correlate well with the MSCEIT ability EI measure or with cognitive ability. We interpret these findings as offering support to the theoretical models of ESE that propose a difference between people's actual emotional skills and their judgements of these abilities. It seems that it is one thing to have EI ability but another to believe that you can use this in everyday encounters. However, both may be important in terms of the behaviour they are able to predict. Future research will want to determine the separate roles of ESE, trait EI, and ability EI in predicting life outcomes.

ACKNOWLEDGEMENT

We wish to thank N. Mohamed, F. Mathieson, and J. Lee for their help with data collection and entry.

Notes

1. Correlation corrected for attenuation and range restriction.