پرسشنامه تنظیم احساسات برای استفاده در ورزشکاران

Objectives Three studies examine the factorial validity, internal consistency, test–retest stability, and criterion validity of the Emotion Regulation Questionnaire (ERQ: Gross & John, 2003) for use with athletes. Design Factorial validity, internal consistency, test–retest stability and criterion validity of the ERQ were examined over three stages, using three separate samples. Method In stage 1 the factorial validity and internal consistency of the ERQ subscales were examined based on responses from 433 sport participants. In stage 2, 176 sport participants completed the ERQ on two occasions separated by an interval of two weeks. In stage 3, the criterion validity of the ERQ was examined. Sport participants (n = 88) completed the ERQ and reported the intensity, frequency and direction of a range of emotions experienced when competing in sport. Results Confirmatory factor analysis results lend some support to a two-factor model when reappraisal and suppression are allowed to correlate. Alpha coefficients were acceptable. Test–retest stability analyses indicated poor agreement and a greater influence of situational, as opposed to trait factors, in the variance of item scores on the second test administration. In addition, results were partially consistent with findings of Gross and John (2003): reappraisal scores were associated with pleasant emotions, but suppression scores were not associated with unpleasant emotions. Conclusion Results provide mixed support for the validity of the ERQ in sport. Because the ERQ is intended to assess stable patterns of emotion regulation, the instability of items is a concern and reasons for this require further investigation.

Whether it is anxiety about returning after an injury, embarrassment about making a mistake, anger at a contentious decision by an umpire, or excitement at the prospect of winning, athletes1 experience a range of emotions prior to, during, and after competitive sport (Hanin, 2000; Lane & Terry, 2000; Uphill & Jones, 2007). Emotions are not only an intrinsic part of competitive sport, they are widely believed to influence sport performance. Consequently, the ability to regulate emotions successfully is regarded by many as an important psychological skill (Gould & Maynard, 2009; Jones, 2003; Robazza, Pellizzari, & Hanin, 2004). Although most humans are able to recognise, and report experiencing, a variety of “emotions”, because emotion is a term derived from everyday language, identifying the necessary and sufficient conditions for something to qualify as an emotion has been plagued with difficulty (Gross & Thompson, 2007). Increasingly researchers have adopted prototype definitions (e.g., Russell & Fehr, 1994), with the essence of emotion being described by Frederickson (2001) as a cognitively appraised reaction to an event, either conscious or non-conscious that “triggers a cascade of response tendencies across loosely coupled component systems, such as subjective experience, facial expression, cognitive processing, and physiological changes” (p.218). In addition, some researchers include a behavioural component (e.g., action tendencies) in the emotional response (see Gross, 1998; Russell, 2003). These features of emotion have a range of intra-personal and interpersonal consequences (see Jones, 2003; Uphill, McCarthy, & Jones, 2009; Vallerand & Blanchard, 2000). For example, heightened muscle tension could lead to a decrement in fine motor control (Noteboom, Fleshner, & Enoka, 2001), changes in what information is being attended to might influence decision-making (Schwarz, 2000), and how emotions are displayed and communicated may influence social functioning (Gross & John, 2003). Emotion regulation can be defined as the evocation of thoughts or behaviours that influence the emotions that people experience, when people have them, and how people experience or express these emotions (Richards & Gross, 2000; Uphill et al., 2009). Emotion regulation broadly encompasses attempts to evoke, diminish, prolong or intensify emotional experience, cognition, expression and/or physiology (cf. Gross, 1999; Gross & Thompson, 2007). The vast majority of literature examining emotion regulation has been conducted in domains of psychology other than sport (Gross & Thompson, 2007). Viewing sport as just one of many arenas in which individuals may wish to regulate their emotions, we contend that (a) such empirical work has utility in this domain, and (b) because sport represents a rich and dynamic laboratory for the assessment of emotion regulation, examining emotion regulation could lead to advances applicable to other domains of psychological inquiry. The ways in which emotions can be regulated are almost limitless with over 400 strategies being identified in the literature broadly (Augustine & Hemenover, 2009; Richards & Gross, 2000). Based on the premise that emotions unfold over time, Gross (1998) proposed a model of emotion regulation in which certain strategies have their primary impact at different stages of the emotion-generation process (cf. Balzarotti, John, & Gross, 2010). The first four of these categories (situation selection, situation modification, attention deployment, cognitive change) have been classified as “antecedent-focussed” emotion regulation, which includes attempts to generate and/or inhibit the arousal of emotional responses. In contrast, the fifth category (response modulation) involves attempts to change the characteristics of emotions that have been triggered, effectively “mopping-up” or altering the “emotional punch” (Gross, 1998, 2008; Richards & Gross, 2000). Within these two broad classes of emotion regulation more subtle differences may be made: antecedent-focussed strategies may involve reappraisal (type of cognitive change) whilst response-focussed strategies may include suppression (Gross, 1998).2 Antecedent-focussed emotion regulation is purported to have different consequences than response-focussed emotion regulation (Gross, 1998). In the regulation of unpleasant emotions for example, antecedent-focussed regulation has been demonstrated to decrease the intensity of subjective, physiological and expressive symptoms. For suppression there is a tendency for the behavioural expression of emotion to decrease as opposed to the subjective or physiological components of emotion (cf. John & Gross, 2007). In addition, emotional suppression has been demonstrated to exact a comparatively heavy cognitive cost (Richards & Gross, 2000). Specifically, the self-regulatory processes involved in inhibiting the expression of emotion depleted attentional resources available for memory, whereas reappraisal exacted little cognitive cost (Richards & Gross, 2000; for a review see Gross, 2008). Collectively, at least in domains other than sport, evidence suggests that reappraisal has a more adaptive profile compared to suppression. To measure individual differences in use of reappraisal and suppression, Gross and John (2003) developed the Emotion Regulation Questionnaire. Evidence suggests that the ERQ possesses adequate factorial validity, internal consistency, test–retest reliability and criterion validity (Gross & John, 2003), and has been validated in different languages (e.g., Balzarotti et al., 2010). Scores on the ERQ relate to several coping measures and results generally show that reappraisal is associated with positive adaptations including pleasant emotions whereas suppression is associated with negative adaptations and unpleasant emotions (John & Gross, 2007). Given divergent outcomes associated with individual differences in the disposition to use reappraisal and suppression, a measure that could identify athletes' use of both these emotion regulation strategies could be beneficial to sport psychologists both theoretically and practically. Theoretically, it is plausible that emotion regulation strategies may in part mediate the variability in the emotion–performance relationship observed between individuals (cf. Hanin, 2000; Jones & Uphill, 2011). Practically, if reappraisal is a preferred emotion regulation strategy, by using the ERQ to identify athletes' habitual use of reappraisal and suppression, interventions directed towards enhancing suppressors' increased habitual use of reappraisal could be monitored and evaluated. The utility of an instrument depends fundamentally upon its psychometric properties and it would be erroneous to assume that instruments are valid across different populations and contexts (Hagger & Chatzisarantis, 2009). For example, items associated with behavioural emotion regulation demonstrated to be valid in one context (Niven, Totterdell, Stride, & Holman, 2011) were often associated with weak factor loadings in sport (Lane, Beedie, Stanley, & Devonport, 2011). If validity has not been demonstrated, it is hazardous to accept and apply data derived from such measures (Cronbach & Meehl, 1955; Schutz & Gessaroli, 1993). Indeed, although there is widespread use of the ERQ, evidence in support of the stability of the scale is lacking. In the original validity paper, Gross and John (2003) reported a test–retest reliability of .69 over a 3 month period and concluded that individuals' reports about their habitual use of reappraisal and suppression are moderately stable over time. Gross and John found less than 50% common variance between test–retest measures, which could lead to interpreting the stability of the scale differently. The stability of the ERQ is an important feature for sport psychologists. On the one hand, considerable evidence suggests that emotion regulation use can be changed through guided training (Robazza et al., 2004), self-help interventions (Achtziger, Gollwitzer, & Sheeran, 2008), and via self-regulation (Augustine & Hemenover, 2009). If training leads to changes in habitual use of emotion regulation strategies, then it is conceivable that such change should be reflected on changes in scores on the ERQ. On the other hand, in the absence of any training and deliberate attempts to self-regulate, test–retest responses to the ERQ should show stability. There are several approaches to demonstrating stability, each yielding possibly different conclusions. Stability here refers to the idea that item scores retain a degree of resistance to change over time. Reliability is defined as the ratio of true variance to error variance (Cohen, 1960), is typically assessed using correlation, and is not necessarily a good indicator of stability. For example, a perfect correlation (r = 1.00) can be found with no agreement, when measures are unstable. Consider the following example to illustrate this point. Scores taken from three participants at one point in time of 1, 2, and 3 will correlate perfectly with scores recorded at a second point in time of 3, 4, and 5. Although correlational methods are routinely used in sport and exercise psychology, they have been subject to some criticism (e.g., Lane, Nevill, Bowes, & Fox, 2005; Nevill, Lane, Kilgour, Bowes, & Whyte, 2001; Wilson & Batterham, 1999). Both the Pearson correlation and the intraclass correlation are an indication of relationship and do not offer information regarding agreement ( Bland & Altman, 1999; Lane et al., 2005; Nevill et al., 2001) and so are unable to detect test–retest differences from one time to another. Recent research has therefore examined differences between test–retest reliability and test–retest stability (Lane et al., 2005; Nevill et al., 2001). An aspect of stability is the extent to which test–retest scores are reproducible, regardless of environment conditions; that is, if person reports a score of 1”not at all” in one week, he would report the same score at the next administration regardless of the time lag or situation. Nevill et al. (2001) outlined a within individual item-by-item approach to test–retest designs, specifically calculating the proportion of agreement. The proportion of agreement is “based on the proportion of participants that record the same response on two separate occasions” (Nevill et al., 2001, p.273). This method entails the calculation of test–retest differences and then the reporting of the percentage of individuals whose differences are found to be within a reference value of plus or minus 1. The authors suggested that 90% of test–retest differences should be within a range of plus or minus 1. Nevill et al. argued that the nature of self-report means that people might be genuinely unsure of whether to report a score of 1 (not at all) or 2 (a little) and as such, provide a score of 1 on the first completion and 2 on the second. They argued that there should not be test–retest differences greater than 1. They acknowledged that a test–retest agreement of 90% ± 1 is arbitrary, but pointed that all criterion values for acceptance in statistics, including a value of less than .05, are based on judgement. An alternative approach to consideration of the stability of item scores is provided based on Latent State-Trait (LST) theory (e.g., Steyer, Schmitt, & Eid, 1999; Ziegler, Ehrlenspiel, & Brand, 2009).3 Briefly stated, LST theory assumes individuals are always measured in situations and the “true” score (latent state score) observed on a measurement instrument is decomposed into variance explained by the trait (latent trait), variance explained by the occasion (latent state residual), and measurement error. These relations, based on the model of Egloff, Schmukle, Burns, Kohlmann, and Hock (2003) can be examined using structural equation modelling (see Fig. 1 for a simple schematic representation). Importantly insofar as the assessment of stability is concerned, coefficients of specificity, consistency and reliability for individual items can be derived (see Table 3 and Ziegler et al., 2009, for elaboration and calculations). The consistency (CON) of the score equals the amount of total observed variance explained by the latent trait. Occasion specificity (SPE) represents the amount of total observed variance due to the latent state residual and, hence, due to the situation and the situation by person interaction. Finally, the reliability (REL) of the test score is the amount of observed total variance explained by latent trait and latent state residual variance. Collectively, these indices of stability afford arguably a robust assessment of the stability of the ERQ in a sport population. Full-size image (94 K) Fig. 1. Illustration of a multi-trait, multi-state model. Note: For reasons of clarity, all error variables are not included. Figure options Table 1. Factor loadings and error variances for the 10-item ERQ. Subscale Item Factor loading Error variance Reappraisal When I want to feel more positive emotion (such as joy or amusement) I change what I'm thinking about .663 .748 When I want to feel less negative emotion (such as sadness or anger) I change what I'm thinking about .647 .762 When I'm faced with a stressful situation, I make myself think about it in a way that helps me stay calm .495 .869 When I want to feel more positive emotion, I change the way I'm thinking about the situation .775 .631 I control my emotions by changing the way I think about the situation I'm in .671 .741 When I want to feel less negative emotion, I change the way I'm thinking about the situation .728 .686 Suppression I keep my emotions to myself .630 .748 When I am feeling positive emotions I am careful not to express them .461 .888 I control my emotions by not expressing them .828 .561 When I am feeling negative emotions, I make sure not to express them .567 .823 Table options Table 2. Fit indices for a correlated and uncorrelated model of the ERQ. Fit index Sample (n = 433) Uncorrelated Correlated Sattora–Bentler χ2 85.23* 96.61* RCFI .95 94 GFI .95 94 RMSEA .07 06 AIC 156.67 148.81 *Significant (p < .01). Table options Table 3. Descriptive statistics, stability indices, and LST coefficients for the ERQ items. Min Max Test 1 Test 2 r Intraclass % (±1) % ≥ 1 % (0 diff) % ≤ −1 LST coefficients M SD M SD CON (T1, T2) SPE (T1, T2) REL (T1, T2) 1. R 1 7 4.7 1.4 5.0 1.4 .29** .29** 71 28 33 39 .5, .2 .2, .7 .7, .9 2. S 1 7 42 1.7 4.2 1.7 .38** .38** 60 39 26 36 .7, .5 .1, .4 .8, .9 3. R 1 7 4.4 1.4 4.8 1.3 .14 .13* 64 23* 39 37 .6, .3 .2, .7 .8, .9 4. S 1 7 3.4 1.5 3.2 1.6 .22** .22** 68 40 29 31 .7, .5 .1, .4 .8, .9 5. R 1 7 4.5 1.6 4.7 1.5 .42** .42** 67 28 32 39 .5, .2 .2, .6 .7, .8 6. S 1 7 3.8 1.6 3.5 1.6 .23** .22** 56 41 28 30 .7, .5 .1, .4 .8, .9 7. R 1 7 4.6 1.3 4.7 1.3 .32** .32** 75 10 35 34 .6, .2 .2, .7 .8, .9 8. R 1 7 4.3 1.4 4.5 1.2 .30** .30** 73 32 32 34 .6, .2 .2, .7 .8, .9 9. S 1 7 4.0 1.7 4.1 1.6 .37** .37** 65 32 31 36 .7, .5 .1, .4 .8, .9 10. R 1 7 4.3 1.5 4.5 1.4 .22** .21** 64 32 32 36 .6, .2 .2, .7 .8, .9 R = reappraisal item, S = suppression item, *p < .05 – illustrates a systematic positive shift (for Median sign test), **p < .01, T1, T2 = time 1, time 2. Table options In summary, the proposed utility of the theoretical framework and associated measures developed by Gross and John (2003), alongside the importance attached to understanding athletes' attempts at emotion regulation suggest that there is a need to investigate emotion regulation among athletes. Given the cautionary suggestions made by Hagger and Chatzisarantis (2009), we suggest that there is need to cross-validate the ERQ for use with a sports population. This study investigates the validity of the ERQ in sport in three ways. First, we tested the factorial validity of the scale. Specifically, it is important to address the question of whether the measurement model for the ERQ that was supported among individuals generally would also be supported among athletes (cf. Hagger & Chatzisarantis, 2009). Second, we examined the stability of the instrument using correlational measures that are typically used in research, those based on item agreement (see Nevill et al., 2001), and by decomposing the variance in scores based on LST theory. Third, we assessed the criterion validity of the ERQ by examining the extent to which scores on the ERQ related to the intensity, direction and frequency of experienced emotions using the Sport Emotion Questionnaire (SEQ: Jones, Lane, Bray, Uphill, & Caitlin, 2005). Although the SEQ only measures the intensity of athletes' pre-competitive emotions, following similar modifications (Jones, Swain, & Hardy, 1993; Swain & Jones, 1993), a directional and frequency scale was appended to the individual items of the SEQ to assess participants' ratings of how much their emotions either associated with good or poor performance, and how often they experienced the emotion respectively.