Relationships Among the Sport Competition Anxiety Test, the Sport Anxiety Scale, and the Collegiate Hockey Worry Scale

November 21, 2017 | Author: Meagan Lee | Category: N/A
Share Embed Donate


Short Description

Download Relationships Among the Sport Competition Anxiety Test, the Sport Anxiety Scale, and the Collegiate Hockey Worr...

Description

JOURNAL OF APPLIED SPORT PSYCHOLOGY, 13: 411–429,TRAIT 2001 COMPETITIVE ANXIETY Copyright © 2001 by the Association for Advancement of Applied Sport Psychology 1041-3200/01 $12.00 + .00

411

Relationships Among the Sport Competition Anxiety Test, the Sport Anxiety Scale, and the Collegiate Hockey Worry Scale JOHN G.H. DUNN AND JANICE CAUSGROVE DUNN University of Alberta This study examined the degree to which the Sport Competition Anxiety Test (SCAT; Martens, 1977) and the Sport Anxiety Scale (SAS; Smith, Smoll, & Schutz, 1990) shared variance with the four subscales of the Collegiate Hockey Worry Scale (CHWS; Dunn, 1999)—a sport-specific measure of athletes’ dispositional tendencies to worry about performance failure, negative social evaluation, physical danger, and situational uncertainty. Participants were 178 male intercollegiate ice hockey players. Correlation and regression analyses reinforced the links between worries about failure and negative social evaluation to competitive trait anxiety (CTA). However, neither the SCAT nor the SAS shared more than 5.8% of the variance surrounding athletes’ worries pertaining to physical danger and situational uncertainty. Findings are discussed in the context of Martens, Vealey, and Burtons’ (1990) recommendation to develop instruments with separate subscales measuring different situational components of CTA.

The cognitive anxiety component, “worry,” is commonly recognized as a central construct in the competitive sport anxiety process (Martens, 1977; Martens et al., 1990; Smith, Smoll, & Wiechman, 1998). Worry can be defined as “a chain of thoughts [that are] negatively affect laden [and] relatively uncontrollable” (Borkovec, Robinson, Pruzinsky, & DePree, 1983, p. 10), and is often experienced when an individual’s attention narrows on perceived sources of threat or danger in the environment (Mathews, 1990). In other words, people often worry when they perceive an upcoming event as being “aversive, likely, imminent, and resource demanding” (Schwarzer, 1996, p. 105). Thus, a central feature of worry

Manuscript received 26 June 2000; revision submitted 29 March 2001. The first author would like to thank the support of both the University of Alberta and the Social Sciences Research Council of Canada (SSHRC) for providing funding to support this study through an Izaak Walton Killam Memorial Scholarship and an SSHRC doctoral fellowship, respectively. Address correspondence to John Dunn, Faculty of Physical Education and Recreation, University of Alberta, Edmonton, Alberta, Canada T6G 2H9. E-mail: [email protected]

411

412

DUNN AND CAUSGROVE DUNN

is that it is generally future-oriented, implying that it is focused upon anticipated threat (Borkovec, 1994). Worry is frequently cited as a component of “cognitive interference,” which refers to a class of cognitions that are intrusive, unwanted, undesirable, and at times disturbing to the individual (Sarason, Pierce, & Sarason, 1996). Given the intrusive nature of worry, its impact upon performance has been studied extensively in a variety of settings (e.g., academic testing, sport performance). Empirical findings strongly suggest that worry has a debilitating effect upon performance in relation to cognitive tasks (e.g., Sarason & Sarason, 1987) and physical tasks (including sport settings where worry is often referred to as “cognitive anxiety” or “concern”; e.g., Burton, 1988). Worry is also identified as a defining characteristic of trait anxiety (e.g., Eysenck & Van Berkum, 1992; Schwarzer, 1996). That is, people who are prone to experiencing anxiety have a tendency to worry because they have an attentional disposition to perceive situational threats (e.g., Mathews, 1990). For example, Eysenck and Van Berkum found that adults (M age = 32.3 years) who were classified as high trait anxious had a tendency to worry more frequently about a variety of concerns (i.e., threats) than individuals low in trait anxiety. A similar view on the relationship between worry and trait anxiety has been adopted by prominent sport anxiety theorists (e.g., Martens et al., 1990; Smith et al., 1998) who postulate that the degree to which athletes perceive threat (and subsequently experience worry) in competitive sport settings is largely influenced by competitive trait anxiety (CTA)—where CTA is defined as a personality disposition reflecting an individual’s “tendency to perceive competitive situations as threatening and to respond to these situations with A-state” (Martens et al., 1990, p. 11). Given the characteristic features of worry that have been described, it can be implied that the content of an individual’s worry cognitions reflects, to some degree, the source of threat or danger that is perceived in the environment. Mathews (1990) contends that the domains of worry (i.e., content of worry) most relevant or salient to an individual are generally dictated by the situational circumstances of a person’s environment (or life circumstances). For example, Borkovec et al. (1983) found that students most typically reported worries about performance in academic settings, while Wisocki (1988) found that elderly people worried most frequently about physical health. Sport psychology researchers have identified numerous situational threats emanating from a variety of sources within the competitive sport environment, including outcome uncertainty (Fisher & Zwart, 1982), coaching decisions (Dunn & Nielsen, 1996), and physical danger (Dunn & Nielsen, 1993). However, Smith et al. (1998) have suggested that the most salient sources of situational threat in sport relate to performance failure and the associated consequences of performance failure pertaining to negative social evaluation. This view is also in accordance with worry research in non-sport environments, where Borkovec (1994) has suggested that “the central source of the anxious experience of worriers . . . resides in social evaluation issues” (p. 8). General support for Smith et al.’s (1998) contention that fear of failure and fear of social evaluation are the central threats relating to competitive sport anxiety has

COMPETITIVE TRAIT ANXIETY

413

been obtained by a number of researchers. For example, in a study conducted by Gould, Horn, and Spreemann (1983) with elite junior wrestlers (N = 458), athletes were presented with 33 different sources of stress and asked to rate how frequently each source of stress caused them to experience nervousness or worry before competition. Results showed that four of the five most highly ranked sources of stress that contributed to pre-competitive nerves or worry were directly linked to the possibility of performance failure. In a more recent qualitative study, James and Collins (1997) interviewed 20 male and female adult athletes who competed in a variety of sports at a variety of competitive levels. Participants were asked to describe sources of stress that they generally encountered in their respective sports. Results showed that approximately two thirds (67.3%) of all sources of stress identified by the athletes were directly linked to socially-evaluative and self-presentational concerns. In another recent study, Wilson and Eklund (1998) examined the relationship between CTA and self-presentational concerns among a sample of male and female intercollegiate athletes (N = 199). Wilson and Eklund observed that among their sample, “the tendency to experience competitive anxiety, particularly cognitive anxiety in the form of worries, [was] closely associated with the tendency to perceive self-presentational threat during sport competition” (p. 93). Having said this, it is important to recognize that both Wilson and Eklund (1998) and James and Collins (1997) cautioned that not all anxiety experiences in sport are founded upon self-presentational concerns. Not surprisingly, concerns about performance failure and negative social evaluation have been identified as central components of CTA. Research has consistently found that high CTA athletes (from a variety of sports) tend to worry more frequently about factors relating to poor performance and negative social evaluation than their low CTA counterparts (e.g., Gould et al., 1983; Passer, 1983). However, despite strong empirical evidence linking fear of failure and fear of negative social evaluation to CTA, Martens et al. (1990) advocated that competitive anxiety research would benefit from the creation of CTA inventories that contain separate subscales measuring different “components of competitive anxiety (e.g., fear of evaluation, fear of failure, etc.)” (p. 115). This position is supported by numerous sport psychologists who have argued that there is value in conceptualizing fear of failure and fear of negative social evaluation as separate constructs in sport anxiety research (e.g., Dunn, 1999; Hackfort & Schwenkmezger, 1993). Moreover, Hackfort (1986; Hackfort & Schwenkmezger, 1993) has even suggested that specific CTA components might include failure anxiety, disgrace/shame anxiety (i.e., fear of negative social evaluation), physical harm/injury anxiety, and anxiety of the unknown (i.e., fear of the unknown). The two most prominent CTA measures currently in use by sport psychology researchers—the Sport Competition Anxiety Test (SCAT; Martens, 1977) and the Sport Anxiety Scale (SAS; Smith et al., 1990)—do not differentiate between different dimensions of threat (or domains of worry) that may reside within the competitive sport environment. For example, the SCAT contains eight items that describe somatic symptoms of anxiety and two items that describe cognitive elements of anxiety. However, it should be noted that the two cognitive items focus

414

DUNN AND CAUSGROVE DUNN

solely on athletes’ concerns about performance failure: “Before I compete I worry about not performing well” and “When I compete I worry about making mistakes.” The somatic and cognitive items are summed to provide a unidimensional composite measure of CTA (Martens, 1977). Despite being used less frequently over the last decade since the introduction of the SAS as a CTA measure, the SCAT is still recognized as a valuable research tool as it continues to be used in current sport anxiety research (e.g., Wilson & Eklund, 1998). In contrast to the unidimensional assessment of CTA provided by the SCAT, the SAS is comprised of three subscales measuring somatic anxiety (nine items), worry (seven items), and concentration disruption (five items). Examination of the item-content in the worry subscale of the SAS suggests that six of the seven items describe self-doubts and concerns about performance failure, whereas the seventh item (“I’m concerned that others will be disappointed in my performance”) describes concern about negative social evaluation. Thus, when responses on the seven SAS-worry subscale items are summed to provide a composite subscale score, the possibility that there may be a conceptual difference between the athlete’s concerns about performance failure and negative social evaluation (see Dunn, 1999; Gould et al., 1983; James & Collins, 1997; Martens et al., 1990) is apparently overlooked. Given that the item-content and subscale-structures of the SCAT and SAS do not differentiate between various domains of competitive worry, the instruments provide little information about the potential sources of threat that may be causing the anxious response. In other words, although the SCAT and SAS provide sport psychologists with valuable information about the presence of CTA, they provide limited information about athletes’ susceptibility to different situational threats that may be contributing to their CTA levels (see Kroll, 1980, for a similar discussion relating to the SCAT). The purpose of this study was to examine the relationships between the SCAT and SAS and athletes’ dispositional tendencies to experience anxiety that originates from a variety of situational threats. More specifically, the study examines the degree to which intercollegiate ice hockey players’ dispositional tendencies for worrying about performance failure, negative social evaluation, physical danger, and situational uncertainty are associated with their SCAT and SAS scores. The four worry domains closely parallel the CTA dimensions proposed by Hackfort (1986) and were empirically derived by Dunn (1999) in a study of the latent dimensionality of competitive worry in ice hockey. On the basis of previous research findings (e.g., Passer, 1983) and the narrow scope of the worry domain content contained within the items of the SCAT and SAS, it was hypothesized that the SCAT and the worry subscale of the SAS would be strongly related to athletes’ dispositional tendencies for worrying about performance failure, and moderately related to players’ tendencies for worrying about negative social evaluation. Hypotheses concerning the strength of the relationships between the two CTA measures and athletes’ worry dispositions regarding situational uncertainties and physical harm were not offered because the SCAT and SAS contain no items describing these dimensions. However, because CTA

COMPETITIVE TRAIT ANXIETY

415

theory predicts that high CTA individuals are more likely to perceive a wider variety of competitive situations as threatening than low CTA individuals (Martens et al., 1990), positive relationships between SCAT, SAS, and worries relating to situational uncertainties and physical harm were anticipated.

METHOD Participants Participants were 178 athletes (excluding goaltenders) from 10 western Canadian university and college hockey teams (M age = 22.27 years, SD = 1.62).1 Written informed consent was obtained from all athletes, each of whom voluntarily participated in the study. The athletes had an average of 2.06 years playing experience at the intercollegiate level (SD = 1.12). Instruments Athletes completed four self-report instruments. The test-battery consisted of a demographic questionnaire, two CTA inventories—the SCAT and the SAS—and a sport-specific measure of pre-competitive worry tendencies in ice hockey (the Collegiate Hockey Worry Scale [CHWS]; Dunn, 1999). Demographic questionnaire. The demographic questionnaire contained nine items. Questions focused primarily upon the athletes’ age, playing experience, and playing responsibilities (e.g., playing position). Sport Competition Anxiety Test. The SCAT contains 15 items, 10 of which measure symptoms associated with anxiety. The five items that are not scored are included in the inventory to reduce the likelihood of an internal response-set bias. The standard instructions of the SCAT ask respondents to indicate how they “usually feel when competing in sports and games.” However, to make the instrument more contextually relevant to the athletes in this study, the phrase “sports and games” was replaced with the word “hockey.” On a 3-point scale (1 = hardly ever, 2 = sometimes, 3 = often) respondents indicated the frequency with which they generally experienced the 10 anxietyrelated symptoms. The scores for the 10 items are summed to provide an overall measure of CTA, with a high composite score (as opposed to a low composite score) reflecting a greater tendency to experience competitive anxiety. Martens et al. (1990) provided a detailed overview of the validation process pertaining to the SCAT’s development including evidence of high internal consistency (KR-20 1. There is overlap between the present sample and the sample used in a previous publication by Dunn et al. (2000), which examined the factorial structure of the SAS. The current sample is a subsample (constituting only 35%) of the Dunn et al. sample. Dunn et al. combined data from three independent research projects including: (1) the present sample of 178 varsity hockey players plus 24 goaltenders who are not included in this study, (2) 174 high-performance male Canadian Football players, and (3) 128 male NCAA intercollegiate athletes.

416

DUNN AND CAUSGROVE DUNN

values ranging from .95 to .97), high test-retest reliability (M retest reliability = .77), and a list of over 80 published studies that have employed the instrument as a measure of CTA. Sport Anxiety Scale. The SAS is a multidimensional measure of CTA that is comprised of three subscales which measure somatic CTA (nine items), worry CTA (seven items), and concentration disruption CTA (five items). On a 4-point Likert-type intensity rating scale (1 = not at all; 4 = very much so) respondents indicated the degree to which they generally experienced the 21 anxiety-related symptoms before or during competition. Although strong psychometric evidence supporting the internal consistency (a ranged from .74 to .92), test-retest reliability (r ranged from .68 to .71), and factorial validity of the three SAS subscales was reported by Smith et al. (1990), a recent study by Dunn, Causgrove Dunn, Wilson, and Syrotuik (2000) identified problems with two items. Specifically, results of an exploratory factor analysis conducted upon a mixed sample (n = 255) of male intercollegiate level athletes from a variety of sports and male high school age Canadian football players revealed that two items originally designed by Smith et al. to measure concentration disruption actually loaded on the cognitive worry factor. Goodness-of-fit indices obtained from a follow-up confirmatory factor analysis on a similar independent cross-validation sample of athletes (n = 249) also indicated that a better fitting solution was obtained when the two items originally designed to measure concentration disruption loaded on the worry factor (Dunn et al.). Collegiate Hockey Worry Scale. The CHWS (Dunn, 1999) assesses hockey players’ predispositions towards experiencing pre-game worries relating to injury/physical danger, negative social evaluation, personal performance failure, and uncertain/unknown competitive conditions. Each subscale contains four items with the exception of the fear of failure subscale which contains three items. The introductory instructions of the CHWS inform respondents that the purpose of the inventory is to identify the aspects of competition that ice hockey players tend to worry about before competing. A 5-point scale (1 = not at all like me; 5 = very much like me) is used by respondents to rate the degree to which each item (i.e., worry statement) characterizes the way the athlete generally feels or thinks before competition. The CHWS data included in the present study are the same data obtained from the 15-item version of the scale reported by Dunn (1999). Procedures At team meetings scheduled by coaches, all four instruments were administered to athletes in a single session. All instruments were completed mid-season, at least 48 hours prior to each teams’ next competitive game. The demographic questionnaire was always completed first. However, the presentation order of the three remaining inventories (i.e., SCAT, SAS, and CHWS) was counterbalanced across respondents to minimize the likelihood of any presentation order effect biasing responses. Athletes completed the SCAT, SAS, and CHWS in one of six possible presentation orders.

COMPETITIVE TRAIT ANXIETY

417

RESULTS Psychometric Analyses of CTA Instruments Sport Competition Anxiety Test. The unidimensional structure of the SCAT has recently been questioned by Smith et al. (1998) who cited two unpublished factor analytic studies that failed to verify the unidimensional structure of the instrument. Moreover, Karteroliotis (1997) found that a two-factor model (with an eightitem somatic factor and a two-item cognitive factor) produced slightly better goodness-of-fit indices than those obtained for a one-factor model with confirmatory factor analyses of SCAT data provided by Greek physical education students. Given these findings, maximum likelihood confirmatory factor analyses (CFA) using LISREL 8 (Jöreskog & Sörbom, 1996) were conducted on both the original onefactor model proposed by Martens (1977) and the alternative two-factor model suggested by Karteroliotis (1997) and Smith et al. (1998). Model fit was assessed using Jöreskog and Sörbom’s (1981) goodness-of-fit index (GFI), Steiger’s (1990) root mean square error of approximation (RMSEA), Bentler and Bonnett’s (1980) non-normed fit index (NNFI), and Bentler’s (1990) comparative fit index (CFI). All goodness-of-fit indices associated with the two-factor model (GFI = .90; RMSEA = .099; NNFI = .88; CFI = .91) were better than those indices associated with the one-factor model (GFI = .88; RMSEA = .120; NNFI = .82; CFI = .86). Moreover, the c² difference test for nested models (Tabachnick & Fidell, 1996, p. 753) revealed that the alternative two-factor model provided a significant improvement in model-fit compared to the one-factor model, c² difference = 31.47, df = 1, p < .001. Eight items loaded on the somatic anxiety factor; factor loadings ranged in size from .41 to .52. The two remaining items (3 and 5) had factor loadings of .52 and .39 (respectively) on the cognitive anxiety factor. All 10 factor loadings were statistically significant at p < .001. The inter-factor correlation was .46. The internal consistency (a) of each factor was .87 (somatic) and .64 (cognitive). Coefficient alpha for the composite 10-item scale was .85. Given that most sport anxiety researchers have used the 10-item SCAT as a unidimensional measure of CTA, in accordance with Martens’ (1977) original conceptualization of CTA, a composite SCAT score based on the summation of all 10 anxiety-related items is used to reflect athletes’ levels of CTA throughout the paper. However, following the clear identification of a better fitting model with the two-factor solution, correlation results relating to an eight-item SCATsomatic factor are also provided. With only two items loading on the SCAT-cognitive factor, psychometricians would generally consider it inappropriate to create a separate two-item subscale (see Smith et al., 1998, p. 114, for a related discussion pertaining to the two-item SCAT-cognitive factor). Nevertheless, for descriptive purposes, correlations between the two-item SCAT-cognitive subscale and all other subscales contained within the three inventories used in this investigation are provided (see Table 1). Sport Anxiety Scale. Maximum likelihood CFAs were conducted on the SAS

418

a

SCAT Cognitiveb

SAS Somatic

SAS Worry

SAS Con. Dis.

a

= .85

a

= .87

.97***

a

= .64

.34***

.55***

a

= .88

.29***

.78***

.77***

a

= .86

.34***

.72***

.26**

.40***

a

n.s.

= .72

.25**

n.s.

n.s.

-.13

.13

-.17*

-.12

a

n.s.

n.s.

= .89

.20**

.20**

.16

.15*

.14

.16*

M (SD) 8.30 (3.62)

CHWS Injury

CHWS Failure

CHWS Evaluation

a

= .80

.60***

.08

n.s.

.24**

.22**

.18*

.18*

.21**

a

a

= .78

.66***

= .76

n.s.

.17*

.14 .17*

.16*

n.s.

.11

.08

n.s.

.60***

.22**

.52***

.18*

.28***

.67***

.29***

.60***

.30***

.42***

M (SD) M (SD) M (SD) 8.43 (3.28) 10.90 (2.71) 13.29 (3.71)

CHWS Unknown

Note. Correlations are displayed in upper triangular matrix. Scale internal consistency coefficients ( a ) are displayed in main diagonal. a SCAT-somatic reflects the composite score obtained by summing the eight SCAT items that loaded on the somatic anxiety factor following the CFA. b SCAT-cognitive reflects the composite score obtained by summing the two SCAT items that loaded on the cognitive anxiety factor following the CFA. *p < .05. **p < .01. ***p < .001. n.s. Non-significant.

CHWS Evaluation

CHWS Failure

CHWS Unknown

CHWS Injury

SAS Con. Dis.

SAS Worry

SAS Somatic

SCAT-Cognitive

SCAT-Somatic

SCAT-10

SCAT Somatica

M (SD) M (SD) M (SD) M (SD) M (SD) M (SD) 21.16 (4.58) 16.60 (4.07) 4.56 (1.13) 18.68 (5.54) 20.72 (5.38) 5.62 (1.95)

SCAT

Table 1 Means, Standard Deviations, Internal Consistencies (Coefficient Alpha) and Correlations (r) Between the Sport Competition Anxiety Test (SCAT), Sport Anxiety Scale (SAS) Subscales, and Collegiate Hockey Worry Scale (CHWS) Subscales

418 DUNN AND CAUSGROVE DUNN

COMPETITIVE TRAIT ANXIETY

419

data to test the comparative fit of the original model proposed by Smith et al. (1990) and the alternative model proposed by Dunn et al. (2000).2 In the alternative model, two items originally designed to measure concentration disruption load on the worry factor. All goodness-of-fit indices associated with the alternative model (GFI = .84; RMSEA = .073; NNFI = .86; CFI = .88) were superior to those indices associated with the original model (GFI = .80; RMSEA = .092; NNFI = .79; CFI = .82). All factor loadings for the alternative model were statistically significant (ps < .001). Given these results, the factor structure of the SAS reported by Dunn et al. is adopted in this study with nine items comprising the somatic subscale, nine items comprising the worry subscale, and three items comprising the concentration disruption subscale. Internal consistency (a) values for the three subscales were .88 (somatic), .86 (worry), and .72 (concentration disruption). Collegiate Hockey Worry Scale. The present CHWS data were previously factor analyzed by Dunn (1999), with the results suggesting the retention of a threefactor solution (with fear of failure and fear of negative social evaluation items loading on a single factor). However, results of multidimensional scaling analyses in that same study revealed four distinct item-clusters depicting four worry dimensions pertaining to fear of failure, fear of negative social evaluation, fear of injury/ physical danger, and fear of the unknown. Furthermore, because the inventory was established within a theoretical framework that posited four a priori worry dimensions (see Dunn, 1999), the CHWS is treated as a four dimensional instrument in the present study. Internal consistency (a) estimates for the present data set across all four subscales were acceptable: fear of failure = .76, fear of negative social evaluation = .78, fear of injury/physical danger = .89, fear of the unknown = .80. Presentation Order Effects The number of athletes who completed each of the six presentation orders ranged from 26 (14.6%) to 32 (18%). To determine if inventory scores differed systematically between the six presentation-order groups, a MANOVA was conducted on the scores from the eight inventory subscales (i.e., a composite 10-item SCAT score, three SAS subscale scores [a nine-item SAS-somatic subscale, a nine-item SAS-cognitive subscale, and a three-item SAS-concentration disruption subscale], and four CHWS subscale scores) with the six presentation-order formats entered as independent variables. Prior to the analysis, 22 athletes were randomly deleted from the data set so that each of the six presentation-order cells contained equal numbers of participants (n = 26). A non-significant multivariate test statistic was obtained, Wilks L = .726 , F (40, 626.12) = 1.19, p = .20, suggesting the absence of any presentation order effects. 2. Again it is acknowledged that the present participants comprise a subset of the sample used by Dunn et al. (2000). However, the data for the 178 athletes in this study have not been previously factor analyzed as a single data set.

420

DUNN AND CAUSGROVE DUNN

Inter-Scale Relationships Table 1 contains the descriptive statistics for the two CTA measures and the four subscales of the CHWS, and the Pearson product-moment correlations between each of the inventory subscales. The 10-item SCAT had a strong significant positive correlation (r = .77) with the SAS-somatic subscale, and a moderate significant positive correlation (r = .40) with the nine-item SAS-worry subscale. A follow-up statistical analysis to test differences between dependent correlations (Glass & Hopkins, 1984, p. 311) revealed that the rSCAT.SAS-Somatic was significantly greater than rSCAT.SAS-Worry , t (175) = 6.862, p < .001. These findings reinforce previous contentions (e.g., Smith et al., 1998) that the SCAT primarily measures the somatic component of CTA. The correlation between the three-item SAS-concentration disruption subscale and the 10-item SCAT was not significant (p > .05). Although all four CHWS subscales had significant positive correlations with the 10-item SCAT, only one of the four CHWS subscales (CHWS-failure) had a correlation with the SCAT (r = .42) greater than the criterion level of .30 that is indicative of a moderate effect size (Cohen, 1992). Follow-up statistical tests to examine differences in correlations between the SCAT and the four CHWS subscales revealed that rSCAT.CHWS-failure (r = .42) was significantly greater than the correlations between the SCAT and the three remaining CHWS subscales (all ts [175] > 2.40, all ps < .05). These results suggest that the SCAT is more closely related to CTA responses associated with threats surrounding performance failure than threats relating to the three other dimensions of situational worry examined in this study. With the exception of the CHWS-injury subscale, all of the CHWS subscales had small significant positive correlations with the SAS-somatic subscale. None of the CHWS subscales shared more than 8.5% of the variance with SAS-somatic scores, indicating that somatic CTA (as measured by the somatic subscale of the SAS) is not closely related to any of the four CHWS dimensions. Both the CHWS-evaluation and CHWS-failure subscales had fairly high positive correlations with the SAS-worry subscale (r = .60 and .67, respectively), whereas the CHWS-unknown and CHWS-injury/physical danger subscales had relatively low (but statistically significant) positive correlations with the SASworry subscale (r = .20 and .24, respectively). The difference in size between the correlations of the CHWS-failure and CHWS-evaluation with the SAS-worry subscale was not statistically significant. However, both the CHWS-failure and CHWS-evaluation subscales had significantly larger correlations with the SASworry subscale than did the CHWS-injury and CHWS-unknown subscales (all ts [175] > 4.69, all ps < .001). These results suggest that the items contained within the worry subscale of the SAS primarily assess the social evaluation and performance failure aspects of competitive worry. Only one of the four CHWS subscales (fear of injury/physical danger) had a statistically significant relationship with the SAS-concentration disruption subscale (r = .20). The correlation between SAS-worry and the SAS-concentration disruption subscale was also small and significant (r = .25). These results suggest that the

COMPETITIVE TRAIT ANXIETY

421

tendency to lose concentration during competition (as measured by the three-item SAS-concentration disruption subscale) is not strongly associated with competitive worry dispositions among the present sample of male varsity hockey players. Regression of SCAT and SAS Scores Simultaneous multiple regression analyses were conducted to further examine the relationships between the SCAT and SAS and the four CHWS subscales. The CTA measures were regressed on the CHWS subscales in four separate analyses in order to (a) determine the size of the overall relationship between the CTA measures and the CHWS subscales, and (b) examine the unique contribution of each CWHS subscale to the CTA scores. The results of these analyses are summarized in Table 2. Table 2 reveals that the set of CWHS subscales explained 19.4% of the variance in SCAT scores, but only the fear of failure subscale was a significant predictor of SCAT scores (p < .001). Despite having significant (albeit moderate to small) correlations with the SCAT (see Table 1), none of the other CHWS subscales made significant unique contributions to the total variance of the SCAT. Examination of the pattern of intercorrelations among the SCAT and CHWS subscales (Table 1), as well as the semipartial correlations (Table 2) provides insight into this finding. For example, the zero-order correlation reveals that the SCAT was significantly correlated with CHWS-evaluation (r = .28), but the semipartial correlation between SCAT and CHWS-evaluation (sr = .001) indicates that this subscale contributed little unique variance explained in the SCAT over and above that which is already contributed by other CHWS subscales. Further, the correlation between the evaluation and failure subscales of the CHWS (r = .66) suggests that the fear of evaluation subscale is accounting for much of the same variance in the SCAT as the fear of failure subscale. The CHWS subscales explained 11.6% of the variance in SAS-somatic scores. The fear of failure and fear of the unknown subscales were significant predictors of SAS-somatic scores. As was the case with the SCAT, fear of evaluation was not a significant predictor despite having a significant zero-order correlation with the somatic subscale of the SAS (r = .22). The semipartial correlation between SASsomatic and CHWS-evaluation (sr = .016) indicates that the fear of evaluation subscale contributes little unique variance to the total variance in SAS-somatic scores. Again, it appears that much of the contribution of the fear of evaluation subscale to the variance explained in SAS-somatic scores is shared with the fear of failure subscale. The CHWS subscales explained almost 51% of the variance in SAS-worry scores, with both the fear of failure and fear of evaluation subscales making significant unique contributions (Table 2). Although the CHWS-injury and -unknown subscales were significantly correlated with SAS-worry (r = .20 and .24, respectively), the semipartial correlations indicate that neither subscale contributed a significant amount of unique variance to the explanation of SAS-worry scores. The correlation between the injury and unknown subscales (r = .60) suggests that

422

DUNN AND CAUSGROVE DUNN

Table 2 Simultaneous Multiple Regression Analyses of Collegiate Hockey Worry Scale (CHWS) Subscales on Sport Competition Anxiety Test and Sport Anxiety Scale (SAS) Subscales CHWS Subscale

Beta

p

Semipartial Correlation

4.311 .018 .254 1.492

< .001 ns ns ns

.294 .001 .017 .102

2.640 .224 –.180 2.033

< .01 ns ns < .05

.189 .016 –.013 .145

6.621 3.788 .396 1.411

< .001 < .001 ns ns

.354 .203 .021 .075

ns ns < .01 ns

–.008 .075 .191 –.063

t

Sport Competition Anxiety Test, F (4, 173) = 10.481, p < .001, R² = .194 Fear Fear Fear Fear

of Failure of Evaluation of Injury of the Unknown

.392 .002 .022 .128

SAS-somatic, F (4, 173) = 5.667, p < .001, R² = .116 Fear Fear Fear Fear

of Failure of Evaluation of Injury of the Unknown

.251 .021 –.016 .182

SAS-worry, F (4, 173) = 44.116, p < .001, R² = .505 Fear Fear Fear Fear

of Failure of Evaluation of Injury of the Unknown

.472 .270 .026 .095

SAS-concentration disruption, F (4, 173) = 2.436, p < .05, R² = .053 Fear Fear Fear Fear

of Failure of Evaluation of Injury of the Unknown

–.010 .100 .239 –.079

–.106 1.020 2.584 –.851

a large proportion of their contributions to the explained variance in SAS-worry scores may be shared. Finally, the set of CWHS subscales explained only 5.3% of the variance in SAS-concentration disruption scores. Only the fear of injury subscale was a significant predictor of concentration disruption CTA (Table 2). Individual Level Assessment of Worry Tendencies Among High CTA Athletes Although the inter-subscale correlations reported in Table 1 suggest that high CTA athletes (as measured by the SAS and SCAT) are generally more inclined than their low CTA counterparts to worry about all four of the situational threat dimensions measured by the CHWS, it would be inappropriate to conclude that high CTA athletes (or low CTA athletes, for that matter) are characterized by similar dispositional patterns towards worrying about all four CHWS dimensions. To illustrate this point with high CTA athletes, individual worry profiles for all

COMPETITIVE TRAIT ANXIETY

423

athletes (n = 13) scoring > – 24 (out of 36) on both the somatic and cognitive subscales of the SAS are reported in Table 3. Selecting the score of 24 as the criterion value on both subscales was done partly out of convenience, but also on the grounds that this score was more than one standard deviation above the mean on both subscales. Consequently, athletes who achieved scores >– 24 were among the highest 20% of scorers on both CTA dimensions in the present sample. As is clearly seen in Table 3, irrespective of age, playing experience, and playing position, contrasting worry profiles are still evidenced among a number of the athletes who had similar CTA levels on the somatic and cognitive subscales of the SAS. For example, athletes 1 and 10 are both 21-year-old first-year defensemen with identical levels of cognitive worry CTA (as measured by the SAS-worry subscale). Despite these similarities, the two athletes show markedly different predispositions towards worrying about social evaluation, performance failure, physical danger, and situational uncertainties in the sport of hockey. Although it may be argued that the difference in dispositional worry patterns between athletes 1 and 10 could have been caused by differences in their SAS-somatic scores (see Table 3), different dispositional worry profiles can also be seen among athletes with similar CTA levels on all three measures of CTA reported (i.e., SAS-somatic, SAS-cognitive, and the 10-item SCAT). For example, athletes 4 and 7 were similarly aged defensemen with very similar CTA profiles on the two SAS subscales and the SCAT, yet they had markedly different scores on the CHWS-injury and CHWS-unknown subscales.

DISCUSSION The present study examined the relationships between the two most prominent CTA measures and competitive worry dispositions of ice hockey players (as measured by the CHWS). Of the 45 zero-order inter-scale correlations computed in this study (see Table 1), 80% (n = 36) were statistically significant and in a positive direction. This pattern of correlations is encouraging from a convergent validity perspective given that the SCAT, SAS, and CHWS employ different methods (albeit similar self-report paper-and-pencil-test protocols) to purportedly measure aspects of the same construct (i.e., CTA). Moreover, because the three instruments differ in terms of the number of response scale options, descriptors associated with the response scale options, and time-frames considered when responding to the items (i.e., before competition, or before and during competition), the interscale correlations may have been attenuated as a function of these methodological differences. With the exception of the SAS-concentration disruption subscale, the direction and statistical significance of the inter-scale correlations contained in Table 1 suggests that the SCAT, SAS, and CHWS are all measuring related components of the CTA construct. The direction, size, and significance levels of the correlations between the three-item SAS-concentration disruption subscale and the other subscales examined in this study suggests that further research surrounding the conceptualization of concentration disruption as a component of CTA (as measured by the three-item SAS subscale) is required.

424

21.42

19.83

20.00

21.08

22.75

20.67

20.67

21.42

22.17

21.50

21.67

21.83

19.50

1

2

3

4

5

6

7

8

9

10

11

12

13

1

1

1

1

1

3

1

1

1

2

1

1

1

Playing Year

Centre

Winger

Winger

Defense

Winger

Winger

Defense

Defense

Centre

Defense

Defense

Centre

Defense

Playing Position

26

34

28

33

25

27

28

28

24

27

27

24

24

SAS-Somatic a

25

31

25

24

24

28

28

34

29

28

29

31

24

SAS-Cognitive a

27

30

28

30

25

26

26

24

25

27

27

23

22

SCAT b

Competitive Trait Anxiety Subscale Scores

Note. Range of possible composite scores for each subscale: a 9 - 36; b 10 - 30; c 4 - 20; d 3 - 15.

Age (Years)

ID

Demographic Characteristics

14

16

15

19

16

15

16

19

16

15

16

17

14

Fear of Evaluationc

14

13

14

15

12

14

12

15

12

13

13

15

10

Fear of Failured

8

10

5

12

13

5

19

20

8

10

8

7

9

Fear of Injuryc

10

14

7

5

12

8

11

20

10

6

6

11

11

Fear of the Unknown c

Collegiate Hockey Worry Scale Subscale Scores

Table 3 Demographic Playing Characteristics and Anxiety Profiles for all Athletes Scoring –> 24 on Both Somatic and Worry Subscales of the Sport Anxiety Scale (SAS)

424 DUNN AND CAUSGROVE DUNN

COMPETITIVE TRAIT ANXIETY

425

In accordance with theory (e.g., Martens et al., 1990), the correlational findings (see Table 1) indicate that high CTA hockey players are generally characterized as being more prone to experiencing a wider variety of pre-competitive worries (relating to different situational threats) than low CTA players. However, of the 16 correlations computed between the CHWS subscales and the CTA measures (i.e., the 10-item SCAT and the three SAS subscales), only three were > .30 (rSAS-worry.CHWS= .67, r SAS-worry.CHWS-evaluation = .60, rSCAT.CHWS-failure = .42). A value of r = .30 represents a failure medium effect size that describes the functional significance of a product-moment correlation coefficient (Cohen, 1992). None of the correlations between the injury and uncertainty subscales of the CHWS and the CTA measures were > .24, indicating that the SCAT and SAS subscales do not share more than 6% of the variance in worry tendencies stemming from concerns relating to physical danger and situational uncertainties. Thus, the degree to which the SCAT and SAS measure anxiety relating to situational threats beyond those directly relating to performance failure and social evaluation appears limited. It should be acknowledged, however, that the strength of the correlations between the CTA measures (i.e., SCAT and SAS) and the CHWS-injury and CHWSuncertainty subscales may also have been attenuated by the nature of the sport and the competitive level of the sport from which the present sample was selected. For example, Dunn (1999) argued that most hockey players competing at the intercollegiate level in Canada are provided with excellent preparatory information from their coaches prior to competition. This would presumably reduce the degree to which situational uncertainties existed in the performance environment, thereby causing high CTA athletes (as measured by the SCAT and SAS) to report few worries on this dimension. Given these conditions, the strength of the positive correlation between the CTA measures and the CHWS-uncertainty subscale would be greatly reduced. Dunn also speculated that given the physical nature of intercollegiate level hockey in Canada, players competing at this level would not likely choose to play (or, for that matter, be selected to play) had their concerns about getting hurt or injured been excessive. If this is the case, then high CTA athletes (as measured by the SCAT and SAS) would again report few worries on this worry domain, which in turn would reduce the strength of the correlations between the CTA instruments and the CHWS-injury subscale. Research has shown that athletes recognize potentially injurious situations as salient sources of threat (Dunn & Nielsen, 1993) and competitive anxiety (Dunn & Nielsen, 1996). However, few researchers (e.g., Hackfort, 1986; Hackfort & Schwenkmezger, 1993) have conceptualized fear of injury/physical danger as a relevant dimension of CTA. Although the magnitude of the correlations between the CHWS-injury subscale and the SCAT and SAS are small (see Table 1), the consistent positive direction of these relationships lends support to Hackfort’s (1986) contention that fear of injury/physical danger is a dimension of CTA. These findings reinforce the need to develop instruments that directly measure fear of injury as part of the CTA construct (cf. Martens et al., 1990). The fear of the unknown dimension of competitive worry was based in part upon the ambiguous A-trait construct described by Endler, Edwards, and Vitelli

426

DUNN AND CAUSGROVE DUNN

(1991), which measures individual differences in anxiety proneness to “new or unfamiliar situations...[in which a person] is uncertain as to what to expect” (p. 8). Although fear of the unknown has rarely been conceptualized as an independent component of CTA, situational- or outcome-uncertainty has long been recognized as a key factor contributing to competitive anxiety (e.g., Fisher & Zwart, 1982; Martens et al., 1990). Correlations between the CHWS-unknown subscale and SAS-somatic (r = .22), SAS-worry (r = .24), and SCAT (r = .21) indicate that players with higher CTA levels are, on average, more inclined to worry about situational uncertainties than their low CTA counterparts. Furthermore, this variable was a significant predictor of somatic CTA (see Table 2), although it should be recognized that it only contributed 2.1% of the unique variance in SAS-somatic scores. Therefore, although the results lend support to the theoretical conceptualization of CTA proposed by Hackfort (1986) and his colleagues (Hackfort & Schwenkmezger, 1993), if a greater understanding of CTA stemming from fear of the unknown is to be gained, the construction of instruments that directly measure this dimension of CTA is required. The results of the regression analyses in Table 2 indicate that the four CHWS subscales are more closely related to the worry subscale of the SAS (R² = .505) than to the SCAT (R² = .194), SAS-somatic (R² = .116) and SAS-concentration disruption (R² = .053). This finding is not surprising given that (a) both the SASworry subscale and the CHWS are measures of cognitive anxiety, and (b) there is some degree of overlap in the content of the items in the two inventories. For example, two of the three CHWS-failure items (“I worry about playing poorly” and “I worry about not performing up to the best of my ability”) closely parallel two SAS-worry items (“I’m concerned about performing poorly” and “I’m concerned that I may not do as well in competition as I could”) which likely explains the fairly large amount of shared variance (r² = .449) between these two subscales.3 The observed relationship between the SAS-worry subscale and the CHWSevaluation subscale was also quite strong (r = .60). This correlation may seem high given that only one item of the SAS-worry subscale (“I’m concerned that others will be disappointed in my performance”) makes specific reference to a social evaluation concern. However, examination of the regression results and the semipartial correlation between SAS-worry and the CHWS-evaluation susbscale (sr = .203) reveals that fear of evaluation does contribute a significant amount of unique variance to SAS-worry scores (sr² = .041) over and above the unique contribution made by fear of failure (sr² = .125). Taken in conjunction with the results described in the preceding paragraph, these findings appear to support pre3. Although not part of the main analysis, the two-item SCAT-cognitive subscale was also regressed on the CHWS subscales, and a statistically significant regression equation was obtained, F (4, 173) = 27.75, p < .001, R² = .391. The fact that the four CHWS subscales accounted for almost 40% of the variance in the SCAT-cognitive scores provides further evidence suggesting that the CHWS primarily taps the cognitive aspect of CTA given that the four CHWS subscales accounted for less than 20% of the variance in SCAT scores when the 10-item SCAT was regressed against the CHWS (R² = .194).

COMPETITIVE TRAIT ANXIETY

427

vious self-presentational anxiety research (e.g., James & Collins, 1997) suggesting that “an athlete’s fear of failure . . . [may] be conceptualized as part of a larger psychological phenomenon associated with a desire to maintain a favorable level of social comparison” (Dunn, 1999, p. 274). Although the SCAT and SAS both provide valuable information about CTA levels, they provide researchers and applied sport psychologists with limited information about the causes of anxiety because neither instrument gives consideration to the existence of different situational threats that might reside within the competitive environment being studied (cf. Fisher & Zwart, 1982). The information contained in Table 3 clearly illustrates that athletes with similar levels of CTA (as measured by the SAS-somatic and SAS-worry subscales) were not similarly predisposed to worrying about all four classes of situational threat measured by the CHWS. Only when the SCAT and SAS subscale scores were examined in conjunction with the CHWS subscale scores was a greater understanding of why athletes experience anxiety attained. The additional information provided by the CHWS would be highly beneficial to a practitioner attempting to develop a more complete psychological profile of an athlete’s anxiety experiences. Without the benefit of the individual athlete-profile analysis that was conducted on the CHWS subscales (see Table 3), the sport psychologist has almost no information regarding the different situational threats that might be contributing to the level of CTA reported. An excellent illustration of the previous point is seen in the comparison of athletes 4 and 7 (Table 3) that was described in the results section. Both athletes are defensemen with almost identical scores on the somatic and worry subscales of the SAS. However, although they also have quite similar worry dispositions in regards to performance failure and negative social evaluation, they have markedly different dispositional tendencies towards worrying about the potential for physical harm and situational uncertainty. From an applied perspective, if a sport psychologist was designing psychological skills programs aimed at helping these two athletes cope with their anxiety, the content (or direction) of certain aspects of the interventions would have to be targeted at quite different sources of worry. In conclusion, the construction of a CTA instrument that differentiates between anxiety stemming from different situational threat dimensions (or CTA components; Martens et al., 1990) would almost certainly help sport psychology professionals gain a better understanding of why athletes may experience anxiety in the competitive sport setting (see Kroll, 1980). Based on the findings of the present study, it is suggested that established CTA instruments (i.e., the SCAT and SAS)— which currently provide global or macro-level assessments of CTA—be expanded in scope in terms of item-content to include additional domains of worry that may provide more detailed information on why athletes experience anxiety in sport.

REFERENCES Bentler, P. M. (1990). Comparative fit indices in structural models. Psychological Bulletin, 107, 238–246.

428

DUNN AND CAUSGROVE DUNN

Bentler, P. M., & Bonnett, D. G. (1980). Significance tests and goodness-of-fit in the analysis of covariance structures. Psychological Bulletin, 88, 588–606. Borkovec, T. D. (1994). The nature and origins of worry. In G. C. L. Davey & F. Tallis (Eds.), Worrying: Perspectives on theory, assessment, and treatment (pp. 5–33). Chichester, England: Wiley. Borkovec, T. D., Robinson, E., Pruzinsky, T., & De Pree, J. A. (1983). Preliminary exploration of worry: Some characteristics and processes. Behaviour Research and Therapy, 21, 9–16. Burton, D. (1988). Do anxious swimmers swim slower? Reexamining the elusive anxietyperformance relationship. Journal of Sport & Exercise Psychology, 10, 45–61. Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159. Dunn, J. G. H. (1999). A theoretical framework for structuring the content of competitive worry in ice hockey. Journal of Sport & Exercise Psychology, 21, 259–279. Dunn, J. G. H., Causgrove Dunn, J., Wilson, P., & Syrotuik, D. G. (2000). Reexamining the factorial composition and factor structure of the Sport Anxiety Scale. Journal of Sport & Exercise Psychology, 22, 183–193. Dunn, J. G. H., & Nielsen, A. B. (1993). A between-sport comparison of situational threat perceptions in ice hockey and soccer. Journal of Sport & Exercise Psychology, 15, 449–465. Dunn, J. G. H., & Nielsen, A. B. (1996). A classificatory system of anxiety inducing situations in four team sports. Journal of Sport Behavior, 19, 111–131. Endler, N. S., Edwards, J. M., & Vitelli, R. (1991). Endler Multidimensional Anxiety Scales (EMAS): Manual. Los Angeles: Western Psychological Services. Eysenck, M. W., & Van Berkum, J. (1992). Trait anxiety, defensiveness, and the structure of worry. Personality and Individual Differences, 13, 1285–1290. Fisher, A. C., & Zwart, E. F. (1982). Psychological analysis of athletes’ anxiety responses. Journal of Sport Psychology, 4, 139–158. Glass, G. V., & Hopkins, K. D. (1984). Statistical methods in education and psychology (2nd ed.). Englewood Cliffs, NJ: Prentice-Hall. Gould, D., Horn, T., & Spreemann, J. (1983). Sources of stress in junior elite wrestlers. Journal of Sport Psychology, 5, 159–171. Hackfort, D. (1986). Theoretical conception and assessment of sport-related anxiety. In C. D. Spielberger & R. Diaz-Guerrero (Eds.), Cross-cultural anxiety (Vol. 3; pp. 79–91). New York: Hemisphere. Hackfort, D., & Schwenkmezger, P. (1993). Anxiety. In R. N. Singer, M. Murphey, & L. K. Tennant (Eds.), Handbook of research on sport psychology (pp. 328–364). New York: Macmillan. James, B., & Collins, D. (1997). Self-presentational sources of competitive stress during performance. Journal of Sport & Exercise Psychology, 19, 17–35. Jöreskog, K., & Sörbom, D. (1981). LISREL V: Analysis of linear structural relationships by the method of maximum likelihood. Chicago: National Education Resources. Jöreskog, K., & Sörbom, D. (1996). LISREL 8: User’s reference guide. Chicago: Scientific Software International. Karteroliotis, C. (1997). The factorial validity of SCAT in Greek students: Testing the assumption of equivalent structure and mean differences across gender. In R. Lidor & M. Bar-Eli (Eds.), Innovations in sport psychology: Linking theory and practice (Part 1: pp. 345–347). Wingate Institute, Israel: International Society of Sport Psychology. Kroll, W. (1980). The stress of high performance athletics. In P. Klavora & J. V. Daniel (Eds.), Coach, athlete, and the sport psychologist (2nd ed.; pp. 211–219). Champaign,

COMPETITIVE TRAIT ANXIETY

429

IL: Human Kinetics. Martens, R. (1977). Sport Competition Anxiety Test. Champaign, IL: Human Kinetics. Martens, R., Vealey, R. S., & Burton, D. (1990). Competitive anxiety in sport. Champaign, IL: Human Kinetics. Mathews, A. (1990). Why worry? The cognitive function of anxiety. Behaviour Research and Therapy, 28, 455–468. Passer, M. W. (1983). Fear of failure, fear of evaluation, perceived competence, and selfesteem in competitive-trait-anxious children. Journal of Sport Psychology, 5, 172– 188. Sarason, I. G., Pierce, G. R., & Sarason, B. R. (1996). Domains of cognitive interference. In I. G. Sarason, G. R. Pierce, & B. R. Sarason (Eds.), Cognitive interference: Theories, methods, and findings (pp. 139–152). Mahwah, NJ: Erlbaum. Sarason, I. G., & Sarason, B. R. (1987). Cognitive interference as a component of anxiety: Measurement of its state and trait aspects. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol 5; pp. 3–14). Berwyn, PA: Swets North America. Schwarzer, R. (1996). Thought control of action: Interfering self-doubts. In I. G. Sarason, G. R. Pierce, & B. R. Sarason (Eds.), Cognitive interference: Theories, methods, and findings (pp. 99–115). Mahwah, NJ: Erlbaum. Smith, R. E., Smoll, F. L., & Schutz, R. W. (1990). Measurement and correlates of sportspecific cognitive and somatic trait anxiety: The Sport Anxiety Scale. Anxiety Research, 2, 263–280. Smith, R. E., Smoll, F. L., & Wiechman, S. A. (1998). Measurement of trait anxiety in sport. In J. Duda (Ed.), Advances in sport and exercise psychology measurement (pp.105–127). Morgantown, WV: Fitness Information Technology. Steiger, J. H. (1990). Structural model evaluation and modification: An interval estimation approach. Multivariate Behavioral Research, 25, 173–180. Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics (3rd ed.). New York: HarperCollins. Wilson, P., & Eklund, R. C. (1998). The relationship between competitive anxiety and selfpresentational concerns. Journal of Sport & Exercise Psychology, 20, 81–97. Wisocki, P. A. (1988). Worry as a phenomenon relevant to the elderly. Behavior Therapy, 19, 369–379.

View more...

Comments

Copyright � 2017 SILO Inc.