2007_pse_relationship sas and cs

24
Psychology of Sport and Exercise 8 (2007) 1–24 Relationships between sources of acute stress and athletes’ coping style in competitive sport as a function of gender Mark H. Anshel a, , Toto Sutarso b a Department of Health and Human Performance, Middle Tennessee State University, Murfreesboro, TN 37132 USA b Information Technology Division, Murfreesboro, TN 37132 USA Available online 3 March 2006 Abstract Objectives: The objectives of this study were to determine athletes’ sources of acute stress (SAS) perceived as highly intense and experienced during the competitive event, their respective coping styles (CS) for two different (highly intense) stress sources (SAS), the relationship between the acute stressors and their CS (approach and avoidance coping in cognitive and behavioral forms), and the generalizability of the SAS and the CS scales as a function of gender. Method: Athletes (N ¼ 332, 176 males and 156 females, M age ¼ 21.6 years) who were former or current sports competitors for their high school or college team completed a two-part inventory generated for this study. The athletes were asked to indicate their perceived stress intensity for common SAS’s (part 1) and the manner in which they typically coped with two of the stressors perceived as the most intense (part 2). Theory-driven categories of acute stress sources were labelled ‘‘performance-related’’ and ‘‘coach-related,’’ and CS’s were grouped as ‘‘approach-behavioral,’’ ‘‘approach-cognitive,’’ and ‘‘avoidance-cognitive.’’ Intra- reliability (Cronbach alphas) for the stressor and coping style items were .81 and .82, respectively. Results: General CS was significantly related to general sources of acute stress (po:0001). Structural equation models indicated that the athletes’ coping styles were positively related to their respective acute stressors category. The coping stress style three-factor model showed a good fit with the data. The results of the analyses indicated valid and reliable relationships between CS and SAS among the athletes. The results indicated that athletes who experienced intense coach-related acute stress was more likely to use primarily an approach-behavior CS followed by the other CS. Finally, the athlete’s gender was a mediating variable in determining CS in response to selected sources of stress. ARTICLE IN PRESS www.elsevier.com/locate/psychsport 1469-0292/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.psychsport.2006.01.003 Corresponding author. Fax: +1 615 898 5020. E-mail address: [email protected] (M.H. Anshel).

Upload: tosutea

Post on 30-Oct-2014

10 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

Psychology of Sport and Exercise 8 (2007) 1–24

1469-0292/$ -

doi:10.1016/j.

�CorresponE-mail add

www.elsevier.com/locate/psychsport

Relationships between sources of acute stress and athletes’coping style in competitive sport as a function of gender

Mark H. Anshela,�, Toto Sutarsob

aDepartment of Health and Human Performance, Middle Tennessee State University, Murfreesboro, TN 37132 USAbInformation Technology Division, Murfreesboro, TN 37132 USA

Available online 3 March 2006

Abstract

Objectives: The objectives of this study were to determine athletes’ sources of acute stress (SAS) perceivedas highly intense and experienced during the competitive event, their respective coping styles (CS) for twodifferent (highly intense) stress sources (SAS), the relationship between the acute stressors and their CS(approach and avoidance coping in cognitive and behavioral forms), and the generalizability of the SASand the CS scales as a function of gender.Method: Athletes (N ¼ 332, 176 males and 156 females, M age ¼ 21.6 years) who were former or currentsports competitors for their high school or college team completed a two-part inventory generated for thisstudy. The athletes were asked to indicate their perceived stress intensity for common SAS’s (part 1) and themanner in which they typically coped with two of the stressors perceived as the most intense (part 2).Theory-driven categories of acute stress sources were labelled ‘‘performance-related’’ and ‘‘coach-related,’’and CS’s were grouped as ‘‘approach-behavioral,’’ ‘‘approach-cognitive,’’ and ‘‘avoidance-cognitive.’’ Intra-reliability (Cronbach alphas) for the stressor and coping style items were .81 and .82, respectively.Results: General CS was significantly related to general sources of acute stress (po:0001). Structuralequation models indicated that the athletes’ coping styles were positively related to their respective acutestressors category. The coping stress style three-factor model showed a good fit with the data. The results ofthe analyses indicated valid and reliable relationships between CS and SAS among the athletes. The resultsindicated that athletes who experienced intense coach-related acute stress was more likely to use primarilyan approach-behavior CS followed by the other CS. Finally, the athlete’s gender was a mediating variablein determining CS in response to selected sources of stress.

see front matter r 2006 Elsevier Ltd. All rights reserved.

psychsport.2006.01.003

ding author. Fax: +1615 898 5020.

ress: [email protected] (M.H. Anshel).

Page 2: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–242

Conclusions: Structural equation model techniques showed that athletes who experienced acute stress usedtheir respective CS consistently. The CS three-factor model showed a good fit with the data. In addition,gender mediates the relationship between source of stress and subsequent use of CS. Future studies in thisarea are needed to determine whether situational characteristics within sports contests influence theathletes’ coping responses, an additional test of trait and contextual coping theory.r 2006 Elsevier Ltd. All rights reserved.

Keywords: Acute stress; Coping style; Competitive sport; Sources of stress; Gender differences

Sources of acute stress (SAS) and athletes’ coping style (CS) in competitive sport as a function of

gender

Coping with stress in sport has received increased attention by researchers in recent years.While much about the coping process has yet to be understood, it has been generallyacknowledged that effective coping strategies are inherent to successful sport performance andthat maladaptive (ineffective) coping is detrimental to both the performance and satisfaction ofsports competitors (Anshel, Kim, Kim, Chang, & Eom, 2001; Krohne & Hindel, 1988). Plausibleexplanations for poor sport performance following maladaptive coping include heightenedmuscular tension (Anshel, Kim, et al., 2001) and narrowed attentional focusing (Krohne &Hindel, 1988). Given the prevalence of experiencing acute stress in sport, surprisingly little isknown about the coping process following stressful events experienced during the contest, and theextent to which coping styles (CS) can be identified among competitive athletes.Researchers generally agree that coping with acute stress is primarily a function of both personal

and situational characteristics (Kaissidis-Rodafinos, Anshel, & Porter, 1997; Suls & Fletcher,1985), a framework known as the contextual, or interactional, coping theory (Miller, 1992).Personal characteristics in the sport psychology literature include the athletes’ CS (Anshel, 1996;Anshel, Williams, & Williams, 2000; Madden, Kirkby, & McDonald, 1989) and their cognitiveappraisal of stressful events (Lonsdale & Howe, 2004). Situational factors would include sources ofstressful events (Anshel & Delany, 2001; Noblet & Gifford, 2002) and perceived intensity of thestressful event (Madden, Summers, & Brown, 1990). Characteristics of stressful situations havebeen shown to describe and predict a persons’ use of coping strategies (Anshel & Kaissidis, 1997;Carver, Scheier & Weintraub, 1989; Kaissidis-Rodafinos et al., 1997; McCrae, 1984).The degree to which athletes use a particular CS following different sources of acute stress

(SAS) has been examined in the coping theory literature. According to the trait theory of coping,individuals respond consistently to various types of stressful events over time in accordance withtheir CS (Averill & Rosenn, 1972; Leventhal, Suls, & Leventhal, 1993; Moos & Swindle, 1990;Roth & Cohen, 1986). The assumption underlying this framework is that responses to stressfulevents are highly correlated with the person’s CS, at least following similar types of stressors.Contextual coping theory, on the other hand, posits that the combination of personal dispositionsand situational factors are required to reliably predict the person’s CS (Folkman, 1992; McCrae,1992; Terry, 1991).In the non-sport literature, researchers have examined the link between stressful events and a

person’s coping responses to those events as a function of personal and situational factors. For

Page 3: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 3

example, CS is increasingly apparent following stressful events that are appraised as highly intenseor under high threat conditions (McCrae, 1992; Phipps & Zinn, 1986; Skinner, Edge, Altman, &Sherwood, 2003). In another example of personal factors that influence coping, Carver et al.(1989), Folkman and Lazarus (1985), and Scheier, Weintraub, and Carver (1986) each found thatsituations perceived as highly controllable elicited approach-coping strategies (e.g., confrontation,problem-solving, positive reappraisal, accepting responsibility). When the situation was perceivedas not controllable, however, more distancing and escape-avoidance patterns were applied. Thus,the context within which acute stress is experienced at least partly explains or predicts anindividual’s coping response to specific stressful events, particularly if the event is perceived ashighly intense. Less well known, however, is the extent to which CS is a function of the type ofstressful event experienced in competitive sport. Because the number of different stressfulsituations in sport is virtually limitless, according to McCrae (1992), it would be advantageous‘‘to make meaningful generalizations about the influence of the stressor on the choice of copingmechanisms if stressors can be organized in terms of common features or dimensions’’ (p. 66).There is an apparent absence of previous research in the sport psychology literature in whichMcCrae’s suggestion has been followed.One construct in the study of CS in both sport psychology and general psychology literatures

has been approach (also referred to as vigilant, attention, active, sensitization, engagement, activecoping) and avoidance (also labelled non-vigilant, passive, desensitization/repression, disengage-ment, avoidant; Anshel, 1996; Anshel, Williams, & Hodge, 1997; Kaissidis-Rodafinos et al., 1997;Krohne, 1993, 1996; McCrae, 1992; Roth & Cohen, 1986). Approach CS generally refers tobehavioral (i.e., taking action) and cognitive (i.e., mental strategies and self-talk) attempts toresolve stress directed towards the threat or its cognitive and emotional inner interpretations(Krohne, 1996; Skinner et al., 2003). Thus, an athlete who engages with the referee after receivinga penalty, either positively (e.g., asking information about the reason for the penalty) ornegatively (e.g., arguing the call), is using approach coping. Avoidance coping, on the other hand,refers to activity directed away from the threat-related cues (Krohne, 1993, 1996), and may also besub-categorized as behavioral or cognitive. For example, an athlete’s use of avoidance copingupon receiving a penalty from the referee would be to psychologically discount the call bylabelling it ‘‘unimportant,’’ or the referee’s ‘‘mistake,’’ then to quickly attend to forthcoming taskdemands. The plethora of terms to describe the same CS construct of approach- and avoidance-coping reflects a limitation in conceptualizing CS and testing the effectiveness of interventionstrategies in sport.The sport psychology (e.g., Anshel, 1996, 2001; Anshel & Anderson, 2002; Anshel & Wells,

2000; Gaudreau, Blondin, & Lapierre, 2002; Kaissidis-Rodafinos et al., 1997; Krohne & Hindel,1988; Poczwardowski & Conroy, 2002) and general psychology coping literature (e.g., Krohne,1993, 1996; Mullen & Suls, 1982; Roth & Cohen, 1986; Skinner et al., 2003; Suls & Fletcher, 1985;Tobin, Holroyd, Reynolds, & Wigal, 1989) are replete with studies using the approach andavoidance CS framework. In an early sport study, Krohne and Hindel (1988) investigatedathletes’ coping activities during competition. They found that elite table tennis players whoemployed cognitive avoidant strategies to cope with critical situations occurring in the course ofthe match won more games in the important tiebreak situation and were less anxious than playerswho did not use such strategies. Successful table tennis players ‘‘were characterized by fewinterfering anxiety reactions (worry cognitions), little vigilant coping, and an extended use of

Page 4: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–244

cognitively avoidant self-regulatory techniques’’ (p. 225). The researchers explained the tabletennis players could not afford to indulge in self-centered interfering cognitions during the gamefor fear of being distracted from subsequent passages of play. More recently, Anshel andAnderson (2002), in support of Krohne and Hindel, found that approach coping was significantlyrelated to negative affect. In addition, significant correlations between the performers’ use ofapproach- and avoidance-coping strategies and their CS, also categorized as approach andavoidance. While the approach- and avoidance-coping framework has received increasinglyextensive attention by researchers, the use of behavioral and cognitive sub-dimensions of these CShas been virtually ignored.According to previous qualitative (Anshel, 2001) and psychometric research (Krohne et al.,

2000), and reviews of related literature (Anshel, Jamieson, & Raviv, 2001; Anshel, Kim, et al.,2001; Krohne, 1993), CS may be examined as approach-behavioral (AppBeh), approach-cognitive(AppCog), avoidance-behavioral (AvBeh), and avoidance-cognitive (AvCog). AppBeh copingconsists of the conscious use of an overt action in response to stressful appraisal of a stimulus orevent in which the person attends to or confronts the stressor (Anshel, 2001; Krohne, 1993).Examples include soliciting information, arguing, or any observable response in attempting toreduce the stressor’s intensity. AppCog coping, on the other hand, consists of a conscious thoughtor emotion in which the individual is ‘‘oriented toward the threat-related aspects of a situation’’(Krohne, 1993, p. 3). Examples include planning, monitoring, anger, strategizing, imaging, andthoughts that promote cognitive arousal. AvBeh coping consists of the conscious decision tophysically remove oneself from a threatening environment. Examples would be walking awayfrom the stress source, or avoiding a threatening or unpleasant situation, also referred to as socialengineering. AvCog coping reflects ‘‘turning away from threatening cues’’ (Krohne, 1993, p. 3).Examples include filtering out information, selective attention (also called disengagement orrepression), and distraction.While it is intuitively appealing to examine approach and avoidance CS, each consisting of

cognitive and behavioral sub-dimensions, in competitive sport, there is relatively less support toinclude an AvBeh CS. Conceptually, not examining an AvBeh CS is understandable among highlyskilled competitive athletes. In an early conceptual article explaining the approach–avoidance-coping framework, Roth and Cohen (1986) have described the potential costs and benefits ofusing approach and avoidance coping. They suggest that implementing an avoidance strategy maybe more appropriate when the situation is less controllable (e.g., receiving a penalty from thereferee), the source of stress is unclear, the person’s confidence is low to moderate, and outcomemeasures are immediate (e.g., losing the ball while the game is in progress). Approach coping,however, is preferred when the athlete seeks situation-relevant input, when action is required,when the cause of stress is known, and when the athlete has relatively high confidence andpossesses good communication skills. Appbeh coping, then, would be deemed inefficient in ahighly skilled competitive sport environment. In addition, from a research perspective, Krohneet al. (2000) were unable to identify an AvBeh CS dimension in their psychometric studyvalidating the (non-sport) Mainz Coping Inventory.Along these lines, Anshel and Anderson (2002) found that an approach CS was a significant

predictor of performance quality, as opposed to avoidance coping, at least for the first set of trials,a time when acute stress receives the most attention. Thus, failing to attend to the causes ofstressful events, and instead, remaining vigilant in preventing future stressors is anathema to many

Page 5: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 5

highly skilled athletes who prefer to remain in control of the situation and to deal directly with thestressor (Hardy, Jones, & Gould, 1996). Therefore, the AvBeh dimension was not included whengenerating a framework for CS in the present study.Absent from most sport coping research is the extent to which an athlete’s CS is consistent

following different types, or categories, of acute stress. For instance, Kaissidis-Rodafinos et al.(1997) and Anshel and Wells (2000) found that the approach and avoidance CS of basketballreferees and basketball players, respectively, could be significantly predicted only as a function ofthe type of stressor experienced. However, while the stressors differed with respect tochangeability (i.e., a bad call versus pain or injury), these studies were limited to relatively fewSAS, and the stressors were specific, resulting in limited generalizability of the athletes’ CS todifferent classifications of stressful events commonly experienced during the contest.Determining the relationships between categories of stressors and the athletes’ CS that follow

these stressors has apparently not been previously studied. Further research is needed todetermine the extent to which the athletes’ CS can be predicted from categories of, rather thansingle, acute stressors. Learning to respond effectively to a class of stressors, as opposed tolearning unique coping skills following single stressful events, would reduce the information loadrequired for storing and retrieving the proper use of specific coping strategies from memory(Krohne, 1993). To Krohne (1993) and Jackson, MacKenzie, and Hobfoll (2000), the copingprocess and reduced information load are based on principles of self-regulation.The goal of learning to cope with categories of stressors is to reduce information overload and

enhance self-regulation processes. Jackson et al. (2000) contend that CS is a self-regulatoryprocess that is linked to the social context in which stress is experienced, what they call ‘‘self-in-social-setting regulation’’ (p. 294). Coping effectiveness is a combination of a person’s CS and thesocial context within which stress is experienced. Ostensibly, a strong relationship between theindividual’s CS and the stressful event’s social context leads to greater self-mastery of copingefforts. This process is more likely to evolve when CS reflect categories of stressful events.Along these lines, Krohne (1993) contends that coping is more efficient when there is a

‘‘structure’’ in place, consisting of stability and regularity. To Krohne, ‘‘stability of change impliesthat a process is replicable. However, this is only possible when the crucial effect mechanismswhich this process is based on have been previously identified’’ (p. 20). For example, Gaudreau,Lapierre, and Blondin (2001) found that athletes used similar coping strategies at three differentphases of a competitive sports event, one manifestation of CS. Each type of stressor, however,may warrant a different CS.The advantages of identifying athletes’ CS in response to categories, rather than distinct, SAS

enhance self-regulation strategies and reduce information load and greater automaticity in thecoping process. In addition, previous coping in sport studies have focused on determiningathletes’ ‘‘usual’’ coping responses following an acute stressor, the extent to which their CSremains consistent following different categories of stressors has been ignored.Finally, one variable that mediates the use of coping strategies in competitive sport, but has

received only scant attention by researchers, is gender differences. For example, Anshel andKaissidis (1997) found that male and female highly skilled athletes both used more approach-oriented coping strategies as compared to their lower-skilled counterparts. However, for the less-skilled male and female athletes, females applied avoidance coping more often than males. Genderdifferences have also been found for 11- and 12-year-old male and female field hockey players

Page 6: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–246

(Anshel & Delany, 2001). For instance, girls used more confidence building self-talk considerablymore than boys, while boys used resignation (e.g., ‘‘I reminded myself that things could be muchworse’’) more often than girls. Lane, Jones, and Stevens (2002), however, found no genderdifferences on coping with failure and changes in self-efficacy among male and female tennisplayers as a function of self-esteem. In addition, maladaptive coping (e.g., self-blame, behavioraldisengagement) was associated with low self-esteem for both genders. While examining genderdifference among athletes is relatively rare, determining the links between types of stressful eventsand subsequent use of the athlete’s CS as a function of gender is apparently unknown.Thus, the purposes of this study were to determine athletes’ SAS experienced during the

competitive event that male and female athletes perceived as highly intense, their respective CS,the relationship between the acute stressors and their CS (approach and avoidance coping incognitive and behavioral forms), and the generalizability of stress sources and CS as a function ofgender. The primary research question being addressed was to determine the extent to whichcompetitive athletes cope similarly, that is, manifest evidence of CS, for different categories ofacute stress experienced during sport competition, and to examine if gender served as a mediatingrole in this process. The implications of linking the source(s) of acute stress with the athlete’s CSwould result in more effective stress management interventions that target individual differencesin the coping process. It was hypothesized that highly intense SAS would be significantly relatedto the athlete’s respective CS, depicted as approach and avoidance, and that these relationshipswould be a function of gender.

Method

Participants

Participants in this study consisted of 176 males and 156 females (N ¼ 332), ranging in agefrom 18 to 23 yr (M ¼ 21:6 yr, SD ¼ 4.86), who were undergraduate majors in health and physicaleducation in a university located in the southeast U.S. One important criterion for participating inthe study was that each individual had competed on his or her high school sports team, asopposed to engaging in recreational sport. While this criterion did not control exclusively for skilllevel (i.e., starting players were not differentiated from non-starters), it could be assumed thatmembership on a school team in the U.S. represents a relatively moderate to high skill level, andthat these athletes experience relatively similar highly intense SAS. Thus, while respondents werecurrently enrolled in college, the results of this study may be generalized to participation in high(secondary) school sport. Sample size for the statistical procedures used in this study iscommensurate with, and even larger, than previous similar research (e.g., Epitropaki & Martin,2004; Srivastava, Locke, & Bartol, 2001).

Materials

An inventory was developed for this study that consisted of two parts. The first part listed 16SAS that were adapted from previous studies in this area (e.g., Fisher & Zwart, 1982; Maddenet al., 1990). The purpose of part 1 was to identify the athletes’ sources of stress perceived as

Page 7: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 7

highly intense for subsequently determining the degree to which they used similar CS followingtwo different acute stressors.There were two criteria for including a stressor in this list: each item was commonly experienced

in sport settings, and the item had to reflect a sudden (acute) stressful event, as opposed to chronic(long-term) sources of stress. Respondents were asked to ‘‘indicate the level of intensity (i.e.,unpleasantness) you felt after experiencing each of the following stressful events’’ on a Likert-typescale: 1 (not at all), 2 (slightly), 3 (somewhat high), 4 (very high), and 5 (extremely high). Inaccordance with the recommendations of researchers following high stress intensity (as opposed toslightly or moderate intensity level) (Aldwin, 1994; McCrae, 1992), the primary selection criterionfor these stressors was based on the players’ perceptions of experiencing somewhat high, very high,or extremely high stress intensity. CS is more apparent when the sources of stress to whichindividuals respond are perceived as highly intense (McCrae, 1992). Individuals may vary in thedegree of consistency they display in coping with the demands of the situation (Compas, Forsythe,& Wagner, 1988).The 16 listed SAS items that were commonly experienced by competitive athletes were obtained

from the extant coping in sport research literature (e.g., Anshel & Delany, 2001; Anshel & Wells,2000; Madden et al., 1990; Noblet & Gifford, 2002; Rawstorne, Anshel, & Caputi, 2000) and frominterviews with 25 male and female college athletes to determine the types of stressful events theytypically experience. The three most intense sources of stress, rated 3 (somewhat high), 4 (very

high), or 5 (extremely high) formed the frame of reference by which athletes indicated their copingresponses in part 2. These sources were categorized as performance- and coach-related based on100% consensus from two researchers familiar with this literature, forming the SAS’s conceptualframework, listed in Tables 1 and 2. The performance-related factor consisted of the items,‘‘received unfair call from the umpire/referee,’’ ‘‘injured and played in pain,’’ ‘‘received a negativecomment,’’ and ‘‘a cheating opponent was not caught.’’ The coach-related factor consisted of theitems ‘‘arguing with my coach,’’ ‘‘coach was upset with me,’’ and ‘‘treated unfairly by my coach.’’It is important to remember that one primary objective of this study was to determine evidence oflinks between categories of acute stress and their concomitant CS.Part 2 consisted of ascertaining the athletes’ CS, listing approach and avoidance categories in

response to each of the two stressful events. Specifically, respondents were asked to indicate the‘‘usual way’’ in which they responded to each SAS. There have been several studies examining theapproach and avoidance coping framework (e.g., Anshel, 1996; Anshel et al., 1997; Kaissidis-Rodafinos et al., 1997; Krohne, 1996), however, to date, an inventory identifying approach andavoidance coping in sport that has received psychometric scrutiny has not been published. Thus,the CS items used in this study were obtained from several previous studies using this framework(e.g., Anshel & Anderson, 2002; Anshel et al., 1997; Krohne, 1993; Rawstorne et al., 2000).A total of 20 coping items were included for each of two sources of stress.The CS’s consisted of three dimensions, AppBeh, AppCog, and AvCog, based on the descriptive

characteristics of the items. An AvBeh CS was not included in this study because it did not reflectthe athletes’ typical responses to acute stress in competitive sport during the contest (e.g.,exercising, avoiding an opponent), as described in Anshel’s (2001) qualitative study. The AppBeh

factor consisted of four items: ‘‘I discussed the problem with others,’’ ‘‘I asked other people togive me their opinion,’’ ‘‘I complained to a friend or another objective party,’’ and ‘‘I discussedthe problem with another person.’’ Second, the AppCog factor consisted of the next three items:

Page 8: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

Table 1

Results of confirmatory factor analysis and multigroup invariance across gender for source of acute stress

Item Factor loading

CFA Male Female

Acute stress category

Factor 1: Performance-related

Received an ‘‘unfair’’ call from the referee .55 .53 .59

I was injured and played in pain .66 .66 .66

Received a negative comment from others .62 .62 .62

My opponent cheated but was not caught .63 .67 .60

Factor 2: Coach-related

Arguing with my coach .68 .68 .67

Coach was upset with me .65 .59 .70

Treated unfairly by my coach .82 .80 .85

Confirmatory Factor Analysis (CFA) was achieved (w2 ¼ 28:56, df ¼ 13, p ¼ :01, w2=df ¼ 2:20, TLI ¼ .99, CFI ¼ .99,

RMSEA ¼ .07).

Multigroup Confirmatory Factor Analysis (MGCFA) for males and females.

Phase 1. Configural invariance (factor structures across gender) was achieved: the male (w2 ¼ 17:17, df ¼ 13, p ¼ :19,TLI ¼ .98, CFI ¼ .99, RMSEA ¼ .04), female (w2 ¼ 34:02, df ¼ 13, p ¼ :00, TLI ¼ .97, CFI ¼ .99, RMSEA ¼ .07).

Phase 2. Metric invariance (factor loadings) was achieved (Dw2 ¼ 2:46, df ¼ 5, p4:78; DCFI ¼ .00) between

unconstrained (w2 ¼ 61:20, df ¼ 26, p ¼ :00, TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .06) and constrained MGCFA

(w2 ¼ 63:66, df ¼ 31, p ¼ :00, TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .06).

Phase 3. Theoretical structures invariance (factor covariances) was achieved (Dw2 ¼ :68, df ¼ 2, po:71, and

DCFI ¼ .00, with constrained all factor covariances to be group invariance MGCFA (w2 ¼ 64:34, df ¼ 33, p ¼ :00,TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .06).

Cronbach’s alphas: Acute stress category ¼ .81; performance-related ¼ .76, coach-related ¼ .71.

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–248

‘‘I tried to obtain more information, ‘‘I tried to use logic or reason to overcome the problem,’’ and‘‘I thought about what to do next.’’ Third, the AvCog factor consisted of the last three items: ‘‘Iprayed to help me deal with the problem or situation,’’ ‘‘I believed the situation was in God’shands,’’ and ‘‘I thought to myself that things could be worse.’’ The dimensions of the three CS’sand their respective items are linked to each dimension as determined by a conceptual framework.

Procedures

Undergraduate students (N ¼ 15) with previous experience as recreational athletes possessingthe same characteristics as the athletes in the actual study (e.g., age, education level, culturalbackground) participated in a pilot test. The purpose of the pilot test was to assess the inventory’scontent validity, including the extent to which instructions were clear and the applicability ofcoping items in competitive sport. No modifications were made to the instructions and to thecontent of any item based on their input.All sample data in the study (N ¼ 332) were obtained voluntarily from current or former

athletes who were currently competing on their university sports team, or were attending course inthe university’s physical education department. The athletes were administered the inventory

Page 9: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

Table 2

Results of confirmatory factor analysis and multigroup invariance across gender for copying style

Item Factor loading

CFA Male Female

Coping style

Factor 1: Approach-behavior

I discussed the problem with others .65 .62 .68

I asked other people to give me their opinion .77 .69 .85

I complained to a friend or another objective party .70 .66 .71

I discussed the problem with another person .88 .86 .89

Factor 2: Approach-cognitive

I tried to obtain more information .65 .63 .66

I tried to use logic or reason to overcome the problem .63 .62 .66

I thought about what to do next .70 .72 .69

Factor 3: Avoidance-cognitive

I prayed to help me deal with the problem or situation .82 .83 .79

I believed the situation was in God’s hands .74 .67 .83

I thought to myself that things could be worse .48 .52 .44

Confirmatory Factor Analysis (CFA) was achieved (w2 ¼ 96:89, df ¼ 32, p ¼ :00, w2=df ¼ 3:03, TLI ¼ .99, CFI ¼ .99,

RMSEA ¼ .07).

Multigroup Confirmatory Factor Analysis (MGCFA) for males and females.

Phase 1. Configural invariance (factor structures across cultures) was achieved: males (w2 ¼ 90:24, df ¼ 32, p ¼ :00,TLI ¼ .97, CFI ¼ .98, RMSEA ¼ .07); females (w2 ¼ 48:82, df ¼ 32, p ¼ :03, TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .06).

Phase 2. Metric invariance (factor loadings) was achieved (Dw2 ¼ 6:73, df ¼ 7, po:46, and DCFI ¼ .00, between

unconstrained (w2 ¼ 139:05, df ¼ 64, p ¼ :00, TLI ¼ .98, CFI ¼ .99, RMSEA ¼ .06) and constrained MGCFA

(w2 ¼ 145:78, df ¼ 71, p ¼ :00, TLI ¼ .98, CFI ¼ .99, RMSEA ¼ .06).

Phase 3. Theoretical structures invariance (factor covariances) was achieved (Dw2 ¼ 2:6, df ¼ 3, po:46, and

DCFI ¼ .00, with constrained all factor covariances to be group invariance MGCFA (w2 ¼ 148:38, df ¼ 74, p ¼ :00,TLI ¼ .98, CFI ¼ .99, RMSEA ¼ .06).

Cronbach’s alphas: Coping Style ¼ .82; approach-behavioral ¼ .82, approach-cognitive ¼ .70, and avoidance-

cognitive ¼ .71.

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 9

either before a practice session or at the start of a class lecture with permission of the coach orcourse instructor, respectively. Duration of time to complete the inventory ranged from 14 to18min.

Measures

Until now, there has been no single and widely accepted measure of athletes’ SAS and theirrespective CS. Different researchers have developed independent lists of traits to measure athletes’SAS and their companion CS. Moreover, both scales tend to be long and, therefore, time-consuming to complete (e.g., Anshel, 1996; Anshel et al., 1997; Fisher & Zwart, 1982; Kaissidis-Rodafinos et al., 1997; Krohne, 1996; Madden et al., 1990). This can be problematic from anapplied research perspective, especially if these two measurements are widely used in a variety of

Page 10: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–2410

samples. Shorter, less time-intensive scales that are both reliable and valid are generally preferred(Epitropaki & Martin, 2004).Nine of the 16 items from the SAS instrument and 10 of the 20 CS items were dropped due to

‘‘low’’ stress intensity level. A minimum criterion was established because stress intensity levelsappraised as ‘‘low’’ have not been shown to identify a person’s CS (Porter & Stone, 1996).Determining CS is increasingly reliable when the stressor is perceived as highly intense, asopposed to low or moderate stress intensity (Beehr & McGrath, 1996), and (2) stressors perceivedat relatively low intensity often do not require a coping strategy (Beehr & McGrath, 1996; Porter& Stone, 1996). Intra-reliability (Cronbach alphas) for the SAS and CS items, respectively, was.81 and .82.

Data analysis

It is important to examine configural (factor structures) and metric (factor loadings) invariancefor all measures across different samples or contexts to make accurate and valid comparisons, andto achieve generalizability of a measurement (Vandenberg & Lance, 2000). Three statisticaltechniques were used to achieve both configural invariance and metric invariance. First,confirmatory factor analysis (CFA) was used to confirm the factor structures derived from theconceptual frameworks for sources of stress and CS (discussed earlier). Second, multigroupconfirmatory factor analysis (MGCFA) was used to test the generalizability of the two (SAS andCS) scales across gender. Finally, structural equation models were used to determine therelationships between latent variables (factors) of the athlete’s source of acute stress and the latentvariables of CS.Chi-square (w2) statistic, the Comparative Fit Index (CFI), Tucker–Lewis Index (TLI), and the

Root Mean Square Error of Approximation (Bentler, 1990; Schumacker & Lomax, 1998) wereused to assess model fit. Low values and insignificant w2 would indicate a better fit. However, thisstatistic is very sensitive for larger sample sizes. w2 may lead to rejection of a model with good fit inlarger sample sizes (Schumacker & Lomax, 1998). Joreskog (1969) proposed the Normed w2, thatis, w2 adjusted by the degrees of freedom (w2/df) to assess model fit. The Normed w2 values between1.0 and 5.0 are considered to fall within the level of acceptance (Schumacker & Lomax, 1998). TheCFI and TLI are evaluated as indicating good model fit to the data if they equal or greater than.90 (.90 ¼ the lower bound of a good fit, .95 or higher ¼ excellent fit). The Root Mean SquareError of Approximation (RMSEA) values below .08 are considered as indication of good fit (i.e.,.08 is the upper limit of a good fit, whereas .05 or less is an excellent fit; Browne & Cudeck, 1989).

Multigroup confirmatory factor analysis (MGCFA) for genderMGCFAs tested the generalizability of athletes’ source of acute stress scale and their CS scale.

First, a baseline model resulting from the conceptual framework, and confirmed by the CFA forthe total sample for each scale, was established. Second, to obtain configural invariance forgender, the baseline model for each scale was tested separately for males and females. Third, thebaseline model was tested simultaneously across gender. In this stage, three nested models weretested as part of each MGCFA. Model 1 was the baseline model without constrained to be equalacross gender. Model 2 was the baseline model with factor loadings were constrained to be equalacross gender. Model 3 maintained equality constraint on the factor loadings, and constrained all

Page 11: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 11

factor covariances to be gender invariant. An insignificant change in w2 among the three modelswould provide support for generalizability across groups (Byrne, 1993). In summary, theMGCFAs were used to test the equivalency of the number of underlying factors and their itemsconstructed for each factor (configural invariance), item measurements or factor loadings (metricinvariance), and theoretical structures, that is, factor covariances (Byrne, 1993; Vandenberg &Lance, 2000).The statistics used to test invariance in MGCFA were w2 change (Dw2), and fit indexes change

(DCFI or DTLI). The differences between models do not exist if Dw2 was not significant or if the fitindex change is very small (i.e., DCFI or if the DTLI ¼ .01 or less: differences between models donot exist).

Results

Source of acute stress

Sources acute of stress (SAS) were generated and categorized based on a plethora of previous-related studies in which CS followed acute stressors in sport experienced during the contest (e.g.,Anshel, Jamieson, et al., 2001; Anshel, Kim, et al., 2001; Anshel et al., 2000; Rawstorne et al.,2000) Based on a 100% consensus of two sport psychology researchers who were familiar withthis literature, as indicated earlier, the stressors were categorized as ‘‘performance-related’’(M ¼ 3:15, SD ¼ .85) and ‘‘coach-related’’ (M ¼ 3:15, SD ¼ .86).CFA confirmed the factor structures resulting from the earlier conceptual framework of the

SAS. The CFA achieved all the statistical criteria discussed earlier (w2 ¼ 28:56, df ¼ 13, po:01,w2=df ¼ 2:20, TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .07). Factor loadings ranged from .55 to .82. Theitems, factor loadings and the statistical measurements for SAS scale are listed in Table 1.

Generalizability of SAS scale for gender

The MGCFA was computed to test the generalizability across gender concerning the athlete’sSAS scale. These include: (1) the number of underlying factors and their respective itemsconstructed for each factor (configural invariance), (2) item measurements, or factor loadings(metric invariance), and (3) theoretical structures, that is, factor covariances.The student athlete’s SAS scale achieved configural invariance. For Phase 1: Male, w2 ¼ 17:17,

df ¼ 13, po:19, w2=df ¼ 1:32, TLI ¼ .98, CFI ¼ .99, RMSEA ¼ .04; Female, w2 ¼ 34:02, df ¼ 13,po:00, w2=df ¼ 2:62, TLI ¼ .97, CFI ¼ .99, RMSEA ¼ .07. For Phase 2: Metric (factor loadingsmeasurement) invariance using MGCFA based on both the likelihood ratio (LR) test or the w2

change (non-significant Dw2) and fit indexes change (very small DCFI, DTLI, or DRMSEA) wasachieved (Dw2 ¼ 2:46, df ¼ 5, p4:78; DCFI ¼ .00) between unconstrained (w2 ¼ 61:20, df ¼ 26,po:00, TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .06) and constrained MGCFA (w2 ¼ 63:66, df ¼ 31,po:00, TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .06). Phase 3: Theoretical structures invariance, that is,factor covariances test was also achieved (Dw2 ¼ :68, df ¼ 2, po:71, and DCFI ¼ .00, withconstrained all factor covariances to be group invariance MGCFA (w2 ¼ 64:34, df ¼ 33, po:00,TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .06). All items from the MGCFA statistics and factor loadingsfor males and females are listed in Table 1.

Page 12: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–2412

Coping style

The coping style three-factor model showed a good fit with the data. The items, factor loadings,and the statistical measurements for coping style scale are listed in Table 2. The CFA wasachieved with all statistics criteria discussed in the method section (w2 ¼ 96:89, df ¼ 32, po:00,w2=df ¼ 3:03, TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .07). The factor loadings ranged from .48 to .88(see Table 2 under CFA factor loadings). Means and standard deviations for CS were Appbeh(M ¼ 2:88, SD ¼ .92), Appcog (M ¼ 2:90, SD ¼ .89), and AvCog (M ¼ 2:96, SD ¼ 1.09).

Generalizability of CS scale for genderFor the athletes’ CS scale: Phase 1. Configural invariance, males, w2 ¼ 90:24, df ¼ 32, po:00,

w2=df ¼ 2:82, TLI ¼ .97, CFI ¼ .98, RMSEA ¼ .07; females (w2 ¼ 48:82, df ¼ 32, po:03,w2=df ¼ 1:51, TLI ¼ .99, CFI ¼ .99, RMSEA ¼ .06). Phase 2: Metric (factor loadings measure-ment) invariance using MGCFA based on both the Likelihood Ratio (LR) test or the w2 change(non-significant Dw2) and fit indexes change (very small DCFI, DTLI, or DRMSEA) was achieved(Dw2 ¼ 6:73, df ¼ 7, p4:46; DCFI ¼ .00) between unconstrained (w2 ¼ 139:05, df ¼ 64, po:00,TLI ¼ .98, CFI ¼ .99, RMSEA ¼ .06) and constrained MGCFA (w2 ¼ 145:78, df ¼ 71, po:00,TLI ¼ .98, CFI ¼ .99, RMSEA ¼ .06). Phase 3: Theoretical structures invariance, that is, factorcovariances test was also achieved (Dw2 ¼ 2:6, df ¼ 3, po:46, and DCFI ¼ .00, with constrainedall factor covariances to be group invariance MGCFA (Dw2 ¼ 148:38, df ¼ 74, po:00, TLI ¼ .98,CFI ¼ .99, RMSEA ¼ .06). All the items, results of MGCFAs’ statistics and factor loadings formales and females can be seen in Table 2.

Relationships model among source of acute stressors and CS

Structural equation modeling (SEM) was used to test the correlations between the SAS and CSfactors. SEM combined the CFA models of the two (SAS and CS) scales into one model,combining the relationships between factors from the two scales. The first phase consisted ofexamining the relationships between acute stressors and CS by combining the CFA models fromthe two scales (see Fig. 1) for the whole sample. The second phase consisted of examining thecorrelations between the SAS and CS scales. This model was developed based on second degreefactor analysis of each scale, correlating general acute stress with general CS for the whole sample(see Fig. 2). The final phase consisted of determining the relationships between acute stressors andCS for gender, applying the male and female samples simultaneously (see Figs. 3 and 4).

Step one: full sample analysesThe results of SEM (see Fig. 1) revealed three major findings: (1) that the model achieved a

good fit; (2) that the factor loadings from the factors on each scale of the constructed items weresignificant, and (3) that there were correlations between the SAS and CS factors. The w2 of themodel was 210.45 (df ¼ 109) and was acceptable (Normed w2 ½w2=df � ¼ 1:93). The other fit indiceswere also supporting the goodness of fit model with the data (CFI ¼ .97, TLI ¼ .99,RMSEA ¼ .05). Fig. 1 includes the w2 and other fit indices.

Page 13: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

Chi-Square = 210.448df = 109

p = 0.000NFI = 0.984CFI = 0.992RFI = 0.977TLI = 0.989

RMSEA = 0.053

AppBeh

C4e4

0.88C3e3

0.71

CI2e2 0.77

C1e10.65

AppCog

C7e7

C6e6

C5e5

0.71

0.63

0.65

AvCog

C10e10

C9e9

C8e8

0.48

0.74

0.82

Perform

S4e24

S3e23

S2e22

S1e21

Coach

S7e27

S6e26

S5e25

0.63

0.63

0.65

0.55

0.82

0.65

0.68

0.41

0.41

0.19

0.18

0.81

0.24

0.57

0.49

0.41

0.10

Fig. 1. Model of the relationship between acute stressors and coping styles for the full sample.

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 13

Significant factor loadings from each factor. As shown in Fig. 1, the SAS consisted of twofactors, performance-related (Perform) and coach-related (Coach). The factor loadings fromPerform to its constructed items were .55, .65, .63, and 63, respectively; the factor loadings fromCoach to its constructed items were .68, .65, and .82, respectively. The CS scale consisted of three

Page 14: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

Chi-Square = 226.413df = 113

p = 0.000NFI = 0.983CFI = 0.991RFI = 0.977TLI = 0.988

RMSEA = 0.055

AppBeh

I4e4

0.88I3e3

0.70

I2e2 0.77

I1e10.65

AppCog

I7e7

I6e6

I5e5

0.70

0.62

0.66

AvCog

I10e10

I9e9

I8e8

0.47

0.74

0.84

Coping Style

e14

e15

e17

0.69

0.85

0.52

Perform

I4e24

I3e23

I2e22

I1e21

Coach

I7e27

I6e26

I5e25

Acute Stress

e28

e29

0.62

0.64

0.66

0.55

0.83

0.65

0.670.89

0.90

0.48

Fig. 2. Model of the relationship between acute stress and coping style for the full sample.

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–2414

factors: AppBeh, AppCog, and AvCog. The factor loadings from AppBeh to its constructed itemswere .65, .77, .71, and .88, respectively. All factor loadings were significant (po:001). The fouritems consisting of the performance-related factor and the three items that consisted of the coach-related factor are listed in Table 1. The items included in factors AppBeh, AppCog, and AvCogare listed in Table 2.

Correlations among factors in the SAS and CS. All the correlations were displayed by the symbol(2) represent the magnitude of the correlation (see Fig. 1). The correlation between the firstsource of stress (performance-related) and the second source of stress (coach-related) wassignificant and positive, r ¼ :81, po:001. Thus, relationship between sources of stress categorizedas performance- and coach-related for stress intensity was significant. Very high correlationbetween factors could be a problem since it reflects redundancy between dimensions.The correlation between the two sources of stress, however, did not exceed the recommended

Page 15: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

Chi-Square = 352.757

df = 218

p = 0.000

NFI = 0.973

CFI = 0.990

RFI = 0.963

TLI = 0.985

RMSEA = 0.043

AppBeh

C4e4

0.84C3e3

0.67

CI2e2 0.70

C1e10.63

AppCog

C7e7

C6e6

C5e5

0.72

0.62

0.63

AvCog

C10e10

C9e9

C8e8

0.53

0.67

0.82

Perform

S4e24

S3e23

S2e22

S1e21

Coach

S7e27

S6e26

S5e25

0.67

0.62

0.66

0.54

0.79

0.60

0.67

0.35

0.43

0.18

0.19

0.75

0.23

0.61

0.44

0.32

0.02

Fig. 3. Model of the relationship between acute stressors and coping styles for males.

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 15

.85 cut-off point for latent factor correlations (Kline, 1998). For CS, AppBeh coping correlatedsignificantly with AppCog, and AvCog, r’s ¼ :57 and .41 (p’so:001), respectively. AppCogcorrelated significantly with AvCog, r ¼ :49 (po:001).

Page 16: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

Chi-Square = 352.757df = 218

p = 0.000NFI = 0.973CFI = 0.990RFI = 0.963TLI = 0.985

RMSEA = 0.043

AppBeh

C4e4

0.90C3e3

0.71

CI2e2 0.84

C1e10.68

AppCog

C7e7

C6e6

C5e5

0.69

0.67

0.65

AvCog

C10e10

C9e9

C8e8

0.44

0.83

0.79

Perform

S4e24

S3e23

S2e22

S1e21

Coach

S7e27

S6e26

S5e25

0.59

0.64

0.68

0.56

0.84

0.70

0.67

0.48

0.35

0.19

0.10

0.86

0.22

0.55

0.58

0.44

0.16

Fig. 4. Model of the relationship between acute stressors and coping styles for females.

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–2416

Page 17: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 17

Among the three CS’s, AppBeh had the highest correlation with the coach-related stressor,followed by AppCog, and AvCog, r’s ¼ :41 (po:001), .19 (po:01), and .18 (po:02), respectively.These results indicate that an athlete who experienced intense coach-related acute stress was morelikely to use AppBeh CS (e.g., ‘‘I discussed the problemy,’’ ‘‘I complainedy,’’ ‘‘I askedy’’)followed by the other CS.The correlations between performance-related stressors and CS (AppBeh, AppCog, and

AvCog) were .41 (po:001), .24 (po:003), and .10 (po:19), respectively. The results indicate thatathletes who experienced performance-related stressors were more likely to use an AppBeh CS,followed by an AppCog CS. AvCog coping, however (e.g., ‘‘I prayedy,’’ ‘‘I thought abouty’’),was not likely to be used.

Step two: relationships between acute stress and CS

This model was developed based on a second-degree factor analysis of each scale and tocorrelate general acute stress with general CS. A second-order factor analysis indicates the factorloadings from each factor to its constructed items, and the factor loadings from eachmeasurement (i.e., SAS or CS) to its factors. The double arrow symbol in Fig. 2 indicates thatthe correlation between the two scales was established. The data indicated a good fit (w2 ¼ 226.41,df ¼ 113, w2/df ¼ 2.00, CFI ¼ .99, TLI ¼ .99, RMSEA ¼ .06).Factor loadings from the first-order (degree) factor analysis from coach-related factor to three

items in the SAS were .67, .65, and .83, respectively. Factor loadings from the AppBeh factor itemson the CS were .65, .77, .70, and .88, respectively. Factor loadings from the second-order factoranalysis (see Fig. 2) from the SAS for performance- and coach-related factors were .90, and .89,respectively. Finally, the factor loadings from the CS for AppBeh, AppCog, and AvCog were .85,.69, and .52, respectively. All the factor loadings were significant (po:0001). Finally, the correlationbetween the two scales (acute stress 2 CS) was r ¼ :48 (po:0001) indicating that general CS wassignificantly related to general SAS. All statistics and fit indices are located in Fig. 2.

Step three: relationships between acute stressors and CS for gender

The model from Step One to determine the correlations between factors from the SAS and CSscales are illustrated in Figs. 3 and 4 for male and female samples, respectively. The model showeda good fit (w2 ¼ 352.76, df ¼ 218, w2/df ¼ 1.62, CFI ¼ .99, TLI ¼ .99, RMSEA ¼ .04).

Male sample. The factor loadings from the performance-related factor items were .54, .66, .62,and 67, respectively; factor loadings from the coach-related factor items were .67, .60, and .79,respectively. The CS consisted of three factors, AppBeh, AppCog, and AvCog. Factor loadingsfrom the AppBeh items were .63, .70, .67, and .84, respectively. Factor loadings from the AppCogitems were .63, .62, and 72, respectively. Finally, factor loadings from the AvCog items were .82,.67, and 53, respectively. Each of these factor loadings were significant (p’so:001).The correlation between the first source of stress (performance-related) and the second source

of stress (coach-related) for the male sample was significant (r ¼ :75, po:001). For CS, theAppBeh CS correlated significantly with AppCog, and AvCog (r’s ¼ :61 and .32, respectively,p’so001). AppCog correlated significantly with AvCog, r ¼ :44 (po:001).AppBeh was the only CS to correlate significantly with the coach-related stressor (r ¼ :43, po:001).

The other two CS, AppCog, and AvCog, failed to correlate significantly with coach-related stressor,

Page 18: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–2418

r’s ¼ :18 (p ¼ :09), and .19 (p ¼ :07), respectively. This indicates that a male athlete experiencing acoach-related stressor was more likely to use an AppBeh CS than AppCog and AvCog CS’s.The correlations between performance-related sources of stress and the three CS’s from this stressor

(i.e., AppBeh, App-, and AvCog) were .35 (po:001), .23 (po:05), and .02 (p ¼ :85), respectively.These results indicate that male athletes experiencing a performance-related stressor were more likelyto use an AppBeh CS, followed by an AppCog CS. An AvCog CS was not likely to be applied.

Female sample. For sources of stress, the factor loadings from the performance-related factoritems were .56, .68, .64, and 59, respectively, and from coach-related factor items were .67, .70,and .84, respectively. For CS, factor loadings from the AppBeh factor items were .68, .84, .71, and.90, respectively. Factor loadings from the AppCog items were .65, .67, and .69, respectively. And,factor loadings from the AvCog items were .79, .83, and .44, respectively (see Fig. 4). Each ofthese factor loadings was significant (po:001).The correlation between the first source of stress (performance-related) and the second source

of stress (coach-related) was significant, r ¼ :86, po:001. This indicated that the intensity of thesestressors was significantly related for the female sample. The AppBeh CS correlated significantlywith AppCog and with AvCog, r’s ¼ :55 and .44, respectively (p’so001). The AppCog CScorrelated significantly with AvCog, r ¼ :58 (po:001).AppBeh was the only CS to correlate significantly with the coach-related stressor (r ¼ :35,

po:001). The other two CS’s, AppCog and AvCog, failed to correlate significantly with coach-related stressor, r’s ¼ :19 (p ¼ :08) and .10 (p ¼ :33), respectively. This indicates thatfemale athletes experiencing coach-related stressors were more likely to use an AppBeh CS(e.g., ‘‘I discussed the problem with others,’’ ‘‘I asked other people to give me their opinion’’).The correlations between performance-related sources of stress and the AppBeh, AppCog, and

AvCog CS’s were .48 (po:001), .22 (p ¼ :06), and .16 (p ¼ :15), respectively. These results indicatethat female athletes experiencing a performance-related stressor were more likely to applyAppBeh coping (e.g., ‘‘I complained to a friend or another objective party,’’ ‘‘I discussed theproblem with others’’), however, not the other CS.In summary, the structural equation model techniques showed that athletes who experience

acute stress used their respective CS consistently. More specifically, athletes who experiencedperformance- or coach-related acute stressors were more likely to use an AppBeh CS; thecorrelations were the highest between both sources of stress and AppBeh coping. By contrast, theleast likely CS used among the athletes was AvCog.

Discussion

The purposes of this study were to determine athletes’ SAS experienced during the competitiveevent that male and female athletes perceived as highly intense, their respective CS, therelationship between the acute stressors and their CS (approach and avoidance coping in cognitiveand behavioral forms), and the generalizability of the stressors and CS as a function of gender.Specifically, we wanted to ascertain the extent to which athletes’ CS was consistent betweencategories of stressors. It was hypothesized in this study that highly intense SAS would besignificantly related to the athlete’s respective CS, depicted as approach and avoidance, and that

Page 19: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 19

these relationships would be a function of gender. The results of the study provided partialsupport for these predictions.Our categories of acute stressors as performance-related (e.g., ‘‘received an unfair call from the

umpire/referee,’’ ‘‘was injured and played in pain,’’ ‘‘negative comment from others’’) and coach-related (e.g., ‘‘arguing with my coach,’’ ‘‘coach was upset with me,’’ and ‘‘treated unfairly by mycoach’’), and the subsequent coping styles of approach (in both cognitive and behavioral forms) andavoidance (only as a cognitive component) appeared supported. We followed Krohne’s (1993)recommendation that formed the basis for conducting CFA on conceptually formed items. Krohnecontends ‘‘factor analysis must be carried out on the basis of theoretical assumptions and that theresults of such a classification must be again related to those same theoretical assumptions’’ (p. 31).The CS three-factor model showed a good fit with the data. In particular, the type of acute

stressor was associated with the athletes’ respective CS. Both performance- and coach-relatedacute stressors were linked to an AppBeh CS. By contrast, the least likely CS used by athletesfollowing both types of stressors was AvCog. AppBeh was the strongest CS to correlatesignificantly with the coach-related stressor.With respect to gender differences, among female athletes, performance-related stressors were

associated with using an AppBeh CS (e.g., ‘‘I complained to a friend or another objective party,’’‘‘I discussed the problem with others’’), however, not the other CS. Male athletes experiencingperformance-related stressors, on the other hand, were more likely to use an AppBeh CS, followedby AppCog CS, but not likely an AvCog CS.Reliabilities of both scales were satisfactory. In addition, results of the CFAs, as well as the

significant correlation between the SAS and CS, observed in Fig. 2, provided measures ofconstruct validity for each scale. Items and concomitant categories of stressors (i.e., performance-related and coach-related) and CS (AP-B, AP-C, AV-C) conceptually generated in this study,were supported. These results lend at least partial support to trait theory of coping, in whichathletes respond consistently to similar types of stressful events in accordance with their CS, assuggested in the sport psychology (e.g., Anshel, 1996) and general psychology (e.g., Averill &Rosenn, 1972; Leventhal et al., 1993; Roth & Cohen, 1986) literature. The assumption underlyingthis framework is that responses to stressful events are highly correlated with the person’s CS, atleast following similar types of stressors. Future studies in this area are needed to determinewhether situational characteristics within sports contests influence the athletes’ coping responses,an additional test of trait and contextual coping theory.Linking categories of sources of stressors with the athletes’ CS, heretofore not addressed in

published sport psychology research, has two distinct advantages in understanding the coping processand creating effective interventions. First, most single items of behavior (e.g., sources of stress) have ahigh component of measurement error and a narrow range of generality (Epstein, 1980). Whenmeasures of behavior are averaged over a number of events, stability coefficients increase. Copingresponses are more predictable. The second advantage to grouping stress sources is improvedgeneralization of results. Instead of determining the athlete’s coping tendencies following a singlesource of stress, researchers and sport psychology consultants are better able to predict the athlete’sresponses following a set of stressors with similar characteristics, thereby suggesting adaptive copingstrategies–and overcoming maladaptive coping tendencies—that have widespread value.The grouping of stressors into categories (i.e., performance- and coach-related) allows

researchers and practitioners to design coping interventions that link the type of CS that is most

Page 20: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–2420

compatible with unique situational (i.e., stressor) characteristics. Teaching athletes to respond tocategories of stressors with similar characteristics will reduce their information load throughoutthe coping process (e.g., perceiving, appraising, and coping with each stressor). Each of theseprocesses requires a brief period of time, thereby reducing the athletes’ attentional focus on thetask at hand (Krohne, 1993). Athletes then quickly select coping strategies that are consistent withtheir style, or typical manner, of coping. Improved coping effectiveness is more likely as a result(Zeidner & Saklofske, 1996). Beck (1993) refers to this process as ‘‘the first principle in thecognitive model of stress reactions’’ (p. 333).To Beck, the stressed individual develops a ‘‘cognitive set,’’ that is, the construction of a

situation, that allows the individual to make a series of appraisals of the external situation, therisks, costs and benefits of reacting in a certain manner, and finally, the decision to respondaccordingly. The stressful situation is conceptualized as ‘‘how does it affect me?’’ (p. 334). Eachsituation warrants distinct characteristics, and unique appraisals and coping responses. However,grouping stressors according to common characteristics (e.g., performance- or coach-related) hasthree advantages to athletes in the coping process.First, grouping stressors allows for greater coping automaticity in the athlete’s ability to ‘‘read’’

the situation and react quickly to it. Coping skills are then enacted more quickly and efficiently.Second, categorizing stressors strengthens the link between the type of stressor experienced andthe athlete’s CS. This was confirmed in the present study by the significant, positive relationshipsbetween stressors and CS. Third, predicting the athlete’s CS (and subsequent use of copingstrategies) in response to selected stressors is strengthened, promoting more effective copinginterventions.Results of the present study also lent support to the approach–avoidance coping framework

used in numerous previous studies (see Anshel, Jamieson, et al., 2001; Anshel, Kim, et al., 2001,and Krohne, 1996 for reviews). In the present study, this framework was embellished by theaddition of behavioral and cognitive sub-dimensions within the CS of approach and avoidancecoping in cognitive form only. These included AppBeh (e.g., ‘‘I discussed the problem withothers,’’ ‘‘I asked other people to give me their opinion,’’ ‘‘I complained to a friend or anotherobjective party,’’ and ‘‘I discussed the problem with another person’’), AppCog (e.g., ‘‘I tried toobtain more information,’’ ‘‘I tried to use logic or reason to overcome the problem,’’ ‘‘I thoughtabout what to do next’’), AvCog (e.g., ‘‘I prayed to help me deal with the problem or situation,’’‘‘I believed the situation was in God’s hands,’’ and ‘‘I thought to myself that things could beworse’’).Lack of consistency in the use of similar coping strategies in different situational contexts

has been shown in previous studies. For example, Bouffard and Crocker (1992) foundthat individuals with physical disabilities did not consistently use the same coping strategiesacross different settings. Testing the goodness-of-fit between the coping tendencies of adolescent-aged athletes across eight stressful events, Anshel (1996) found that CS were a function oftype of stressor. The findings of other previous studies of basketball referees (Kaissidis-Rodafinoset al., 1997) and basketball players (Anshel & Wells, 2000) indicated that personal dispositionsand situational appraisals jointly determine approach and avoidance coping responses.Taken together, the current results suggest that CS in sport contexts, particularly usingthe approach and avoidance framework, may be helpful in response to selected categories ofstressful events.

Page 21: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 21

This study included selected limitations that provide guidelines for conducting future researchin this area. For example, the sample of college students who had participated on their respectivehigh school sports team may not equate with a more homogeneous group, for instance, athleteswho are currently engaged in competitive sport or older, better skilled competitors. Participatingin competitive sport relatively more recently will help ensure more accurate recall, and higherskilled athletes may exhibit different coping skills than their younger, lower skilled counterparts(Anshel, 1996; Anshel, Jamieson, et al., 2001; Anshel, Kim, et al., 2001). In addition, smaller thandesirable sample sizes are often an inherent limitation in factor analytic research (Vandenberg &Lance, 2000). However, the current sample size was acceptable, and previous published studies(e.g., Srivastava et al., 2001) have included sample sizes even smaller than the current study.Finally, the results confirmed evidence of the approach and avoidance CS framework.Another concern in this study was the relatively high correlations between CS, for example,

AvCog and AppCog. These relationships confirm that, at least within the context of competitivesport, coping is primarily a cognitive rather than a behavioral strategy consisting of eitherapproach and avoidance styles, or categories. Nevertheless, it is important to recognize that:(1) approach and avoidance coping each carries its own set of advantages and disadvantages giventhe context within which each are adopted (Roth & Cohen, 1986), and (2) each of these copingstyles have received extensive attention in antecedent sport psychology research. Finally, there areapparent advantages of grouping sources of stress into categories, as shown in this study, forteaching coping skills to competitive athletes. Rather than instruct athletes to use coping strategiesfor specific stressful events, athletes should be able to create a schema (cf. Schmidt, 1975) thatprovides a framework from which to determine the best coping strategy.A schema is a set of rules that serves to provide the basis for a decision. To Schmidt, developing

a schema consists of extracting information from the environment based on related experiencesand combining information obtained to generate a motor program. In response to stressful events,then, results of the current study suggest that athletes may generate generalized rules ofengagement in response to categories of stressful events and then adopt the approach–avoidancecoping framework. The schema reduces the information load that accompanies the ‘‘best’’ type ofcoping strategy following a particular type of stressful event. Further research is needed to test theconcept of generating a coping schema following different types, or categories, of acute stress asgrouped in the present study.From an applied perspective, future research in this area with athletes should continue to

include the approach and avoidance coping framework in predicting the use of adaptive andmaladaptive coping strategies, and then to offer interventions that help ensure appropriate copingresponses (Zeidner & Saklofske, 1996). Since coping is associated with psychological andphysiological well-being and adaptation, improved understanding of the coping process willprovide improved interventions and heighten participation satisfaction in sport (Hardy et al.,1996; Holahan & Moos, 1987).

References

Aldwin, C. M. (1994). Stress, coping, and development: An integrative perspective. New York: Guilford.

Anshel, M. H. (1996). Coping styles among adolescent competitive athletes. Journal of Social Psychology, 136, 311–324.

Page 22: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–2422

Anshel, M. H. (2001). Qualitative validation of a model for coping with acute stress in sport. Journal of Sport Behavior,

24, 223–246.

Anshel, M. H., & Anderson, D. (2002). Coping with acute stress in sport: Linking athletes’ coping style, coping

strategies, affect, and motor performance. Anxiety, Stress, and Coping, 15, 193–209.

Anshel, M. H., & Delany, J. (2001). Sources of acute stress, cognitive appraisals, and coping strategies of male and

female child athletes. Journal of Sport Behavior, 24, 329–353.

Anshel, M. H., Jamieson, J., & Raviv, S. (2001). Cognitive appraisals and coping strategies following acute stress

among skilled competitive male and female athletes. Journal of Sport Behavior, 24, 75–94.

Anshel, M. H., & Kaissidis, A. N. (1997). Coping style and situational appraisals as predictors of coping

strategies following stressful events in sport as a function of gender and skill level. British Journal of Psychology, 88,

263–276.

Anshel, M. H., Kim, K. W., Kim, B. H., Chang, K. J., & Eom, H. J. (2001). A model for coping with stressful events in

sport: Theory, application, and future directions. International Journal of Sport Psychology, 32, 43–75.

Anshel, M. H., & Wells, B. (2000). Sources of acute stress and coping styles in competitive sport. Anxiety, Stress, and

Coping, 13, 1–26.

Anshel, M. H., Williams, L. R. T., & Hodge, K. (1997). Cross-cultural and gender differences on coping style in sport.

International Journal of Sport Psychology, 28, 141–156.

Anshel, M. H., Williams, L. R. T., & Williams, S. (2000). Examining evidence of coping style following acute stress in

competitive sport. Journal of Social Psychology, 140, 751–773.

Averill, J. R., & Rosenn, M. (1972). Vigilant and nonvigilant coping strategies and psychophysiological stress reactions

during the anticipation of electric shock. Journal of Personality and Social Psychology, 23, 128–141.

Beck, A. T. (1993). Cognitive approaches to stress. In P. M. Lehrer, & R. L. Woolfolk (Eds.), Principles and practice of

stress management (2nd ed, pp. 333–372). New York: Guilford.

Beehr, T. A., & McGrath, J. E. (1996). The methodology of research on coping: Conceptual, strategies, and

operational-level issues. In M. Zeidner, & N. S. Endler (Eds.), Handbook of coping: Theory, research, applications

(pp. 65–82). New York: Wiley.

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

Bouffard, M., & Crocker, P. R. E. (1992). Coping by individuals with physical disabilities with perceived challenge in

physical activity: Are people consistent? Research Quarterly for Exercise and Sport, 63, 410–417.

Browne, M., & Cudeck, R. (1989). Simple sample cross validation indices for covariance structures. Multivariate

Behavioral Research, 24, 445–455.

Byrne, B. M. (1993). The Maschlach burnout inventory: Testing for factorial validity and invariance across

elementary, intermediate, and secondary teachers. Journal of Occupational and Organizational Psychology, 66,

197–212.

Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping strategies: A theoretically based approach.

Journal of Personality and Social Psychology, 56, 267–283.

Compas, B. E., Forsythe, C. J., & Wagner, B. M. (1988). Consistency and variability in causal attributions and coping

with stress. Cognitive Therapy and Research, 12, 305–320.

Epitropaki, O., & Martin, R. (2004). Implicit leadership theories in applied setting: Factor structure, generalizability,

and stability over time. Journal of Applied Psychology, 89, 293–310.

Epstein, S. (1980). The stability of behavior. II. Implications for psychological research. American Psychologist, 35,

790–806.

Fisher, A. C., & Zwart, E. F. (1982). Psychological analysis of athletes’ anxiety responses. Journal of Sport Psychology,

4, 139–158.

Folkman, S. (1992). Making the case for coping. In B. N. Carpenter (Ed.), Personal coping: Theory, research, and

application (pp. 31–46). Westport, CT: Praeger.

Folkman, S., & Lazarus, R. S. (1985). If it changes it must be a process. Journal of Personality and Social Behavior, 48,

150–170.

Gaudreau, P., Blondin, J. P., & Lapierre, A. M. (2002). Athletes’ coping during a competition: Relationship of coping

strategies with positive affect, negative affect, and performance-goal discrepancy. Psychology of Sport and Exercise,

3, 125–150.

Page 23: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–24 23

Gaudreau, P., Lapierre, A. M., & Blondin, J. P. (2001). Coping at three phases of a competition: Comparison between

pre-competitive, competitive, and post-competitive utilization of the same strategy. International Journal of Sport

Psychology, 32, 369–385.

Hardy, L., Jones, G., & Gould, D. (1996). Understanding psychological preparation for sport: Theory and practice of elite

performers. New York: Wiley.

Holahan, C. J., & Moos, R. H. (1987). Personal and contextual determinants of coping strategies. Journal of Personality

and Social Psychology, 52, 946–955.

Jackson, T., MacKenzie, J., & Hobfoll, S. E. (2000). Communal aspects of self-regulation. In M. Boekaerts,

P. R. Pintrich, & M. Zeidner (Eds.), Handbook of self-regulation (pp. 275–300). San Diego: Academic Press.

Joreskog, K. G. (1969). A general approach to confirmatory maximum likelihood factor analysis. Psychometrika, 34,

183–202.

Kaissidis-Rodafinos, A., Anshel, M. H., & Porter, A. (1997). Personal and situational factors that predict coping

strategies for acute stress among basketball referees. Journal of Sports Sciences, 15, 427–436.

Kline, P. (1998). The new psychometrics: Science, psychology, and measurement. Florence, KY: Taylor &

Francis.

Krohne, M. W. (1993). Vigilance and cognitive avoidance as concepts in coping research. In M. W. Krohne (Ed.),

Attention and avoidance: Strategies in coping with aversiveness (pp. 19–50). Seattle: Hogrefe & Huber.

Krohne, H. W. (1996). Individual differences in coping. In M. Zeidner, & N. S. Endler (Eds.), Handbook of coping:

Theory, research, applications (pp. 381–409). New York: Wiley.

Krohne, H. W., Egloff, B., Varner, L. J., Burns, L. R., Weidner, G., & Ellis, H. C. (2000). The assessment of

dispositional vigilance and cognitive avoidance: Factorial structure, psychometric properties, and validity of the

Mainz Coping Inventory. Cognitive Therapy and Research, 24, 297–311.

Krohne, H. W., & Hindel, C. (1988). Trait anxiety, state anxiety, and coping behavior as predictors of athletic

performance. Anxiety Research, 1, 225–234.

Lane, A. W., Jones, L., & Stevens, M. J. (2002). Coping with failure: The effecs of self-esteem and coping on changes in

self-efficacy. Journal of Sport Behavior, 25, 331–345.

Leventhal, E. A., Suls, J., & Leventhal, H. (1993). Hierarchical analysis of coping: Evidence from life-span studies. In

M. W. Krohne (Ed.), Attention and avoidance: Strategies in coping with aversiveness (pp. 71–118). Seattle: Hogrefe &

Huber.

Lonsdale, C., & Howe, B. L. (2004). Stress and challenge appraisals of acute taxing events in rugby. International

Journal of Sport and Exercise Psychology, 2, 7–23.

Madden, C. C., Kirkby, R. J., & McDonald, D. (1989). Coping styles of competitive middle distance runners.

International Journal of Sport Psychology, 20, 287–296.

Madden, C. C., Summers, J. J., & Brown, D. F. (1990). The influence of perceived stress on coping with competitive

basketball. International Journal of Sport Psychology, 21, 21–35.

McCrae, R. R. (1984). Situational determinants of coping responses: Loss, threat, and challenge. Journal of Personality

and Social Psychology, 46, 919–928.

McCrae, R. R. (1992). Situational determinants of coping. In B. N. Carpenter (Ed.), Personal coping: Theory, research,

and application. Westport, CT: Praeger.

Miller, S. M. (1992). Individual differences in the coping process: What to know and when to know it. In

B. N. Carpenter (Ed.), Personal coping: Theory, research, and application. Westport, CT: Praeger.

Moos, R. H., & Swindle, R. W., Jr. (1990). Person-environment transactions and the stressor-appraisal-coping process.

Psychological Inquiry, 1, 30–32.

Mullen, B., & Suls, J. (1982). The effectiveness of attention and rejection as coping styles: A meta-analysis of temporal

differences. Journal of Psychosomatic Research, 26, 43–49.

Noblet, A. J., & Gifford, S. M. (2002). The sources of stress experienced by professional Australian footballers. Journal

of Applied Sport Psychology, 14, 1–13.

Phipps, S., & Zinn, A. B. (1986). Psychological responses to amniocentesis: II. Effects of coping style. American Journal

of Medical Genetics, 25, 143–148.

Poczwardowski, A., & Conroy, D. E. (2002). Coping responses to failure and success among elite athletes and

performing artists. Journal of Applied Sport Psychology, 14, 313–329.

Page 24: 2007_PSE_Relationship SAS and CS

ARTICLE IN PRESS

M.H. Anshel, T. Sutarso / Psychology of Sport and Exercise 8 (2007) 1–2424

Porter, L. S., & Stone, A. A. (1996). An approach to assessing daily coping. In M. Zeidner, & N. S. Endler (Eds.),

Handbook of coping: Theory, research, applications (pp. 133–150). New York: Wiley.

Rawstorne, P., Anshel, M. H., & Caputi, P. (2000). Toward evidence of individual differences in coping with acute

stress in sport. Australian Journal of Psychology, 13, 1–8.

Roth, S., & Cohen, L. J. (1986). Approach, avoidance, and coping with stress. American Psychologist, 41, 813–819.

Scheier, M. F., Weintraub, J. K., & Carver, C. S. (1986). Coping with stress: Divergent strategies of optimists and

pessimists. Journal of Personality and Social Psychology, 51, 1257–1264.

Schmidt, R. A. (1975). A schema theory of discrete motor skill learning. Psychological Review, 82, 225–260.

Schumacker, R. E., & Lomax, R. G. (1998). A beginner’s guide to structural equation modelling. Hillsdale, NJ: Erlbaum.

Skinner, E. A., Edge, K., Altman, J., & Sherwood, H. (2003). Searching for the structure of coping: A review and

critique of category systems for classifying ways of coping. Psychological Bulletin, 129, 216–269.

Srivastava, A., Locke, E. A., & Bartol, K. M. (2001). Money and subjective well being. Journal of Personality and Social

Psychology, 80, 959–971.

Suls, J., & Fletcher, B. (1985). The relative efficacy of avoidant and nonavoidant coping strategies: A meta-analysis.

Health Psychology, 4, 249–288.

Terry, D. J. (1991). Coping resources and situational appraisals as predictors of coping behavior. Personality and

Individual Differences, 12, 1031–1047.

Tobin, D. L., Holroyd, K. A., Reynolds, R. V., & Wigal, J. K. (1989). The hierarchical factor structure of the coping

strategies inventory. Cognitive Therapy and Research, 13, 343–361.

Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance literature: Suggestions,

practices, and recommendations for organizational research. Organizational Research Methods, 3, 4–69.

Zeidner, M., & Saklofske, D. (1996). Adaptive and maladaptive coping. In M. Zeidner, & N. S. Endler (Eds.),

Handbook of coping: Theory, research, applications (pp. 505–531). New York: Wiley.