ترس از شکست، هدف دستاورد 2 × 2 و خود ناتوان سازی: بررسی مدل سلسله مراتبی از انگیزه پیشرفت در تربیت بدنی

Abstract In this study, the hierarchical model of achievement motivation [Elliot, A. J. (1997). Integrating the “classic” and “contemporary” approaches to achievement motivation: A hierarchical model of approach and avoidance achievement motivation. In P. Pintrich & M. Maehr (Eds.), Advances in motivation and achievement (Vol. 10, pp. 143–179). Greenwich, CT: JAI Press] is used to investigate the motivational mechanism behind the relationship between fear of failure and self-handicapping adoption. A cross-sectional design was employed. The participants were 691 college students enrolled in physical education in Taiwan. Students completed the Performance Failure Appraisal Inventory (PEAI-S; Conroy, D. E., Willow, J. P., & Metzler, J. N. (2002). Multidimensional measurement of fear of failure: The performance failure appraisal inventory. Journal of Applied Sport Psychology, 14, 76–90), the Chinese 2 × 2 Achievement Goal Questionnaire for Physical Education (CAGQ-PE; Chen, L. H. (2007). Construct validity of Chinese 2 × 2 achievement goal questionnaire in physical education: Evidence from collectivistic culture. Paper presented at the 5th conference of the Asian South Pacific Association of Sport Psychology. Bangkok, Thailand) and the Self-Handicapping Scale (SHS; Wu, C. H., Wang, C. H., & Lin, Y. C. (2004). The exploratory and confirmatory factor analysis of self-handicap scale for sport. Journal of Higher Education in Physical Education, 6(1), 139–148). Structural equation modeling was conducted. Generally, the results showed that mastery-avoidance and performance-avoidance goals partially mediated the relationship between fear of failure and self-handicapping. The results are discussed in terms of the hierarchical model of achievement motivation, and its implications for physical education are also highlighted.

1. Introduction In achievement settings such as physical education classes, the maladaptive behavior of self-handicapping can sometimes be observed. At times, students may claim that they are ill or may provide some unfounded excuse just before executing a challenging task. At other times, they might reduce their effort on purpose in a competitive achievement setting to mask their possible incompetence. Although self-handicapping may protect self-worth in the short term, research indicates that it has high long-term costs for the individual (Zuckerman & Tsai, 2005). Zuckerman and Tsai’s longitudinal studies found that self-handicapping led to worse health and well-being, lower competence satisfaction, lower intrinsic motivation, more frequent negative moods and symptoms, and higher self-reported use of various substances. Given the wide range of negative effects associated with self-handicapping, it behooves researchers to better investigate why self-handicapping occurs. Therefore, the main aim of current study was to investigate the motivational antecedents of self-handicapping in physical education classes. In the current study, we have adopted the hierarchical model of achievement motivation (Elliot, 1997, Elliot, 1999, Elliot, 2006, Elliot and Church, 1997 and Elliot, 1999) to account for the motivational process that triggers self-handicapping in Physical Education (PE). Past research suggests that PE is an ideal setting for examining the link between achievement motivation and self-handicapping among students (see Ommundsen, 2001) given that PE lessons often require students to overtly display their physical abilities. Therefore, incompetence can potentially be readily observed by others in such a setting. As such, in PE classes, students might be highly motivated to adopt self-handicapping to prevent themselves from being perceived unfavorably in public ( Chen et al., 2008). 1.1. Theoretical rationale Berglas and Jones (1978) defined self-handicapping as the action of claiming or creating obstacles to account for poor performance. An individual embraces self-handicapping strategies to protect his or her self-worth or competent image in public. Self-handicappers believe that such acts can mask the relationship between performance and evaluation should they fail to perform (Siegel, Scillitoe, & Parks-Yancy, 2005). Thus, through self-handicapping, personal self-worth can be protected. In a way, the use of self-handicapping alleviates the threat to the self in the short run (Urdan & Midgley, 2001). According to the definition of self-handicapping, it can be argued that the major antecedent of self-handicapping is fear of failure. This is because fear of failure, defined as the dispositional tendency to avoid situations with possible negative outcomes due to the risk of feeling ashamed of failure, whether it is real or putative (Elliot & Thrash, 2004), will lead an individual to adopt a self-handicapping strategy to avoid a decrease in personal self-worth. It is logical to expect and find a positive relationship between fear of failure and self-handicapping. In fact, Elliot and Church (2003) previously reported a positive relationship between fear of failure and self-handicapping. However, for psychologists, the important issues are these: How does fear of failure influence self-handicapping, and what is the mechanism behind the negative impact of fear of failure on self-handicapping? Based on the hierarchical model of achievement motivation (Elliot, 1997), achievement goals can be regarded as playing a crucial role in the mechanism behind the negative impact of fear of failure on self-handicapping. In the hierarchical model of achievement motivation, Elliot (1997) integrated the class motives (e.g., need for achievement and fear of failure) and the contemporary approach–avoidance achievement goal perspective. The motives in Elliot’s model represent an individual difference in affective dispositions that derives energy to drive people’s actions in the general achievement setting. The goal provides the direction (either approach or avoidance) that complements the motives that account for the individual’s intention in a specific condition. Finally, the behavior is performed in response to the goals adopted ( Fryer & Elliot, 2007). Briefly, the motive exerts its influence on the achievement-relevant outcomes through the achievement goal, and the achievement goal is expected to account for achievement-related behavior more directly than the distal motives ( Conroy & Elliot, 2004). Thus, the hierarchical model describes an achievement-striving process that stems from individual differences to the goals adopted and the end states in the achievement situation. According to the hierarchical model, fear of failure is regarded as a motive, while self-handicapping represents achievement behavior, and thus, the mechanism that links fear of failure and self-handicapping is the achievement goal. However, different kinds of achievement goals may play different roles in the relationship between fear of failure and self-handicapping. Achievement goals, which are defined as “the purposes for engaging in competence-relevant behavior” (Moller & Elliot, 2006, p. 308), are primarily focused on how an individual’s competence is defined by differentiating mastery and performance goals. Individuals who endorse mastery goals evaluate their competence according to an absolute or intrapersonal standard. On the contrary, those who subscribe to performance goals focus on attaining a normative standard (Ames, 1992, Ames and Archer, 1988 and Pintrich, 2000). Elliot and Church (1997) integrated the valence of competence into the achievement goal to represent the approach and avoidance tendencies. They proposed the trichotomous model; this model crosses the two aspects of competence, distinguishing between the mastery (approach) goal (MAp, focused on attaining intrapersonal or task-based competence), performance-approach goal (PAp, focused on attaining normative competence), and performance-avoidance goal (PAv, focused on attaining normative incompetence). Recently, Elliot and McGregor (2001) also differentiated the mastery goal into approach and avoidance forms and proposed the mastery-avoidance goal (MAv, focused on attaining intrapersonal or task-based incompetence), forming a 2 × 2 achievement goal framework. Previous empirical research had documented that the achievement goals in the framework have different patterns of antecedents and consequences, providing support for the utilization of the 2 × 2 framework over the dichotomous and trichotomous models (for a review, see Moller and Elliot (2006); see also Roberts, Treasure and Conroy (2007)). Accordingly, based on the hierarchical model, the current study aims to examine the possible links between fear of failure and self-handicapping by means of the four achievement goals. First, in order to protect self-worth, fear of failure orients individuals either to not performing worse than their peers or to not doing worse than they have in the past (Elliot and Church, 1997 and Elliot and Thrash, 2004). Therefore, it is likely that fear of failure would prompt an individual to adopt negative goal forms such as PAv and MAv because fear of failure orients people to negative and undesirable possibilities (Conroy and Elliot, 2004 and Elliot and Church, 1997). Furthermore, some individuals might seek achievement as a result of wanting to avoid failure (Elliot & Church, 1997); hence, fear of failure is also expected to be positively related with PAp. Finally, MAp would be unrelated to fear of failure because of its purely appetitive nature. These hypothesized relationships between fear of failure and the four achievement goals have been discussed and demonstrated in several studies within the 2 × 2 achievement goal framework (e.g., Conroy, 2004, Conroy and Elliot, 2004, Conroy et al., 2003a and Elliot and McGregor, 2001). Further, regarding the relationships between the four achievement goals and self-handicapping, we expected that MAp and PAp would negatively associate with self-handicapping since MAp pertains to an absolute/intrapersonal competence that may reduce the need for a self-protection process. On the other hand, PAp concentrates on pursuing achievement, which requires that an individual not do anything to potentially impede his or her performance, such as self-handicapping behaviors (Elliot, Cury, Fryer, & Huguet, 2006). Thus, we expected that self-handicapping would be triggered by avoidance of achieving normative incompetence, which is the core consideration of PAv, and that self-handicapping may serve as a channel to release the individual’s evaluation threats. These hypotheses have been supported by previous studies (Elliot et al., 2006, Midgley and Urdan, 2001 and Ommundsen, 2004). However, it is difficult to make a hypothesis on the relationship between MAv and self-handicapping because MAv contains both a positive definition and a negative valence of competence. Based on the findings that self-handicapping is motivated by avoidance motivation (Elliot and Church, 2003 and Rhodewalt and Vohs, 2005) and MAv is associated with avoidance of help seeking and executive help seeking (Karabenick, 2003 and Karabenick, 2004), we predicted that MAv would be positively linked to self-handicapping. Urdan, Ryan, Anderman, and Gheen (2003) indicated that these two behaviors are conceptually similar to self-handicapping. 1.2. The present study Based on the prior work discussed above, it is interesting to examine this relationship in Taiwanese students because of the significant cultural differences between Eastern and Western cultures (McCarthy, 2005). It has been suggested that individuals from the East are more sensitive to negative self-relevant information. Socialization processes in Eastern cultures emphasize the importance of not making mistakes or not losing for establishing a positive view of the self, while Western cultures construct the positive self by doing one’s best or by striving to win (Heine, Lehman, Markus, & Kitayama, 1999). This avoidance tendency rooted in Eastern cultures renders individuals from Eastern cultures more susceptible to fear of failure (Eaton & Dembo, 1997) and to adopting avoidance goals (Elliot, Chirkov, Kim, Sheldon, 2001). Potential cultural differences between Eastern and Western individuals in the attainment of a positive view of the self should be highlighted. In addition, two forms of self-handicapping behaviors (reducing effort and making excuses) were differentiated in the current study. These two forms of self-handicapping separate into behavioral handicaps and handicaps that focus on individuals’ claims (Arkin and Baumgardner, 1985, Elliot et al., 2006, Rhodewalt, 1990, Ryska et al., 1999 and Snyder and Smith, 1982). This distinction is important because an individual might not necessarily make a verbal claim of her or his reduced effort, and conversely, an individual may not act out the excuse of decreased effort that has been offered (Ryska, Yin, & Boyd, 1999). Ryska et al., 1999, Wang, 2003 and Wang, 2005 used these two forms of self-handicapping and Nicholls, 1984 and Nicholls, 1989 framework of goal orientation in achievement motivation to investigate the relationships between self-handicapping and goal orientation. They found that athletes with a task-involvement orientation have a negative correlation with the “reducing effort” form of self-handicapping, but a positive or zero correlation with the “making excuses” form. Their results suggest that a task-involvement orientation prevents athletes from reducing their efforts, but may lead to a greater tendency to make excuses when a self-handicapping strategy is elicited. Since Ryska et al. (1999) and Wang, 2003 and Wang, 2005 suggested that the two forms of self-handicapping behaviors may have different relationships with achievement orientation, we differentiated between these two forms of self-handicapping behaviors here to examine whether these two kinds of self-handicapping behaviors have different (or the same) relationships with other research variables (i.e., fear of failure and the four achievement goals). In summary, the main purpose of this study was to examine the relationships among fear of failure, 2 × 2 achievement goals and the two kinds of self-handicapping behaviors based on the hierarchical model of achievement motivation (Elliot, 1997 and Elliot and Church, 1997). It was expected that achievement goals would mediate the relationship between motives and achievement-related outcomes (i.e., self-handicapping). Specifically, it was expected that fear of failure would positively predict PAp, PAv and MAv. In addition, MAv and PAv were expected to have positive relationships with the two self-handicapping strategies, whereas the two approach-based goals were expected to have negative relationships with both types of self-handicapping behaviors. Furthermore, because the hierarchical model of achievement motivation proposes that the 2 × 2 achievement goals cannot fully mediate the relationship between fear of failure and outcome (Fryer & Elliot, 2007), in the following analysis we also added a direct effect of fear of failure on the two self-handicapping behaviors. Fig. 1 presents the hypothesized model. Structural equation modeling (SEM) was conducted to examine the model. Hypothesized model of current study. Dotted line indicates the negative paths. ... Fig. 1. Hypothesized model of current study. Dotted line indicates the negative paths. FF: fear of failure; MAp: mastery-approach goals; MAv: mastery-avoidance goals; PAp: performance-approach goals; PAv: performance-avoidance goals; ME: making excuses; and RE: reducing efforts.

. Results 3.1. Descriptive and correlation analysis Table 1 presents the means, standard deviations and correlations of the seven research variables. Correlation results showed that FF is positively related to MAv, PAv, ME and RE (r = .27–.40, all ps < 0.01). The four achievement goals were significantly related to each other. Specifically, the correlation between MAp and PAv was negative (r = −.10, p < 0.05). The correlations between the four achievement goals were positive (r = .21–.37, all ps < 0.01). For the self-handicapping subscale of making excuses, it was found that ME positively correlated with MAv and PAv (r = .12 and .43, ps < 0.01). However, MAp had a significant negative correlation with ME (r = −.25, p < 0.05). As for reducing efforts, MAv, PAp and PAv had significant positive correlations with RE (r = .15–.31, ps < 0.01). Finally, the two kinds of self-handicapping were positively correlated (r = .44, p < 0.01) 1. 3.2. Structural equation modeling analysis The research model was examined using SEM. First, the measurement model of seven constructs was examined. In this model, each factor was indicated by its items, and errors of items were uncorrelated. Variances of latent factors were set as one to fix their scales. The seven factors were allowed to be correlated. Although the skewness and kurtosis for items are in the range between −1 and 1, the Mardia kurtosis for multivariate normality statistics is 61.30. In order to account for the non-normality, a robust estimator with Satorra–Bentler scaled chi-square was used with Mplus (Muthén & Muthén, 2007). The Satorra–Bentler chi-squared value of this model was significant (SB-χ2 = 627.125, df = 231, p < 0.05) because of the large sample size, however, various fit indices indicated that this model is acceptable (TLI = 0.92; CFI = 0.93; RMSEA = 0.059; SRMR = 0.07). All factor loadings were significant in the model, revealing that measures of the seven constructs are appropriate. Further, the hypothesized model displayed in Fig. 2 was examined by specifying hypothesized paths based on the measurement model. In this model, FF has a direct effect on three achievement goals (MAv, PAp, and PAv) and both kinds of self-handicapping (ME and RE). All four achievement goals (MAp, MAv, PAp, and PAv) have direct effects on both kinds of self-handicapping (ME and RE). In addition, because past studies (Conroy et al., 2003a, Elliot and McGregor, 2001, Nien and Duda, 2008 and Wang et al., 2007) and the correlation analysis in the previous section suggested that the four achievement goals were related, the four achievement goals were also specified as being correlated with each other. As the three achievement goals (MAv, PAp, and PAv) were endogenous latent variables in the model, the correlations among the four achievement goals were specified by correlating their residual variances. This model was estimated using a maximum likelihood estimator with Satorra-Bentler scaled chi-square in Mplus (Muthén & Muthén, 2007) as well. Structural diagram of the hypothesized model. Only significant paths are ... Fig. 2. Structural diagram of the hypothesized model. Only significant paths are presented, and the measurement part was omitted for simplicity. Figure options Although the Satorra–Bentler chi-squared value of this model was significant (SB-χ2 = 683.18, df = 233, p < 0.05) because of the large sample size, various fit indices indicated that this model is acceptable (TLI = 0.91; CFI = 0.92; RMSEA = 0.053; SRMR = 0.07). Estimates in the measurement part are all significant. Reliability (coefficient H) values computed from standardized factor loadings ( Hancock & Mueller, 2001) of these seven constructs are 0.85 for FF, 0.87 for MAp, 0.87 for MAv, 0.90 for PAp, 0.73 for PAv, 0.76 for ME, and 0.67 for RE (see Table 1). However, in the structural part, some parameters were not significant. Specifically, the direct effect of MAp on ME and MAv on RE was insignificant. Among the correlations between the four achievement goals, PAv was not related to MAp. All other relationships were significant. Specific results of this model are presented in Table 2. In addition, in order to empirically test the non-significant relationship between FF and Map, we added this path in the model without changing other specifications in the model. The results showed that the path from FF and Map is not significant and results on overall model fit are not changed, because adding this path did not contribute to data fit. 3.3. Mediation effect of achievement goals Mediation effects of achievement goals were tested by the indirect effects from fear of failure to ME and RE through the three types of goals (MAv, PAp and PAv) in Mplus (Muthén & Muthén, 2007). It was found that only MAv and PAv had significant mediation effects on the relationships between fear of failure and ME (unstandardized effect = 0.05, standardized effect = 0.04, p < 0.05 for MAv; unstandardized effect = 0.17, standardized effect = 0.13, p < 0.01 for PAv), and only PAv had a significant mediation effect on the relationship between fear of failure and RE (unstandardized effect = 0.31, standardized effect = 0.21, p < 0.01). This finding reveals that the three types of goals have different mediation effects. When this model was compared to the full mediation model, in which the two direct effects of fear of failure on ME and RE were dropped, a Satorra–Bentler chi-squared difference test showed that this model was different from the full mediation model (SB-χ2 (2) = 38.61, p < 0.01), suggesting that the two direct effects of fear of failure on ME and RE cannot be deleted. Hence, by model comparison, it can be concluded that effects of FF on ME and RE are not completely mediated by achievement goals.