Sport environments can provide excessive stress for athletes, and long-term exposure to the stress may cause burnout [1
]. Raedeke [3
] first proposed the multidimensional construct of burnout by adopting Maslach and Jackson’s definition of burnout (i.e., a psychological syndrome of emotional exhaustion, depersonalization, and reduced personal accomplishment) [4
]. However, because depersonalization in Maslach and Jackson’s definition is negative feelings and reactions towards clients and is less applicable to athletes, Raedeke [3
] replaced depersonalization with sport devaluation in order to explain athletes’ negative feelings and attitudes toward their sports and defined athlete burnout as a syndrome composed of emotional and physical exhaustion, sport devaluation, and a reduced sense of accomplishment. Due to the lack of a valid and reliable questionnaire to measure burnout in athletes, Raedeke and Smith [5
] later developed a sport-specific burnout questionnaire (i.e., Athlete Burnout Questionnaire; ABQ). The development of the ABQ advanced burnout research in the athlete population [6
]. Early signs of burnout include emotional and physical tiredness or fatigue, mood disturbance, lack of enjoyment, loss of motivation, and perceptions of inadequate social support [3
]. Smith [8
] proposed four stages of the cognitive-affective model to better understand athlete burnout. The first stage is related to “interactions between the environmental demands and personal and environmental resources” [8
] (p. 41). High competitive demands, low social support, and low autonomy can increase or decrease the environmental demands. In the cognitive appraisal stage, each athlete appraises the situational demands unequally. The imbalance between demands and resources causes stress. The athlete shows physiological responses (e.g., tension, anxiety, depression, and fatigue) when perceiving the demand as threatening. The physiological responses lead athletes to the last stage, coping and task behaviors, including decreased performance and withdrawal from activities. Within the cognitive-affective burnout model, withdrawal from sports would be one of the behavioral consequences, but burnout is not “the primary cause of sport withdrawal” [9
] (p. 277). Recent studies in various settings showed work-related stress was a significant predictor of burnout [10
], and studies in sport-specific settings also provided supportive results that chronic stress is highly related to burnout [9
Anxiety is a reaction by an individual to a stressful situation [17
], and athletes in competitive sports possibly have a great deal of performance-related stress. In early studies, researchers modified and used general anxiety measures to examine anxiety in sports, but they found sport-specific anxiety measures to be better predictors of athletes’ behavior. For example, the Sport Anxiety Scale-2 has been used to examine the multidimensional trait anxiety in sport contexts [18
]. Theses sport-specific anxiety measures have helped researchers to obtain valid and reliable data in order to investigate the effects of anxiety on athletic performance, injury, and burnout in athletes.
The relationship between anxiety and burnout was predictable from Smith’s cognitive-affective model of athlete burnout [8
]. As mentioned in the model, burnout is related to a complex cognitive process and is a consequence of chronic stress which can be intimately related to anxiety, especially cognitive anxiety. Studies have supported that anxiety would be one of the predictors of burnout in athletes [14
]. Athletes reported feelings of frustration, lack of confidence, and concentration problems as mental symptoms of burnout that could also be interpreted as cognitive anxiety [13
]. Trait anxiety as a dispositional characteristic was the best predictor of burnout among all intrapersonal and situational predictors, and the cognitive appraisal of and physiological responses to stress could influence the development of burnout [16
]. Cremades et al. [19
] also found several significant correlations between trait anxiety and burnout in collegiate athletes. The higher the levels of trait anxiety, the more risk an athlete has of becoming burned out [20
There are diverse sources of stress for athletes, and identifying and understanding them helps to understand the development and prevention of anxiety as well as burnout in athletes. One of the potential sources can be coaches’ behaviors as an interpersonal factor. In sports, coaches play an influential role in affecting anxiety in athletes [21
], and previous studies found a significant relationship between perceived coaching behaviors and anxiety. Gould and Weinberg [22
] suggest the use of the interaction approach, which includes interpersonal and situational factors to understand anxiety and burnout in competitive sport environments. Ryska and Yin [23
] examined the relationship between the athlete’s perceptions of coach support and precompetitive anxiety in high school athletes and found high coach support lowered precompetitive anxiety. Baker, Côté, and Hawes [24
] have highlighted that negative perceptions of coaching behaviors had a positive correlation with trait anxiety levels in athletes. Baker et al. also suggested levels of anxiety in athletes would increase if a negative relationship between a coach and athlete exists. Studies also found a significant relationship between perceived coaching behaviors and burnout. Autonomy-supportive coaching behaviors were negatively related to athlete burnout, whereas controlling coaching behaviors were positively related to athlete burnout [25
]. Less controlling and more autonomy coaching behaviors might lower levels of burnout or avoid the development of burnout in elite athletes [25
]. Athletes with more autonomy-supportive environments also felt less anxious and burned-out [27
]. Recent studies in non-sport settings also showed consistent results that perceptions of leaders’ and supervisors’ support significantly predicted burnout [11
]. Vealey and colleagues [15
] examined the effect of athletes’ perceptions of coaching behaviors on burnout and competitive anxiety and also tested the relationship between athletes’ levels of competitive (trait) anxiety and burnout. All competitive trait anxiety subscales (i.e., somatic anxiety, worry, and concentration disruption) significantly predicted burnout in athletes. Somatic anxiety was a weaker predictor of burnout than cognitive anxiety (i.e., worry and concentration destruction).
In sum, from the results of previous research, some correlations and causal relationships were revealed between perceived coaching behaviors, anxiety, and perfectionism. However, these results were inconsistent. For example, although Vealey et al. [8
] did not find any relationship between perceived coaching behaviors and anxiety, they found that perceived coaching behaviors significantly influenced anxiety [23
]. The inconsistent findings have required further research; however, there was no study investigating causal relationships. Therefore, the purpose of this study was simultaneously to examine the mediating effects of competitive trait anxiety on the relationship between perceived coaching behaviors and athlete burnout.
The purpose of this study was to examine the relationship between athletes’ perception of their coaches’ coaching behaviors, competitive trait anxiety, and burnout. Specifically, we aimed to investigate the medication effects of competitive trait anxiety on the relationship between perceived coaching behaviors and athlete burnout.
First, consistent with previous studies, trait anxiety in athletes had a significant correlation with athlete burnout (r
= 0.25, p
< 0.01), as well as significant pathways (β
= 0.25, p
< 0.05). This finding confirmed that trait anxiety caused by chronic stress anxiety is an antecedent of burnout in athletes [14
]. All three components of trait anxiety were significantly related to physical and emotional exhaustion and sport devaluation, but intriguingly there was no significant correlation between all three trait anxiety components and a reduced sense of accomplishment. This finding is inconsistent with previous studies [15
], showing that all three components of trait anxiety were significantly related to accomplishment as well as the other two dimensions of athlete burnout and, thus, more trait-anxious athletes felt less personal accomplishment.
The results of the correlations for the composite scores indicated that controlling coaching behaviors were significantly related to athlete’s competitive trait anxiety (r
= 0.14, p
< 0.01), whereas autonomy-supportive coaching behaviors were not significantly related to trait anxiety (r
= −0.03, p
= 0.57). The structural equation modeling results also showed a significant positive pathway from controlling coaching behaviors to trait anxiety (β
= 0.18, p
< 0.05), but not from autonomy-supportive coaching behaviors to trait anxiety (β
= 0.03, p
= 0.66). The current findings were consistent with Baker et al.’s findings [24
] that negative coaching behavior was the strongest predictor of trait anxiety in athletes; however, it is surprising that autonomy-supportive coaching behaviors were not predictive of anxiety and consistent with previous studies. As Vealey et al. [15
] and Baker et al. [24
] explained, this unexpected finding might be due to the different scales used to measure perceived coaching behaviors. The core construct of both measures used in this study heavily relies on self-determined motivation, whereas other studies used measures based on leadership and other constructs. Additionally, some components in the controlling coaching behavior scale contain similar or the same constructs as other scales used in previous studies. For example, intimidation (“My coach shouts at me in front of others to make me do certain things”) used in this study is similar to the construct of negative rapport (“uses fear” and “yells when angry”) [24
]. It may also be due to the notion that the unidimensional construct of the autonomy-supportive coaching behavior questionnaire was not able to measure various positive coaching behaviors.
Excessive personal control was significantly related to all three subscales in trait anxiety, and intimidation was related to two (i.e., somatic anxiety and concentration disruption), but negative conditional regard did not have significant correlations with any anxiety subscales. Overall, these findings support the notion that “certain coaching behaviors are better predictors of sport anxiety” [24
] (p.116). As shown in Table 2
, excessive personal control had stronger relationships with three more trait anxiety components than intimidation and negative conditional regard had. The participants in this study were collegiate athletes, and they have many different kinds of stress, such as athletic stress and academic stress. When they felt that their coaches tried to control not only sport-related matters but also their life and free time outside of sports, athletes can perceive these behaviors as “over-intrusive behaviors” [27
] (p.197) which can also be a strong precursor to increase anxiety levels in athletes.
Both autonomy-supportive and controlling coaching behaviors were significantly related to athlete burnout (r
= −0.33, p
< 0.01 and r
= 0.41, p
< 0.01, respectively). All components of controlling coaching behaviors were also significantly correlated with all three dimensions of athlete burnout. Excessive personal control had the strongest correlations with the three athlete burnout dimensions. The structural equation modeling result indicated that both autonomy-supportive and controlling coaching behaviors had significant pathways to athlete burnout (β
= −0.21, p
< 0.05 and β
= 0.32, p
< 0.001, respectively). The findings support previous research that indicated perceived coaching behaviors was the main precursor of burnout in collegiate athletes [15
]. The bootstrapping results indicated a significant indirect pathway from controlling coaching to athlete burnout via competitive trait anxiety, and this is inconsistent with Vealey et al.’s findings that perceived coaching behaviors had only a direct effect to burnout but did not have indirect effects on burnout via trait anxiety [16
]. As mentioned, this may be due to the fact that each study used different scales.
There are several limitations to generalize the current findings. First, in this study, a cross-sectional design was used to collect the data. The cross-sectional approach is limited in providing clear causal relationships between variables. Athlete burnout may change before, during, and after seasons. It is possible that coaches provide more autonomy-supportive coaching before the season and then more controlling coaching after a season starts. Therefore, future research using a longitudinal approach is needed to examine how perceived coaching behaviors affect athlete burnout and to test the mediating effects of trait anxiety on the relationship between perceived coaching behaviors and athlete burnout. Second, due to relatively small samples, we could not conduct invariance analyses across sex (i.e., males vs. females) and types of sports (i.e., individual vs. team sports). There are usually more numbers of athletes in team sport than in individual sports (e.g., soccer team vs. golf team). Due to bigger numbers in team sports, a coach in a team sport might employ more controlling coaching to manage the number of athletes compared to a coach in an individual sport. Therefore, future research needs to have equal or similar sample sizes in groups and at least 200 participants in each group to conduct invariance analyses. As previous researchers pointed out [15
], there are various scales to measure athletes’ perceptions of their coaches’ behaviors. This caused inconsistent findings in the relationship between perceived coaching behaviors and anxiety. A new questionnaire development to measure perceived coaching behaviors more precisely or a re-evaluation of the current scales should be conducted. This study tested one interpersonal factor (perceived coaching behaviors) and one intrapersonal factor (trait anxiety) that could influence athlete burnout. Other previous studies examined the relationship between other interpersonal factors (e.g., parent–athlete relationship and peer support) and/or intrapersonal factors (e.g., perfectionism) and athlete burnout. Future research should investigate simultaneously the relationship between athlete burnout and multi intrapersonal and interpersonal factors in order to clearly understand the antecedents of burnout and their relationships.