Burnout and turnover intentions in Australian coaches as related to organisational support and perceived control

Abstract

Coach burnout and turnover intentions, as potentially modified by organisational support, perceived control, and coping strategies, were explored in a sample of 406 team and individual sport coaches. Multiple regression analyses revealed that higher perceived organisational support was associated with lower coach burnout scores. Further, coaches' internal locus of control and use of approach coping strategies predicted lower levels of burnout, especially reduced sense of accomplishment. Higher perceived organisational support was also a negative predictor of coaches' turnover intentions, whereas all three burnout dimensions were strong positive predictors. The focus of coach burnout literature has traditionally been on individual factors; this study shows that organisational factors also play an important role for understanding and preventing coach burnout and turnover. We therefore encourage sporting organisations to consider these findings and how they can work towards protecting their coaches against burnout and from leaving the organisation.

Keywords

Burnout coping strategies locus of control organisational support

Introduction

The coaching role has been described as very rewarding, yet very stressful at times; both personal and organisational factors seem to affect perceptions.^1,2 In sport, burnout has been described as a debilitating problem due to its negative physical, emotional, and psychological consequences for individuals and teams alike.^3,4 Burnout often presents with reduced satisfaction and diminishing commitment, which in turn increases turnover intentions and thereby threatens long-term coach retention.^2,5

Although burnout may result from chronic stress, not everyone burns out.⁶ This can possibly be explained by the conservation of resources (COR) model of burnout,^7–9 which proposes that humans are motivated to acquire and preserve resources. Burnout is thereby less likely to occur in coaches with greater personal and environmental resources at their disposal.

Failure to gain resources may therefore according to Hobfoll and colleagues^7–9 contribute to burnout; for example, coaches investing considerable time and energy into their coaching while foregoing family time.¹ Failure to receive desired or expected recognition, pay, or job stability following this resource investment can lead to demotivation and burnout.¹⁰ Chronic stressors such as an enduring losing streak can result in a net loss of resources, for example, loss of esteem, feelings of shame,¹¹ and the threat of losing ones' coaching contract.¹

In contrast, supportive work environments can protect from burnout, particularly when employees feel their contributions are valued and their organisation cares about their wellbeing.¹² The principle of reciprocity – that work-effort will be rewarded by the receipt of resources such as money, praise, or esteem¹³ – is upheld in supportive organisations.^7,8 Supportive work environments are less stressful because they are more stable, predictable, and collaborative places to work.¹⁴

Qualitative studies have consistently revealed that a lack of organisational support contributes to coach stress and burnout.^1,3,15 Organisational stressors identified in the above studies have included interpersonal conflict, poor management, and inferior communication within the organisation, inadequate resources to conduct programs, job insecurity, and a lack of understanding from organisational management about the demands placed on coaches. Given the pervasiveness of organisational stressors in qualitative research and their deleterious effects on coaches, quantitative research is required to examine whether higher perceived organisational support helps protect coaches from burnout.

According to COR theory, people are motivated to invest resources to either prevent or recover from resource loss, an antecedent of burnout.^7–9 Consequently, people will make active attempts to control their environment to influence personal outcomes – a behavioural tendency more common for internals.¹⁶ This relates back to Rotter's¹⁷ theory of locus of control: ‘internals’ perceive they have control over their environment whereas ‘externals’ believe that what happens to them is caused by fate, luck, or powerful others. Studies in organisational settings^18–20 and with athletes^21–23 reveal that lower stress and burnout is observed in internals and those with higher perceived control. Coaches have repeatedly verified that perceived lack of control over resources, equipment, funding, how programs are run, and over their future career prospects cause them considerable stress.^1,15,24 These findings suggest that coaches who possess an external locus of control, or low perceived control are more susceptible to burnout.

One way to affect the stress-burnout relationship is through coping.²⁵ Coaches utilising a variety of approach-based coping strategies including being organised, undertaking proactive contingency planning, seeking support and advice from other coaches, using mentors, focusing on what is controllable, seeking ongoing training and development, and using psychological skills are less likely to burnout.^3,26–28 Approach-based coping has previously in non-coaching populations been associated with lower rates of reduced accomplishment, depersonalisation, and exhaustion¹⁸ as well as better health outcomes and resolution of stressors.^29,30 Only two studies^28,31 have assessed how different coping responses relate to stress outcomes like burnout in coaches. Given that coach burnout can negatively influence performance of coaches and their athletes^3,27; further investigation of the relationship between coping responses and coach burnout is warranted.

As alluded to earlier, a negative consequence of burnout is a heightened turnover risk,³² but turnover has surprisingly received little attention in coach burnout research. Instead, factors influencing coach retention have been explored in isolation.³³ Research from non-sport settings nevertheless reveals that burnout is related to turnover intentions^32,34,35 as are problems dealing with organisational administration.² In contrast, supportive organisational practices contribute to coach retention.^33,36 The relationship between locus of control and turnover is less clear; while some studies indicate that turnover would be more likely among internals,^37,38 others report that internals become more embedded in their organisations and therefore are less likely to leave.³⁹

Based on the above, we hypothesised that higher perceived organisational support, higher internal locus of control, and greater use of approach coping would predict lower levels of coach burnout. Additionally, we hypothesised that higher perceived organisational support would be a negative predictor of turnover intentions, whereas burnout would be a positive predictor. Due to inconsistencies of previous research, the relationship between locus of control and turnover intentions was explored but no specific hypothesis proposed.

Method

Participants

Participants were sought using direct email contact or by snowball convenience sampling with responses received from 453 coaches. After removing 47 cases with unusable data, the final sample contained 406 Australian coaches aged between 15 and 77 years (M = 43.4, SD = 12.7). The sample consisted of 294 males (72%) and 112 females (28%). Participants coached a mix of 36 different team and individual sports (see Table 1). Table 2 outlines the key sample demographics. While there was a reasonably even split between paid and volunteer coaches, the majority of coaches coached part-time, also engaging in other employment outside of coaching. The largest proportion (41%) of participants coached at the community or club level. Participants spent an average of 18.3 hours per week engaged in coaching and coaching related activities (SD = 15.2) and had been involved in coaching their chosen sport for an average of 12 years (SD = 9.7).

Table 1.

Frequencies of sports represented by coaches in the sample.

Sport	n (%)	Sport	n (%)
AFL	128 (30.8)	Rowing	118 (28.4)
Equestrian	20 (4.8)	Gymnastics	16 (3.8)
Basketball	14 (3.4%)	Swimming	11 (2.6%)
Water polo	9 (2.2%)	Other	83 (20.0)
Invalid or not stated	17 (4.1)	Total	406^a

AFL: Australian Football League.

Categories above are not mutually exclusive as 10 coaches indicated that they coached multiple sports.

Table 2.

Key sample demographics.

Demographic variable	n	%
Coaching status
Full-time	78	19.2
Part-time	328	80.8
Payment status
Paid coach	178	43.8
Volunteer	228	56.2
Coaching level
Beginner	23	5.7
School/youth coach	62	15.2
Local community level	166	40.9
State level	65	16.0
National level	47	11.6
International level	31	7.6
Other	12	2.9
Total (N)	406	100

Measures

Demographics

Coaches were asked to provide demographic details about themselves including their age, gender, and relationship status. They were also asked to identify their sport, coaching qualifications, years of experience, and full- or part-time status. The number of hours engaged in coaching and other employment was also recorded.

Coach burnout

Coach burnout was measured using the Coach Burnout Questionnaire (CBQ). The CBQ is a sports-specific burnout measure adapted from the Athlete Burnout Questionnaire⁴⁰ with questions reworded to be more specific to coaching.⁴¹ The CBQ measures three dimensions of coach burnout: physical/emotional exhaustion, reduced sense of accomplishment, and devaluation based on Raedeke's model of coach burnout. Example items for each of the respective subscales include ‘I am exhausted by the mental and physical demands of coaching’; ‘I am not achieving much in coaching’; and ‘I have negative feelings toward coaching’. Participants were asked to think about their sport participation and to indicate how frequently they experienced the feelings listed during the previous season. Responses were given on 5-point rating scales ranging from 1 (almost never) to 5 (almost always). Negatively worded items were reverse-scored, so that higher scores indicated higher levels of burnout. The CBQ has been recommended as the most valid measure of coach burnout as it reliably measures context-specific aspects of coach burnout that other burnout scales do not (see Lundkvist et al.⁴² for details of convergent and discriminant validity, respectively). Previous studies have also shown satisfactory internal consistency, Cronbach's α around .8 (e.g., Malinauskas et al.⁴³).

Locus of control

The Work Locus of Control Scale⁴⁴ was used to measure generalised locus of control within the workplace. Items assessed participants’ general beliefs about jobs. Half of the items measured internal beliefs (Cronbach's α .70; Tong and Wang⁴⁵), for example, ‘Promotions are given to those who perform well on the job’ and the other half measured external beliefs, for example, ‘It takes a lot of luck to be an outstanding employee on most jobs’ [Cronbach's α .8; Tong and Wang⁴⁵]. Responses were given on 5-point rating scales ranging from 1 (strongly disagree) to 5 (strongly agree). Higher scores represented more external responses.

Perceived organisational support

The 16-item short-form of the Survey of Perceived Organisational Support¹² measured coaches’ global perceptions that their sporting organisation appreciated their work and cared about their welfare. An example item is ‘Help is available from the organisation when I have a problem’. The unidimensional scale contained seven negatively worded and nine positively worded questions. Responses were given on a 5-point rating scale ranging from 1 (strongly disagree) to 5 (strongly agree). After recoding negatively worded items, higher scores indicated greater perceptions of organisational support (Cronbach's α .93; Duffy and Lilly⁴⁶).

Approach coping

The Logical Analysis and Problem Solving subscales from the Coping Responses Inventory (CRI, Billings and Moos⁴⁷) measured coaches’ use of active approach coping responses to stressful coaching situations encountered during the last 12 months. Example items included ‘Try to step back from the situation and be more objective’ and ‘Try at least two different ways to solve the problem’. The two subscales were combined to produce a global approach subscale, and after removing one item (item 36 from the original CRI), the scale exhibited acceptable reliability (α = .79). Responses were given on 4-point scales with coaches indicating whether they used a particular strategy, and if so, how often, from not at all (0) to fairly often (3).

Turnover Intentions

Three items adapted from Meyer et al.⁴⁸ measured coaches’ intentions to leave their current coaching position. Coaches were asked to rate their level of agreement with the following questions: ‘I think a lot about quitting coaching because of job stress’; ‘I am actively searching for a position outside of coaching because of job stress’; and ‘I plan to leave the coaching profession in the next year due to job stress’. A fourth option, ‘Not applicable – planning to leave for reasons other than job stress’ was also given. Participants were asked to provide details if they selected this response. Responses were given on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree).

Procedure

Approval to conduct this research was received from the Human Research Ethics Committee, at the University of Canberra (#14-99). Convenience sampling was used whereby 294 coaching liaison contacts and State Sport and Recreation Department officers were identified through State and local sporting organisations websites. These individuals were contacted by email, provided with information about the research and a link to the online questionnaire and were asked to distribute the email to their coaching colleagues.

The questionnaire was housed online on the Qualtrics web survey tool. It was available during June and July 2014, allowing coaches to respond anonymously. Upon completion, participants were thanked for their participation and were given the opportunity to opt-in to receive a summary of the study's results. They were also given chance to enter into a draw to win a sports store voucher as a token of appreciation for their participation.

Statistical analyses

Data were collated and analysed using SPSS (ver. 22). After recoding negatively worded items, composite scores were calculated as the average of the items within each subscale. A correlational design was used, with three hierarchical multiple linear regression (MLR) analyses conducted to determine the extent to which the individual and environmental factors predicted burnout. For each regression model coaches' gender, age, years of experience, weekly hours coached, and weekly hours worked served as control variables, and were entered in step 1. The predictor variables of interest were: perceived organisational support (POS), locus of control (LOC), and approach coping. Because no theoretical causal sequence was identified, the main predictors were entered simultaneously in step 2. Reduced accomplishment, exhaustion, and devaluation, served as the dependent variables for each of the three regression models. A fourth hierarchical regression was conducted, with turnover as the dependent variable; demographic variables were again entered in step 1 as control variables, POS and LOC were entered in step 2 to determine their effect on turnover above that of the control variables. Finally, the three burnout dimensions, reduced accomplishment, exhaustion, and devaluation were entered in step 3 to investigate their unique influence on turnover intentions.

Results

Missing values analysis revealed relatively small proportions of missing data. Little's MCAR test was not significant, indicating that data were missing at random. The SPSS default methodology of listwise deletion was employed for all subsequent analyses. Data were screened for quality and univariate outliers; 16 univariate outliers with z-scores greater than 3.29 were identified during univariate data screening, subsequently excluded from all analyses. Reliability analyses were conducted on the study subscales (see Table 3 for Cronbach's α-coefficients and descriptive statistics). Reliability coefficients (α) were considered acceptable as they all exceeded .70⁴⁹ and were consistent with previous studies.^41,48,50,51 Skewness and kurtosis statistics for the subscales were all below an absolute value of one indicating that the data were reasonably normally distributed.

Table 3.

Descriptive statistics for study variables.

Variable	M	SD	Skewness	Kurtosis	α ^a
Perceived org. support	3.56	.76	−.55	.09	.95
Locus of control^b	2.27	.42	.37	.47	.82
Approach coping	3.04	.47	−.54	−.19	.79
CBQ: Reduced accomplishment	2.02	.62	.47	−.14	.80
CBQ: Exhaustion	2.39	.95	.49	−.31	.93
CBQ: Devaluation	1.97	.73	.60	−.34	.84
Turnover intentions	2.03	.90	.54	−.43	.86

Cronbach's α.

Higher scores represent more external locus of control.

Three hierarchical regression analyses were then conducted to explore the influence of the organisational and individual predictor variables, with each of the three burnout dimensions as the criterion. Demographic variables were entered into the model at step 1 to control for their possible influence on the burnout subscales. As there was no theoretical causal sequence identified for the main predictors, POS, LOC, and approach coping were entered together in step 2.

Table 4 shows the intercorrelations between the predictor variables and each dependent variable. POS and LOC were significantly correlated with each of the criterions in the expected direction. Higher POS was significantly negatively correlated with each burnout subscale, while external LOC was significantly positively correlated with each burnout subscale. Approach coping was significantly negatively correlated with reduced accomplishment and devaluation but was not significantly correlated with exhaustion.

Table 4.

Intercorrelations between predictor variables and burnout dimensions.

Variable	1	2	3	4	5	6	7	8	9	10	11	12
1. Gender	–
2. Age	.22**	–
3. Years experience	.10	.46**	–
4. Hours coached	.11*	−.01	.32**	–
5. Hours worked	.22**	.15**	−.14**	−.47**	–
6. POS	.07	.09	−.08	−.04	.03	–
7. WLOC^a	−.16**	−.05	−.01	−.05	−.06	−.39**	–
8. Coping	.07	.12*	.13*	.14**	−.02	.14**	−.19**	–
9. RA	−.18**	−.14**	−.09	−.06	−.04	−.39**	.28**	−.23**	–
10. Exhaustion	−.19**	−.22**	−.01	.25**	−.17**	−.32**	.20**	.00	.44**	–
11. Devaluation	−.09	−.09	.04	−.02	−.05	−.37**	.19**	−.13*	.68**	.55***	–
12. Turnover	−.10	−.11*	−.01	.05	−.04	−.46**	.24**	.09	.59**	.57***	.67***	–

POS: perceived organisational support; LOC: locus of control; Coping: approach coping; RA: reduced accomplishment.

p < .05;

p < .01 (two tailed).

Higher scores represent more external locus of control.

Assumptions of MLR were tested prior to interpreting final models for each dependent variable. All tolerance statistics were greater than 0.1 and VIF statistics were less than 10, indicating no multicollinearity problems. Inspection of Cook's D showed that there were no influential cases. Durbin-Watson statistics were 1.89, 2.08 and 2.02 for each analysis respectively, demonstrating independence of errors. Inspection of the histograms, the normal probability plot of standardised residuals, and the scatterplot of standardised residuals against standardised predicted values showed that the assumptions of normality and linearity of residuals were met.

Predicting reduced accomplishment

One outlier in the solution with a standardised residual greater than 3.29 SDs from the mean, and two multivariate outliers with Mahalanobis distance greater than the critical value for χ² for df = 8 (α = .001) of 26.13, were identified and removed before interpreting the final model predicting reduced accomplishment. The scatterplot of standardised residuals against standardised predicted values showed a relatively even spread of data indicating that homoscedasticity was assumed.

Gender was the only significant predictor of reduced accomplishment in step 1, t(357) = −2.54, p = .01 with males reporting less reduced accomplishment than females. The step 1 model was significant, F(5,357) = 3.39, p = .005, explaining 5% of the variance in reduced accomplishment. In step 2, the addition of POS, LOC, and approach coping accounted for an additional 19% of variance, ΔF(3,354) = 29.81, p < .001.

The final model (see Table 5) explained 24% of the variance in reduced accomplishment, adjusted R² = .22, F(8,354) = 13.81 p < .001, and with a large effect (f² = .31; Cohen, 1992). POS, LOC, coping, and gender were all significant predictors in the final model (Table 5). POS was the strongest predictor, t(354) = −6.46, p < .001, (β = −.33), uniquely explaining 9% of the variance in reduced accomplishment, approach coping, t(354) = −3.25, p = .001, (β = −.16) uniquely explained 2% of the variance, and LOC (β = .11), uniquely explained 1% of variance, t(354) = 2.11, p = .04. Coaches with a more internal LOC, who perceived their work environment as highly supportive, who employed active coping strategies, or who were male, were less likely to experience reduced accomplishment.

Table 5.

Non-standardised and standardised regression coefficients and squared semi-partial correlations (sr²) for each predictor in the final step of a hierarchical regression model predicting burnout.

	Reduced Accomplishment			Exhaustion			Devaluation
Predictor variable	B [95% CI]	β	sr ²	B [95% CI]	β	sr ²	B [95% CI]	β	sr ²
Gender	−.14 [−.28, .00]	−.10	.01*	−.39 [−.60, −.17]	−.18	.03**	−.05 [−.22, .12]	−.03
Age	−.00 [−.01, .00]	−.03		−.01 [−.02, −.00]	−.14	.01**	−.01 [−.01, .00]	−.08
Years' experience	−.00 [−.01, .00]	−.06		−.01 [−.02, .01]	−.05		.01 [−.00, .02]	.08
Hours' coached	−.00 [−.01, .00]	−.03		.02 [.01, .03]	.31	.06***	−.00 [−.01, .00]	−.05
Hours worked	−.00 [−.00, .00]	−.02		.00 [−.00, .01]	.05		−.00 [−.01, .00]	−.02
POS	−.27 [−.35, −.19]	−.33	.09***	−.36 [−.48, −.23]	−.28	.06***	−.34 [−.44, −.24]	−.36	.10***
WLOC	.16 [.01, .31]	.11	.01*	.14 [−.09, .37]	.06		.01 [−.17, .20]	.01
Coping	−.20 [−.33, −.08]	−.16	.02**	.11 [−.09, .30]	.05	.01	−.10 [−.26, .05]	−.07

CI: confidence interval; POS: perceived organisational support; LOC: locus of control; Coping: approach coping.

p < .05;

p < .01;

p < .001 (two-tailed).

Table 6.

Non-standardised and standardised regression coefficients and squared semi-partial correlations (sr²) for each predictor in the final step of a hierarchical regression model predicting turnover intentions.

Predictor variable	B [95% CI]	β	sr ²
Demographic variables
Gender	−.01 [−.17, .15]	−.00
Age	.00 [−.00, .01]	.03
Years' experience	−.01 [−.01, .00]	−.05
Hours' coached	.00 [−.00, .01]	.05
Hours worked	.00 [−.00, .01]	.04
POS	−.22 [−.32, −.12]	−.19	.02***
WLOC	.02 [−.15, .19]	.01
Burnout variables
RA	.24 [.08, .40]	.16	.01**
Exhaustion	.23 [.14, .32]	.24	.03***
Devaluation	.44 [.30, .57]	.36	.05***

***

CI: confidence interval; POS: perceived organisational support; LOC: locus of control; Coping: approach coping.

p < .01;

p < .001 (two-tailed).

Predicting exhaustion

Two multivariate outliers with Mahalanobis distance greater than the critical value for χ² for df = 8 (α = .001) of 26.13 were identified and removed before interpreting the final model predicting exhaustion. The scatterplot of standardised residuals against standardised predicted values indicated some heteroscedasticity, but as regression is reasonably robust to violations of homoscedasticity,⁵² the model was still interpreted.

In the second hierarchical regression analysis, gender, t(358) = −3.85, p < .001, age, t(358) = −3.11, p = .002, and hours coached per week, t(358) = 5.16, p < .001, were significant predictors of exhaustion at step 1. The model in step 1 was significant, explaining 15% of the variance in exhaustion, F(5,358) = 12.57, p < .001. The addition of POS, LOC, and coping at step 2 together explained an additional 9% of the variance in exhaustion, ΔF(3,355) = 14.45, p < .001.

The final model (see Table 5) explained 24% of the variance in exhaustion, adjusted R² = .23, F(8,355) = 14.16, p < .001, f² = .32, which according to Cohen⁵³ constitutes a large effect. Gender, age, and POS were all significant negative predictors of exhaustion in the final model, with POS the strongest negative predictor, t(355) = −5.51, p < .001, (β = −.28), uniquely explaining 6% of the variance in exhaustion. Hours coached was a significant positive predictor of exhaustion, t(355) = 5.33, p < .001, (β = .31), uniquely explaining 6% of the variance. Coaches with lower perceived organisational support, with higher coaching loads, who were younger, or who were female, reported higher levels of exhaustion. Neither LOC nor coping emerged as significant predictors of exhaustion.

Predicting devaluation

Two multivariate outliers with Mahalanobis distance greater than the critical value for χ² for df = 8 (α = .001) of 26.13, were identified and removed before interpreting the final model predicting devaluation. Some heteroscedasticity was evident, but as regression is reasonably robust to violations of homoscedasticity,⁵² the model was still interpreted.

In the third hierarchical regression analysis, at step 1 the demographic variables accounted for a non-significant 3% of the variance in devaluation, adjusted R² = .01, F(5,358) = 1.83, p = .11. In step 2, POS, LOC, and coping accounted for a further 14% of the variance in devaluation, ΔF(3,355) = 19.21, p < .001.

The final model (Table 5) was significant, F(8,355) = 8.52, p < .001, explaining 16% of the variance in devaluation, adjusted R² = .14, with a medium to large effect size (f² = .19). POS was the only significant predictor of devaluation, t(355) = −6.59, p < .001, (β = −.34), uniquely explaining 10% of the variance. Coaches with higher perceived organisational support experienced less sport devaluation.

Predicting turnover intentions

All assumptions for MLR were assumed; however, six multivariate outliers with Mahalanobis distance greater than the critical value for χ² for df = 10 (α = .001) of 29.58, were identified and removed before interpreting the final model.

In a hierarchical MLR predicting turnover intentions, the demographic variables entered at step 1 failed to explain any variance in turnover intentions, F(5,353) = 2.10, p = .07, R² = .03, adjusted R² = .02. The addition of POS and LOC in step 2 explained a significant 21% of the variance in turnover intentions, ΔF(2,351) = 47.81, p < .001. The three burnout dimensions added in step 3 explained a further 32% of variance in turnover intentions, ΔF(3,348) = 84.86, p < .001.

The final model (Table 5) was significant, F(10,348) = 44.16, p < .001, explaining 56% of the variance in turnover intentions, adjusted R² = .55, representing a very large effect (f = 1.27). POS (β = −.19), was a significant negative predictor of turnover intentions, t(348) = − 4.39, p < 001. The three burnout dimensions all positively predicted turnover intentions; devaluation was the strongest predictor, t(348) = 6.29, p < .001, (β = .36), uniquely explaining 5% of the variance. Locus of control was not a significant predictor of turnover intentions, t(348) = .24, p = 81 (Table 6). Coaches with lower perceived organisational support, lower feelings of accomplishment, higher levels of exhaustion, or higher levels of devaluation, reported greater intentions to leave.

Discussion

Our primary aim was to explore organisational and personal factors contributing to coaches' experiences of burnout. Whilst they were not the main focus of the analyses, results nevertheless revealed that some demographic variables were related to coach burnout. For example, female coaches experienced higher levels of reduced accomplishment and exhaustion than males. This finding parallels gender differences reported using the Maslach Burnout Inventory (MBI).⁵⁴ Age also predicted coaches' levels of exhaustion, with younger coaches more susceptible. A possible reason for this is that over time older coaches may have developed resilience to burnout by building up coping resources necessary to effectively deal with the stressors of coaching.⁵⁴ Survival bias may also be a contributing factor as coaches who burn out early in their career may have already left, leaving the remaining survivors who consequently have lower levels of burnout.³² Coaches who spent more hours per week coaching had significantly higher exhaustion levels. Similarly, American teacher–coaches, whose dual role regularly requires long working hours, reported that it led them to feel overwhelmed and physically and emotionally fatigued.²⁶ Although the above demographic variables were all significant predictors of coach burnout, they only explained 1–6% of variance in burnout scores indicating that other factors were more important contributors.

As hypothesised, coaches with higher perceived organisational support had lower levels of burnout, as evidenced by negative relationships with all three subscales of the CBQ. Organisational support was in fact the strongest negative predictor in each of the regression models. This is congruent with the idea that the greater availability of resources in supportive organisations reduces coaches' vulnerability to resource loss, stress, and burnout,⁸ whereas lack of financial and staffing resources can contribute to burnout.⁵⁵ It also corroborates findings of qualitative studies that reported organisational factors were some of the most salient stressors experienced by coaches.^1,24

Higher internal locus of control was associated with lower levels of reduced accomplishment but was not a significant predictor of exhaustion or devaluation; thus, only partially supporting our initial hypothesis. The demonstrated relationship between locus of control and the reduced accomplishment dimension of burnout may be due to an underlying conceptual likeness, which is not evident between locus of control and the other two burnout-dimensions. Reduced sense of accomplishment assesses coaches' self-evaluations of their coaching skills and achievements, whereas exhaustion and devaluation assess coaches' levels of fatigue and growing detachment from their coaching role.²³ Other research has reported that the reduced accomplishment dimension is more distinct than the other two burnout dimensions and shows a different pattern of relationships with burnout correlates.^1,56

A relationship with self-efficacy is another possible explanation; it has been suggested that self-efficacy, locus of control, self-esteem, and neuroticism form a higher order trait labelled ‘core self-evaluations’, which is negatively related to job stress and strain.⁵⁷ If these constructs do form part of a core self-evaluations trait, it follows that coaches who have an internal locus of control may also possess a stronger self-efficacy in their coaching ability. Therefore, they may be less vulnerable to the erosion of their sense of accomplishment by occasional failures or stressors within the coaching environment⁵⁸; this merits future research.

Partial support was received for the hypothesised relationship of approach coping and burnout. Greater use of approach coping strategies predicted lower levels of reduced accomplishment in coaches. It did not however significantly predict either exhaustion or devaluation, despite being negatively correlated with devaluation; a possible explanation is that stress mediates the relationship between coping and burnout.²⁵ Stress has previously been shown to mediate the effect of several other personal variables like hardiness, anxiety, and perceived support on coach burnout.^59,60 As coaches' perceived stress was not measured in this study, future research could investigate whether a stress-mediated model accounts for the lack of direct relationships between coping and burnout observed in this study.

The coping measure used in this study did not explore the perceived effectiveness of the coping responses used.²⁸ Research indicates that the ability to tailor coping responses to situational requirements is related to greater perceived effectiveness of coping strategies.⁶¹ Future studies could explore coaches' coping flexibility or how they rate the perceived effectiveness of the coping strategies they use.²⁸

Some studies of coach burnout have only examined the exhaustion dimension, and it has been argued that study findings would not differ substantially if only the emotional exhaustion subscale was used.⁶² Our findings – that personal variables such as locus of control, approach coping, and many of the demographic variables were differentially related to the three dimensions of the CBQ – challenge the above statement. Future research should work to further validate the factor structure of the CBQ⁴² and should further examine if specific resources relate differentially to each burnout dimension.

A secondary aim of this study was to explore the link between coach burnout and turnover intentions. Consistent with the hypothesis of this study, and the findings of others,^33,36 higher organisational support was related to lower turnover intentions. As hypothesised, burnt out coaches reported higher turnover intentions. The devaluation dimension of burnout was the most strongly related to turnover intentions, which is unsurprising as it is the dimension that represents coaches' disengagement from coaching.²³ This finding is consistent with the idea that when stress related costs of coaching begin to outweigh the positives aspects, coaches are likely to withdraw and cease their involvement.⁴ This then further supports a multidimensional view on measuring burnout.

However, turnover intentions were measured in this study rather than actual turnover; albeit behavioural intentions are considered a key cognitive antecedent and are reasonably indicative of a person's likelihood of performing a specific behaviour.⁶³ Nevertheless, longitudinal studies could better establish the causal sequencing between burnout and turnover. For example, coach burnout could be measured and organisational records could subsequently be used to follow up coaches who later left the organisation to ascertain if their reason for leaving was related to low organisational support, burnout, or a combination of these and other factors.

Even if organisational and individual factors are considered equally important contributors to burnout, many burnout interventions have focused on changing the individual rather than the organisation.³² Interventions that combine individual training with wider organisation changes have, however, shown initial promise for increasing employees' perceptions of workplace resources, and reducing burnout.^64,65 As organisational support was an important predictor of lower coach burnout in this study, similar interventions could be successful for increasing perceptions of resources and ameliorating burnout in coaches; however, future research is needed to clarify this.²²

Interventions that allow coaches to participate in organisational decision making and enhance open communication with organisational management would help foster a sense of collaboration and perceived organisational support in line with Self-Determination Theory.⁶⁶ This may in turn lead to lower levels of burnout and higher retention rates as stressors experienced by coaches are addressed proactively at the organisational level.^24,33

The strength of this study is that it examined burnout among a large sample of full-time and part-time coaches in both paid and volunteer roles with varying coaching qualifications. This broad sample increases the possibility to generalise more widely compared to a narrower sample; however, it may have also contributed to heterogeneity in the results. Future research could compare burnout experiences of coaches across the different subpopulations, for example, comparing elite versus community level coaches. This would help clarify if the current findings are more relevant within different subsamples. Hjälm et al.⁵⁵ previously suggested that differences in the availability of staffing and financial resources between coaches of professionalised and less professional sports might contribute to coach burnout. This tentatively implies that community coaches may be at higher risk for burnout, but future research is required to confirm whether that is actually the case.

The convenience sample used in this study resulted in over half of the sample being comprised of Australian Rules Football and rowing coaches. It is possible that coaches from these sports were more motivated or interested in the research area than non-participants hence their higher rates of participation. However, it also limits the generalisability of the findings to other sports being less represented within the sample. The possibility of a survival bias should also be noted, whereby those coaches with higher levels of burnout have already left coaching, meaning that true levels of coach burnout may be underestimated.

Finally, the cross-sectional correlational research design and exclusive reliance on self-report measures limits the ability to establish causality among the variables studied. It is possible that organisational support may reduce burnout; alternatively, those with lower burnout may perceive their organisations as more supportive. Longitudinal or quasi-experimental multi-method research designs may help to further establish possible causal links among coach burnout and the personal and organisational factors that are related to it.

Conclusion

Our results suggest that coaches who feel supported by their organisation are less likely to experience burnout; they are also less likely to contemplate leaving coaching. We therefore suggest that sporting organisations should engage in open collaborative conversations with coaches about the organisational stressors they experience, and how these can be addressed. This will likely make coaches feel more supported and involved, thereby increasing the likelihood of retaining them within the organisation long term.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Lundkvist

Gustafsson

Hjälm

. An interpretative phenomenological analysis of burnout and recovery in elite soccer coaches. Qual Res Sport Exerc Health 2012; 4: 400–419. DOI: 10.1080/2159676x.2012.693526.

Raedeke

Warren

Granzyk

. Coaching commitment and turnover: a comparison of current and former coaches. Res Q Exerc Sport 2002; 73: 73–86. DOI: 10.1080/02701367.2002.10608994.

Olusoga

Butt

Maynard

. Stress and coping: a study of world class coaches. J Appl Sport Psychol 2010; 22: 274–293. DOI: 10.1080/10413201003760968.

Smith

. Toward a cognitive-affective model of athletic burnout. J Sport Psychol 1986; 8: 36–50.

O'Connor

Bennie

. The retention of youth sport coaches. Change Transformations Educ 2006; 9: 27–38. Retrieved from http://sydney.edu.au/education_social_work/research/publications/change.shtml (accessed 19 February 2016).

Pines

Burnout: an existential perspective. In: Schaufeli

Maslach

Marek

(eds). Professional burnout: recent developments in theory and research, Washington: Taylor and Francis, 1993, pp. 33–51.

Hobfoll

. Conservation of resources: a new attempt at conceptualizing stress. Am Psychol 1989; 44: 513–524. DOI: 10.1037/0003-066X.44.3.513.

Hobfoll

. The influence of culture, community, and the nested-self in the stress process: advancing conservation of resources theory. Appl Psychol Int Rev 2001; 50: 337–421. DOI: 10.1111/1464-0597.00062.

Hobfoll

Freedy

Conservation of resources: a general stress theory applied to burnout. In: Maslach

Schaufeli

Marek

(eds). Professional burnout: recent developments in theory and research, Washington: Taylor and Francis, 1993, pp. 115–129.

10.

McLean

Mallett

. What motivates the motivators? An examination of sports coaches. Phys Educ Sport Pedagog 2012; 17: 21–35. DOI: 10.1080/17408989.2010.535201.

11.

Read AD. When coaches' burnout: understanding the phenomenon from the perspective of high performance coaches. Master Thesis, University of Ottawa, Ottawa, ON, 2007. Retrieved from http://www.ruor.uottawa.ca/ee\7 (accessed 19 February 2016).

12.

Eisenberger

Huntington

Hutchison

. Perceived organizational support. J Appl Psychol 1986; 71: 500–507. DOI: 10.1037/0021-9010.71.3.500.

13.

Levinson

. Reciprocation: the relationship between man and organization. Admin Sci Q 1965; 9: 370–390. DOI: 10.2307/2391032.

14.

Cropanzano

Howes

Grandey

. The relationship of organizational politics and support to work behaviors, attitudes, and stress. J Organ Behav 1997; 18: 159–180. DOI: 10.1002/(SICI)1099-1379(199703)18:2%3C159::AID-JOB795%3E3.0.CO;2-D.

15.

Thelwell

Weston

Greenlees

. Stressors in elite sport: a coach perspective. J Sports Sci 2008; 26: 905–918. DOI: 10.1080/02640410801885933.

16.

Spector

. Behavior in organizations as a function of employee's locus of control. Psychol Bull 1982; 91: 482–497. DOI: 10.1037/0033-2909.91.3.482.

17.

Rotter

. Generalized expectancies for internal versus external control of reinforcement. Psychol Monogr Gen Appl 1966; 80: 1–27. DOI: 10.1037/h0092976.

18.

Ito

Brotheridge

. Resources, coping strategies, and emotional exhaustion: a conservation of resources perspective. J Vocat Behav 2003; 63: 490–509. DOI: 10.1016/S0001-8791(02)00033-7.

19.

McKnight

Glass

. Perceptions of control, burnout, and depressive symptomatology: a replication and extension. J Consult Clin Psychol 1995; 63: 490–494. DOI: 10.1037/0022-006X.63.3.490.

20.

TWH

Sorensen

Eby

. Locus of control at work: a meta-analysis. J Organ Behav 2006; 27: 1057–1087. DOI: 10.1002/job.416.

21.

Cresswell

Eklund

. The athlete burnout syndrome: possible early signs. J Sci Med Sport 2004; 7: 481–487. DOI: 10.1016/S1440-2440(04)80267-6.

22.

Goodger

Gorely

Lavallee

. Burnout in sport: a systematic review. Sport Psychol 2007; 21: 125–151. Retrieved from www.humankinetics.com/ (accessed 19 February 2016).

23.

Raedeke

. Is athlete burnout more than just stress? A sport commitment perspective. J Sport Exerc Psychol 1997; 19: 396–417. Retrieved from journals.humankinetics.com/jsep (accessed 19 February 2016).

24.

Olusoga

Butt

Hays

. Stress in elite sports coaching: identifying stressors. J Appl Sport Psychol 2009; 21: 442–459. DOI: 10.1080/10413200903222921.

25.

Raedeke

Smith

. Coping resources and athlete burnout: an examination of stress mediated and moderation hypotheses. J Sport Exerc Psychol 2004; 26: 525–541. Retrieved from journals.humankinetics.com/jsep (accessed 19 February 2016).

26.

Drake

Herbert

. Perceptions of occupational stress and strategies for avoiding burnout: case studies of two female teacher-coaches. Phys Educat 2002; 59: 170–183. Retrieved from http://js.sagamorepub.com/ (accessed 19 February 2016).

27.

Frey

. College coaches' experiences with stress – “Problem Solvers” have problems, too. Sport Psychol 2007; 21: 38–57. Retrieved from www.humankinetics.com/ (accessed 19 February 2016).

28.

Levy

Nicholls

Marchant

. Organisational stressors, coping, and coping effectiveness: a longitudinal study with an elite coach. Int J Sports Sci Coach 2009; 4: 31–45. Retrieved from http://www.multi-science.co.uk/sports-science&coaching.htm (accessed 19 February 2016).

29.

Moos

. Development and applications of new measures of life stressors, social resources, and coping responses. Eur J Psychol Assess 1995; 11: 1–13. DOI: 10.1027/1015-5759.11.1.1.

30.

Moos

Brennan

Fondacaro

. Approach and avoidance coping responses among older problem and non-problem drinkers. Psychol Aging 1990; 5: 31–40. DOI: 10.1037/0882-7974.5.1.31.

31.

Kosa

. Teacher-coach burnout and coping strategies. Phys Educat 1990; 47: 153–158. Retrieved from http://js.sagamorepub.com/ (accessed 19 February 2016).

32.

Maslach

Schaufeli

Leiter

. Job burnout. Annu Rev Psychol 2001; 52: 397–422. DOI: 10.1146/annurev.psych.52.1.397.

33.

Surujlal

Grobler

WCJ

. Retaining sport coaches: a qualitative assessment of retention strategies at sport clubs in South Africa. Mediterr J Soc Sci 2014; 5: 651–658. DOI: 10.5901/mjss.2014.v5n2p651.

34.

Goodman A. Retention and attrition factors of female certified athletic trainers in the National Collegiate Athletic Association Division-I Football Bowl Subdivision Setting. Doctoral Dissertation, Retrieved from ProQuest Dissertations & Theses Global database (UMI No. 3296658), 2007.

35.

Scanlan

Still

. Job satisfaction, burnout and turnover intentions in occupational therapists working in mental health. Aust Occup Therap J 2013; 60: 310–318. DOI: 10.1111/1440-1630.12067.

36.

Cunningham

Sagas

. Occupational turnover intent among assistant coaches of women's teams: the role of organizational work experiences. Sex Roles 2003; 49: 185–190. DOI: 10.1023/A:1024469132536.

37.

Allen

Weeks

Moffitt

. Turnover intentions and voluntary turnover: the moderating roles of self-monitoring, locus of control, proactive personality, and risk aversion. J Appl Psychol 2005; 90: 980–990. DOI: 10.1037/0021-9010.90.5.980.

38.

Blau

. Locus of control as a potential moderator of the turnover process. J Occup Psychol 1987; 60: 21–29. DOI: 10.1111/j.2044-8325.1987.tb00238.x.

39.

TWH

Feldman

. Locus of control and organizational embeddedness. J Occup Organ Psychol 2011; 84: 173–190. DOI: 10.1348/096317910X494197.

40.

Raedeke

Smith

. Development and preliminary validation of an athlete burnout measure. J Sport Exerc Psychol 2001; 23: 281–306. Retrieved from journals.humankinetics.com/jsep (accessed 19 February 2016).

41.

Harris BS. Coach and athlete burnout: the role of coaches' decision-making style. Master Thesis, West Virginia University, Morgantown, WV, 2005.

42.

Lundkvist

Stenling

Hassmén

. How to measure coach burnout: an evaluation of three burnout measures. Meas Phys Educ Exerc Sci 2014; 18: 209–226. DOI: 10.1080/1091367X.2014.925455.

43.

Malinauskas

Malinauskiene

Dumiciene

. Burnout and perceived stress among University Coaches in Lithuania. J Occup Health 2010; 52: 302–307.

44.

Spector

. Development of the work locus of control scale. J Occup Psychol 1988; 61: 335–340.

45.

Tong

Wang

. Work locus of control and its relationship to stress perception, related affections, attitudes and behaviours from a domain-specific perspective. Stress Health 2012; 28: 202–210. DOI: 10.1002/smi.1423.

46.

Duffy

Lilly

. Do individual needs moderate the relationships between organizational citizenship behaviour, organizational trust and perceived organizational support? J Behav Appl Manag 2013; 14: 185–197.

47.

Billings

Moos

. The role of coping responses and social resources in attenuating the stress of life events. J Behav Med 1981; 4: 139–157. DOI: 10.1007/BF00844267.

48.

Meyer

Allen

Smith

. Commitment to organizations and occupations: extension and test of a three-component conceptualization. J Appl Psychol 1993; 78: 538–551. DOI: 10.1037/0021-9010.78.4.538.

49.

Allen

Bennett

. SPSS for the health & behavioural sciences, Victoria: Thomson, 2008.

50.

Kagee

Review of the coping responses inventory. In: Plake

Impara

(eds). The Fourteenth Mental Measurements Yearbook, Lincoln: Buros Institute of Mental Measurements, 2001.

51.

Rhoades

Eisenberger

. Perceived organizational support: a review of the literature. J Appl Psychol 2002; 87: 698–714. DOI: 10.1037/0021-9010.87.4.698.

52.

Tabachnick

Fidell

. Using multivariate statistics, Boston: Pearson Education, 2007.

53.

Cohen

. A power primer. Psychol Bull 1992; 112: 155–159. DOI: 10.1037/0033-2909.112.1.155.

54.

Caccese

Mayerberg

. Gender differences in perceived burnout of college coaches. J Sport Psychol 1984; 6: 279–288.

55.

Hjälm

Kenttä

Hassmén

. Burnout among elite soccer coaches. J Sport Behav 2007; 30: 415–427.

56.

Lee

Ashforth

. A meta-analytic examination of the correlates of the three dimensions of job burnout. J Appl Psychol 1996; 81: 123–133. DOI: 10.1037/0021-9010.81.2.123.

57.

Bono

Judge

. Core self-evaluations: a review of the trait and its role in job satisfaction and job performance. Eur J Pers 2003; 17: S5–S18. DOI: 10.1002/per.481.

58.

Bandura

. Self-efficacy: toward a unifying theory of behavioral change. Psychol Rev 1977; 84: 191–215. DOI: 10.1037/0033-295X.84.2.191.

59.

Kelley

. A model of stress and burnout in collegiate coaches: effects of gender and time of season. Res Q Exerc Sport 1994; 65: 48–58. DOI: 10.1080/02701367.1994.10762207.

60.

Kelley

Eklund

Ritter-Taylor

. Stress and burnout among collegiate tennis coaches. J Sport Exerc Psychol 1999; 21: 113–130. Retrieved from journals.humankinetics.com/jsep (accessed 19 February 2016).

61.

Cheng

. Assessing coping flexibility in real-life and laboratory settings: a multi-method approach. J Pers Soc Psychol 2001; 80: 814–833. DOI: 10.1037/0022-3514.80.5.814.

62.

Raedeke

Granzyk

Warren

. Why coaches experience burnout: a commitment perspective. J Sport Exerc Psychol 2000; 22: 85–105. Retrieved from journals.humankinetics.com/jsep (accessed 19 February 2016).

63.

Fishbein

Ajzen

. Predicting and changing behavior, New York: Psychology Press, 2010.

64.

Hätinen

Kinnunen

Pekkonen

. Comparing two burnout interventions: perceived job control mediates decreases in burnout. Int J Stress Manag 2007; 14: 227–248. DOI: 10.1037/1072-5245.14.3.227.

65.

van Dierendonck

Schaufeli

Buunk

. The evaluation of an individual burnout intervention program: the role of inequity and social support. J Appl Psychol 1998; 83: 392–407. DOI: 10.1037/0021-9010.83.3.392.

66.

Deci

Ryan

. The “What” and “Why” of goal pursuits: human needs and the self-determination of behavior. Psychol Inq 2000; 11: 227–268. DOI:10.1207/S15327965PLI1104_01.