Coaching under pressure: An exploration of stress,burnout,and coaching behaviors using self determination theory

Abstract

Youth sport coaches play a critical role in influencing developmental outcomes for athletes, including prolonged engagement in sport and physical activity, life skill development, and resilience. However, due to lack of knowledge, limitations in organizational support, or personal and professional experiences, coaches do not always engage in coaching behaviors that promote this growth and well-being. Within this body of literature, there are gaps in examining how specific coach experiences, like stress and burnout, influence coaching behaviors. Using Self Determination Theory (SDT), the current study utilized a cross-sectional design to explore the relationships among stress, burnout, need-supportive behaviors, and need-thwarting behaviors. Researchers collected survey responses from 353 youth sport coaches across the United States. Data were analyzed using Structural Equation Modeling (SEM). Results (χ² = 1,869.67; CFI = .96; SRMR = .07; RMSEA = .045) revealed that burnout was a stronger predictor of need-supportive and need-thwarting behaviors than stress. Coaches reporting higher levels of burnout also reported less use of need-supportive behaviors (p < .01) and more use of need-thwarting behaviors (p < .01). Stress was not found to be a strong predictor even though most coaches in the sample reported experiencing moderate to high levels of stress. Findings indicate a need to support youth sport coaches by utilizing strategies to reduce rates of coach burnout and support coaches experiencing burnout.

Introduction

Youth sports represent an important developmental context for young athletes. Sport settings can be leveraged to promote healthy developmental outcomes across the life span including positive youth and life skill development,^1–3 physical activity, ² and resilience. ⁴ However, sport entities cite missed opportunities for young athletes to gain these benefits given increasing rates of drop out from sport.^5,6 Researchers have explored factors associated with positive sport experiences that lead to continued participation in sport, finding motivation and enjoyment as key components for preventing drop-out.^7–10 When addressing environments that promote intrinsic motivation and sport enjoyment, researchers often highlight the role of adult social agents in youth sport settings like coaches.^8,11–13 Oftentimes, youth sport coaches serve as precursors to athletes seen later in more competitive systems and, importantly, to athletes who disengage or drop out of sport altogether.

It is well demonstrated that various coaching behaviors, philosophies, and interaction styles influence a myriad of athlete outcomes in sport,^3,14,15 as well as motivational outcomes of athletes such as engagement, continued participation, and enjoyment.^8,9,16,17 Researchers seeking to understand how coaches influence youth outcomes often look to self-determination theory (SDT). Specifically, SDT demonstrates how coaching behaviors designed to meet the needs of autonomy, relatedness, and competence are most effective at promoting athlete motivation.^14,17–19

SDT in sport

SDT proposes motivational orientation toward participation in activities is influenced by social conditions and by whether three basic psychological needs—autonomy, competence, and relatedness—are satisfied.^20,21 Autonomy is best understood as the ability to independently contribute and make decisions, and competence is defined as the ability to master skills, gain knowledge, and apply learnings.^20,21 Finally, relatedness considers an individual's sense of belonging. According to SDT, variations in meeting these three needs predict motivational orientation and are influenced by environmental factors like psychosocial relationships and organizational pressures.^21–23

SDT is a widely used framework in sport contexts. Researchers promote the use of coaching behaviors aligned with the meeting of each of the three psychological needs, or more broadly, need-supportive behaviors.^21–24 Need-supportive strategies include those that facilitate a sense of choice among athletes, promote goal setting and achievement, and foster meaningful relationships.^17,19,21,24 More specifically, Mageau and Vallerand ¹⁷ provide a framework of important autonomy-supportive behaviors, defining seven key coaching behaviors that promote the meeting of athletes’ needs for autonomy, including 1) providing choice within limits, 2) explaining the “why” or rationale for expectations, drills, and decisions, 3) acknowledging others’ emotions and views, 4) providing opportunities for independent contribution and decision-making, 5) giving mastery-oriented feedback, 6) avoiding criticisms, controlling behaviors, and tangible rewards, and 7) preventing ego-driven coaching philosophies. ¹⁷ This framework is supported by scholarship from other sport psychology researchers like Zakrajsek and colleagues ²⁵ that emphasize the need to provide meaningful choices, ask the opinions of others, and provide transparency to meet the needs of not only athletes but other adults within sport settings. By engaging in these types of behaviors, athletes may engage further in their sport, take on leadership opportunities and feel a larger sense of responsibility toward the team, which is likely to result in more self-determined motivation for participation and engagement.^14,22,24,26 Several examples of autonomy-supportive behaviors have been found in qualitative studies including:

Allowing athletes to decide game play or strategies in practice and at scrimmages^27–29

Asking athletes for their opinions and feelings on drills, uniforms, and team activities^28,30

Allowing athletes to reflect on their performances and identify goals and areas for improvement^17,28

Allowing athletes to make decisions, even if they may lead to failure ²⁹

Competence-supportive behaviors also are important for coaches hoping to increase self-determined motivation among athletes. Competence-supportive strategies are closely aligned with a mastery-oriented, or growth, mindset which emphasizes the importance of individualized instruction for skills, knowledge sharing, and creation of opportunities to improve self-efficacy and confidence.^18,25,31,32 Behaviors aligned with this orientation may include encouraging athletes to master sport skills that match their current talent, emphasizing individual improvement over comparison to others, and providing constructive and positive feedback. ¹⁹ Specifically, coaches might use the following competence-supportive strategies cited by researchers:

Acknowledging successes by using verbal feedback or providing a reward like canceling practice after a game where all athletes gave effort^27,28,30

Implementing goal setting at the start of the season through either written responses or verbal discussions^27,28

Celebrating incremental improvements in and out of sports regardless of outcomes^28,29

Reflecting on practices and competitions to identify successes and areas for improvement ²⁸

Designing drills to meet athletes where they are at by starting with drills that allow athletes to start with successes then work through challenges ²⁹

Last, relatedness-supportive behaviors improve athletes’ senses of belonging by demonstrating care and understanding, acknowledging athletes’ feelings, providing opportunities in and out of practice for relationship-building, and fostering trust with athletes.^31,33 Athletes who have quality relationships with coaches and peers often are more motivated to continue playing with their teams and demonstrate dedication.^20,21,34 Examples of relatedness-supportive strategies include:

Organizing team-building activities like team meals or external events^28,29

Asking athletes personal questions about their lives outside of sports (e.g., related to school or family)^28,29

Understanding each athletes’ motivations and individualizing motivational techniques ²⁹

Using humor appropriately to connect with athletes ²⁹

Conversely, coaches who do not engage in these types of behaviors, or use need-thwarting coaching behaviors, can negatively impact athlete outcomes and motivation. Examples of need-thwarting behaviors include emphasizing mistakes or faults, doubting athletes’ abilities, ignoring athletes, using non-active listening skills, fostering feelings of guilt among athletes, and using controlling behaviors.^17,19,31 When need-thwarting occurs, motivational orientations can shift toward controlled or amotivation wherein athletes engage in activities to avoid negative emotions (e.g., avoid punishment) or disengage from sport.^20,21,34 Ultimately these coaching behaviors can result in decreased motivation and in some cases burnout among athletes. ²⁴ Specific need-thwarting behaviors are cited below:

Yelling at athletes when they make mistakes at practices ²⁸

Setting unrealistic expectations for athletes both individually and at team levels ³⁰

Providing too many critiques at one time, being unclear in feedback, and/or delaying feedback (e.g., stating, “You’re not good,” versus, “If you adjust your hand position, you might have a better chance at hitting the ball.”) ³⁵

Providing high-stakes rewards or using physical punishments as behavioral reinforcements ³⁵

Appearing distant from athletes at practices whether physically or through limited communication

SDT research to-date provides support for coaches’ use of autonomy-, competence-, and relatedness-supportive behaviors and avoidance of need-thwarting behaviors. For example, Gagné and colleagues ²² found that youth gymnasts (ages 7–18) who perceived their coaches used need-supporting behaviors also reported increased motivation and continued participation in gymnastics. More recently, Larson and colleagues ²⁶ found that youth competitive swimmers were more likely to report commitment, enjoyment, and intention to continue swimming when they perceived coaches using need-supportive behaviors as opposed to need-thwarting behaviors. Beyond motivation and prolonged engagement, evidence suggests that the use of need-supportive behaviors aligns with increased life skill development,^36,37 positive vitality, ²⁴ self-esteem, ³⁸ and sport performance. ³⁹ The use of need-supportive behaviors has the potential to positively influence a multitude of athlete outcomes in sport and in life.

Though research supports the use of need-supportive behaviors, evidence suggests that coaches do not always adopt these behaviors. Iachini and colleagues ²⁸ found that high school sport coaches often did not intentionally use need-supportive behaviors. High school coaches in this study demonstrated limited understanding of how need-supportive strategies benefitted athletes and struggled to identify and define need-supportive behaviors specific to their coaching practices. Findings also highlighted contextual barriers influencing coaching practices such as structure (e.g., role), type of sport (e.g., individual or team), limited time (e.g., sport season, practice time), and athlete characteristics (e.g., gender, ability to trust athletes). As an example, some coaches were less likely to use relatedness-supportive strategies like organizing team bonding if they had an assistant coach, and individual sport coaches were less likely to use all three types of need-supportive behaviors. ²⁸ Still the exploration of what influences need-supportive coaching behaviors remains a growing priority.

Predictors of need-supportive behaviors

When exploring predictors of need-supportive behaviors, although limited, research has identified factors relevant to three areas, including organizational contexts, personal orientations and perceptions, and psychological needs. In addition to the work of Iachini and colleagues, ²⁸ studies have explored organizational variables that influenced whether coaches used need-supportive or need-thwarting behaviors. For example, Rocchi and Pelletier ⁴⁰ and Stebbings and colleagues ⁴¹ both found that coaches who perceived less opportunities for professional development were less likely to use need-supportive behaviors, whereas coaches who perceived tension between work and life commitments were more likely to use need-thwarting behaviors. Other organizational factors positively influencing coaches’ use of need-supportive behaviors include perceived support from administration, ⁴⁰ job security, ⁴¹ positive motivational climate, ⁴² and less bias in the work environment. ⁴³

Coaches’ own personal orientations have the potential to influence behavior choices. These factors include personality characteristics and perceptions, such as coaches’ egos, values, views of team satisfaction and engagement, levels of certification, and self-efficacy beliefs.^17,28,44,45 As an example, Pulido and colleagues ⁴⁵ found that Spanish coaches who were more satisfied with their teams were more likely to use need-supportive behaviors. Iachini ⁴⁴ suggested that coaching efficacies (i.e., beliefs about impact on athletes) positively and ego-involvements (i.e., athlete performances impacting coach self-confidence) negatively influenced the use of autonomy-supportive behaviors among high school coaches.

Finally, the degree to which coaches use need-supportive coaching behaviors also is due to the meeting of their own personal needs. For example, Rocchi and colleagues ⁴⁶ found that youth basketball coaches were more likely to use autonomy-supportive behaviors when they perceived their own needs for autonomy in decision-making were met. Another study found that coaches in the United Kingdom who reported higher satisfaction of competence and autonomy also were more likely to use autonomy-supportive behaviors. ⁴⁷ Given the relationships among need-supportive coaching behaviors and athletes' motivations and outcomes, understanding other factors and psychological experiences that influence coach behaviors remains a priority.

Stress and burnout

One factor which may greatly impact coaching behaviors, including need-supportive ones, is stress. Stress is best understood as a depletion in psychological resources which often leads to less investment and empathy toward others. ¹⁷ According to Smith's ⁴⁸ Cognitive-Affective Stress Model (CASM) that has since been adapted by Gould, ⁴⁹ stress involves a series of cognitive, physiological, and behavioral responses to situational demands and resources within environments. As individuals evaluate demands and resources to meet demands, the use of task-oriented behaviors occurs. These behaviors can be both adaptive and maladaptive, meaning that stress does not always lead to engagement in appropriate behavioral responses.^48,49 Over time, Gould ⁴⁹ contends that chronic experiences of stress can progress toward burnout. Burnout is characterized by several stages: (a) negative cognitive appraisals of demands, resources, and accomplishments, (b) physiological responses such as feelings of anxiety or depression, insomnia, fatigue, and illness, and (c) behavioral changes such as disengagement, interpersonal challenges, and decreased performance. ⁴⁹ While this model was originally designed for athletes, it can be applied to coaches given connections to research on burnout among professionals.

Among professionals, burnout has been described as a psychological response to chronic stress characterized by emotional exhaustion, reduced personal accomplishment, and depersonalization. ⁵⁰ Research from Maslach and Jackson, ⁵⁰ among others, posits that individuals in helping professions, especially demanding ones, are at higher risk for burnout as a result of increased focus on the needs and feelings of others.^51,52 Certainly coaching is a demanding profession wherein coaches are often expected to go beyond teaching sport skills and tactics.^1,53–55 Even further, youth sport coaches are often required to navigate complex athletic systems, work long hours, adopt multiple roles (e.g., administrative, teaching, fundraising), and adapt to ever-changing environments often with little to no pay.^54–58 Identified stressors such as challenges with athletes, feelings of isolation, work–life balance difficulties, pressures to win, and negative interactions with parents and other stakeholders are commonplace. ⁵⁹ Coaches also are humans who may face stressors in their personal lives, and can be exacerbated by the demands of their jobs.

No doubt navigating these challenges, multiple demands, and resultant stress areinherent parts of coaching,^60,61 leaving coaches physically and emotionally exhausted. Recent data suggests coach stress is at an all-time high, with 77% of youth sport coaches in the United States National Coach Survey reporting moderate to extreme stress associated with their coaching jobs.^54,62 Over time, experiences of stress can lead to burnout. ⁵⁰ Feelings of being drained, numb, and distant from work and activities may result. ⁵⁰ Further, reduced senses of accomplishments accompanying burnout typically are coupled with lower self-efficacies, which also can lead to cycles of de-investment and decreased satisfaction in coaching. As these cognitive and psychological processes occur, coaches are at risk for adopting rigid, inappropriate behaviors such as yelling, reducing personal choices, or disengaging from athletes.^48–51

To date however, explorations of the relationships among stress, its counterpart burnout, and coaching behaviors have been limited among youth sport coach populations, although scholars have suggested the importance. Researchers such as Mageau and Vallerand ¹⁷ and Iachini ²⁸ propose that coaches experiencing stress related to coaching, general stress, and burnout would have less psychological resources and energy to engage in behaviors aligned with SDT (although not the central foci for their studies). Additionally, there is some evidence supporting the role of stress and burnout in relation to other types of coaching behaviors. For example, one study revealed that high school coaches who reported experiencing emotional exhaustion were less likely to focus on life skill development, provision of structure within the team, instruction of skills and tactics, and relationship-building in their coaching practices. ⁶³ Additionally, Vealey and colleagues ⁶⁴ found that burnout negatively influenced positive coaching behaviors like demonstrating empathy and providing positive praise among college sport coaches. Another study by Alahayneh ⁶⁵ found college coaches with higher perceived emotional exhaustion and lower senses of personal accomplishment were less likely to provide positive feedback, social support, and democratic decision-making. Indeed coaches are not immune to stress and burnout, which could impact how they show up on the field, court, or mat.

To the researchers’ knowledge, no research exists concurrently exploring whether stress and burnout influence need-supportive and need-thwarting behaviors among youth sport coaches. Iachini ⁴⁴ did test a conceptual model of predictors of autonomy-supportive coaching behaviors that included stress among high school coaches, but results suggested stress was not a significant predictor of autonomy-supportive behaviors. This finding opposes theoretical underpinnings of stress and burnout which suggest that stress and burnout both influence behavioral outcomes, concurrently and independently,^49–51 as well as research on other coaching behaviors which suggest that aspects of burnout may be significant predictors of coaching behaviors.^63–65 These inconsistencies warrant additional explorations of stress, burnout, and need-supportive and need-thwarting behaviors. What's more, this type of inquiry is needed among youth sport coaches given the role they play in not only influencing sport skill acquisition, but broader developmental outcomes such as mental well-being, physical activity engagement, and life skill development.

To address these gaps, the primary purpose of this exploratory study wasto examine relationships among coach stress, burnout, need-supportive behaviors, and need-thwarting behaviors among youth sport coaches. Researchers used structural equation modeling (SEM) to explore the relationships among coach stress, burnout, need-supportive behaviors, and need-thwarting behaviors. One key research question guided the study: (a) What are the relationships among stress, burnout, need-supportive behaviors, and need-thwarting behaviors among youth sport coaches? Specifically, researchers hypothesized that coaches experiencing lower levels of stress (both in general and sport-specific) and coach burnout would be more likely to utilize need-supportive behaviors and less likely to use need-thwarting behaviors. Alternatively, we proposed coaches experiencing higher levels of these two types of stress and coach burnout would be more likely to utilize need-thwarting behaviors and less likely to use need-supportive behaviors. A SEM model was tested to explore these relationships as shown in Figure 1. Understanding these relationships can help sport organizers, coaches and other stakeholders to further promote positive coaching behaviors predictive of motivation and other athlete outcomes.

Figure 1.

Hypothesized model.

Methods

Recruitment and procedures

All study procedures were approved by the authors’ Institutional Review Board of record. Researchers began by developing a comprehensive list of sport entities with connections to coaches over 18 years old. This list included contacts of sport leagues, coach associations, high school or school districts, professional associations, and non-profits and for-profit sport clubs used in previous research.^5,54,66 Recruitment emails were then sent to sport administrators and leaders in these sport entities asking them to disseminate a recruitment script to coaches in their networks via email communications, newsletters, social media, and media outreach activities. Coaches interested in the study were asked to indicate whether they were over 18 years old. Those who met inclusion criteria reviewed online consent language and completed a survey via Qualtrics. Surveys took approximately 25–30 min to complete, and no incentives were provided as a part of this study.

Sample

A total of 415 youth sport coaches consented to participate and began the online survey. Coaches completing less than 30% of the survey (following guidance from Newman) ⁶⁷ and/or those coaching collegiate teams were removed, leaving 353 responses in the final analytic sample. Table 1 overviews the demographic characteristics of the sample. Most participants identified as male (66.3%), White (90.9%), reported coaching a team sport (82.7%), coached in a school setting (77.9%), and coached male athletes (43.6%).

Table 1.

Coach demographics (N = 353).

Demographics	% (n)
Coach Demographics
Gender
Male	66.3 (224)
Female	33.7 (114)
Age of Coach in Years
20–29	14.4 (49)
30–39	23.2 (79)
40–49	37.1 (126)
50–59	15.9 (54)
60+	9.4 (32)
Race
White	90.9 (308)
Black/African American	2.7 (9)
Asian/Pacific Islander	1.5 (5)
Multiracial	1.5 (5)
Hispanic/Latinx	0.9 (3)
Other	0.9 (3)
Native American or American Indian	0.3 (1)
Sport Context
Years Coached
Less than 10	42.2 (145)
11–20	32.2 (116)
21+	25.7 (91)
Primary Sport Type
Team	82.7 (286)
Individual	17.3 (60)
Gender of Athletes of Primary Team
Male	43.6 (154)
Female	38.2 (135)
Equally Male and Female	11.0 (39)
Co-ed	7.1 (25)
Primary Sport Setting
School	77.9 (275)
Competitive	15.9 (56)
Recreational	3.1 (11)
Developmental	1.7 (6)
Other	1.4 (5)

Measures

Independent variables

General and Coaching-Related Stress. General stress was measured using the Perceived Stress Scale-10 (PSS-10).^68–70 Coaches were asked how often in the last month they experienced 10 items descriptive of feelings of stress on a 5-point scale. Response options were on a 5-point Likert-type scale ranging from Never to Very Often with higher scores indicating higher levels of stress. Authors recommended creating an index variable to represent stress, with scores ranging from 0–13 implying low stress, scores between 14–26 implying moderate stress, and scores 27 and above implying high stress.^68–70 Example items included “how often have you felt nervous and stressed?” and “how often have you found that you could not cope with all the things that you had to do?” Consistent with previous studies, items from this scale demonstrated strong reliability with a Cronbach's alpha of 0.86 (Cronbach's αs were between .78 and .91). ⁶⁸

While the PSS-10 captured general stress that may be related to challenges within coaching or life, researchers wanted to capture a variable related to sport-specific stress. Given no coach-specific stress scale exists, an additional stress variable used in previous research asked coaches to rate “how stressful is the coaching experience?” on a 5-point Likert-type scale. ⁵⁴ Response options ranged from Not Stressful at all to Extremely Stressful.

Coach Burnout. Burnout was measured using The Coach Burnout Questionnaire (CBQ), ⁷¹ an adapted version of the Athlete Burnout Questionnaire (ABQ). ⁷² This tool was originally adapted from the ABQ to measure burnout among a sample of swimming coaches but has since been used to measure broader coaching contexts.^{41,71,73–76} Given the sample for this study, items were adapted from the original to reflect a broader coaching context. The adapted 15-item measure assessesed three sub-scales of burnout (i.e., devaluation, decreased sense of accomplishment, and emotional exhaustion) by asking coaches to rate how often they felt certain ways while coaching (e.g., “I feel successful at coaching.”; “I feel overly tired from coaching.”) on a 5-point scale. Response options ranged from Almost Never to Almost Always, and two items were reverse coded per measure guidelines. Comparable to prior studies, the adapted measure demonstrated strong reliability with a Cronbach's alpha of 0.89.^41,71,75

Dependent variables

Need-Supportive and Need-Thwarting Coaching Behaviors. The Interpersonal Behaviors Questionnaire-Self (IBQ-Self) ⁴⁰ was used to assess both need-supportive and need-thwarting behaviors. The IBQ-Self consisted of 24-items representing six factors (i.e., autonomy support, autonomy thwart, competence support, competence thwart, relatedness support, relatedness thwart) and was adapted to a coaching context by modifying the stem to state, “When I am with my student-athletes, I…” Coaches were asked to rate how much they agreed with each of the 24 items on a 7-point scale. Response options ranged from Do Not Agree at All to Completely Agree. Example items included “when I am with my student-athletes I support their decisions” and “when I am with my student-athletes I encourage them to improve their skills”. Internal reliability was demonstrated for each of the six factors and overall (Cronbach's αs > .75), consistent with previous studies examining samples of students, student-athletes, and coaches.^40,77,78 Following measure guidelines, the 12 items measuring autonomy-, competence-, and relatedness-thwarting were used to represent need-thwarting behaviors, and the 12 items measuring autonomy-, competence- and relatedness-support were used to represent need-supportive behaviors. ⁴⁰

Analytic approach

Before conducting SEM analyses, descriptive statistics and bivariate correlations examined trends among study variables. Data first were exported to and analyzed in SPSS (IBM SPSS Data Editor for Windows, Version 28.0. Armonk, NY: IBM Corp). The first step was to clean the data and remove cases missing over 30% of variables in the dataset. ⁶⁷ Then, missing data for key variables were examined in the dataset. Overall, missing data for key variables ranged from 3.4% to 17.3%. Results from Little's ⁷⁹ Missing Completely at Random (MCAR) test indicated data were MCAR (p = .73).

After addressing missing variables, descriptive statistics explored coaches’ reports of general stress (stress in the last month), coaching-related stress, burnout, need-supportive behaviors, and need-thwarting behaviors. Secondly, bivariate correlations explored relationships among latent and observed variables to assist with model specification. Then SEM using Mplus 8.4 was used to examine models in line with previous social science and coaching literature.^14,80 As recommended by Kline ⁸¹ the following procedures were performed to assess models: model specification, model identification, estimation, evaluation of fit, and model modifications. In the model specification phase, a model was specified based on SDT and tested for meeting necessary requirements for identification. ⁸¹ Specifically, the researchers confirmed the model had at least 1 degree of freedom (df > 1000) and assigned reliable and valid scales to each of the four latent variables. ⁸¹ For model estimation, weighted least squares means and variables (WLSMV) estimator was appropriate for use given the types of variables in the model and sample size.^81,82 When using the WLSMV estimator, Mplus used pairwise deletion to address missing data and chi-squared statistics for difference testing.^81,83 Multiple fit indices were used to evaluate the adequacy of the measurement and structural models, including the significance of χ², the root mean square error of approximation (RMSEA), the standardized root mean squared residual (SRMR), and the comparative fit index (CFI). Criteria indicating good fit of the model to the data were based on Hu and Bentler ⁸⁴ and included a non-significant (p < .05) χ² value. In addition, a RMSEA < 0.05, a SRMR < 0.08, and a CFI > 0.95 indicated acceptable fit. ⁸⁴

Measurement model

The measurement portions of the models were examined using confirmatory factor analyses. ⁸¹ The measurement model included four latent factors represented by 49 observed variables. Fit indices, factor loadings, and measurement error terms were examined for each latent factor including general stress, burnout, need-supportive behaviors, and need-thwarting behaviors. In addition to fit indices, factor loadings above .40 were deemed acceptable to include in the final measurement model. ⁸⁵

Structural model

After assessing the measurement model, the next step included specifying the structural model with the addition of one observed structural variable for coaching-related stress. Directional and non-directional relationships for endogenous latent variables were specified. The specified model tested whether general stress, coaching-related stress, and burnout predicted need-supportive and need-thwarting behaviors. Using criteria from Hu and Bentler ⁸⁴ to assess model fit, the model was specified and then modified until acceptable fit of the model was reached. Modification indices and results from bivariate correlations were used to make decisions for modifications to the model. ⁸¹

Results

Descriptive statistics

Table 2 presents the descriptive statistics and correlations for key variables in the study. Overall, 96.5% of coaches in the sample reported moderate or high levels of perceived general stress (M = 24.87; SD = 6.22). In the sample, 80.4% of coaches reported that coaching was somewhat or extremely stressful (M = 3.81; SD = .86). In terms of burnout, the mean scale scores indicated low to moderate levels of burnout across the entire sample (M = 34.41; SD = 11.28; scores on the measure ranged from 0 to 75). Coaches in the sample also reported high use of need-supportive behaviors (M = 73.12; SD = 7.45), and mean scores for need-thwarting behaviors indicated low to moderate use of need-thwarting behaviors (M = 26.24; SD = 8.81).

Table 2.

Descriptive statistics.

	Overall mean (SD)	Range
Need-Supportive Behaviors	73.12 (7.45)	0–84
Need-Thwarting Behaviors	26.24 (8.81)	0–84
Burnout	34.41 (11.28)	0–75
General Stress	24.87 (6.22)	0–50
Coaching-Related Stress	3.81 (.86)	1–5

Bivariate correlations examined relationships among study variables as presented in Table 3. Moderate, statistically significant positive correlations were found among burnout and need-thwarting behaviors (p < .01), general stress (p < .01), and coaching-related stress (p < .01). The two stress variables also were moderately, and statistically significantly positively correlated (p < .01). A moderate, statistically significant negative correlation was found between need-supportive behaviors and need-thwarting behaviors (p < .01). Similarly, weak, statistically significant negative correlations were found between need-supportive behaviors and burnout (p < .01) and general stress (p < .01). Finally, coaching-related stress was not statistically correlated with need-supportive behaviors (p = .49), and need-thwarting behaviors (p = .22).

Table 3.

Correlations among study variables.

Variable	Need-Supportive behaviors	Need-Thwarting behaviors	Burnout	Coaching-Related stress	General stress
Need-Supportive Behaviors	–
Need-Thwarting Behaviors	−.41**	–
Burnout	−.27**	.31**	–
Coaching-Related Stress	−0.04	.08	.35**	–
General Stress	−.15**	.07	.54**	.35**	–
Mean(std)	73.12(7.45)	26.24(8.61)	34.41(11.28)	3.81(.86)	24.76(6.23)

Note. *Significance is found at the p < .05 level

**Significance is found at the p < .01 level

Measurement model

The initial measurement model demonstrated adequate fit. Table 4 shows model fit indices statistics with RMSEA = 0.045, SRMR = .07, and CFI = .96. Notably, the χ² value was significant (p < .05), as is often the case with large sample sizes. ⁸¹ Additionally, all factor loadings among latent factors were statistically significant, with values above (see Table 5). ⁴⁰

Table 4.

Measurement model fit indices (N = 339).

Model	χ2	p	df	RMSEA	SRMR	CFI
Measurement Model	1869.67	.00	1112	.045	.07	.96

Note. χ² = chi-square value; DF = degrees of freedom; RMSEA = root mean square error of approximation; SRMR = standardized root mean squared residual; CFI = comparative fit index.

Table 5.

Standardized factor loadings and reliability indices from the measurement model.

Variable	Factor loading	$R^2$
General stress
PSS1	.71	.50
PSS2	.72	.52
PSS3	.72	.52
PSS4	.49	.24
PSS5	.75	.57
PSS6	.69	.47
PSS7	.69	.47
PSS8	.54	.29
PSS9	.77	.60
PSS10	.60	.37
Burnout
Accomplishment (5 items)	.90	.81
Emotional exhaustion (5 items)	.84	.71
Devaluation (5 items)	.90	.81
Need-supportive behaviors
Autonomy-supportive behaviors (4 items)	.42	.18
Competence-supportive behaviors (4 items)	.80	.64
Relatedness-supportive behaviors (4 items)	.86	.74
Need-thwarting behaviors
Autonomy-thwarting behaviors (4 items)	.49	.24
Competence-thwarting behaviors (4 items)	.78	.61
Relatedness-thwarting behaviors (4 items)	.97	.94

Note. All values were significant at p < .001.

Structural model

The initial structural model (Model 1) examined relationships among general stress, coaching-related stress, burnout, need-supportive behaviors, and need-thwarting behaviors. The hypothesized structural model did not indicate adequate fit (SRMR > .08; RMSEA = .05), so using initial modification indices and effect estimates, coaching-related stress was removed from the model given modification indices ranged between 1.57 and 149.92 and the effects of coaching related stress on need-supportive and need-thwarting behaviors were low (standardized coefficients were 0.037 and 0.042). ⁸¹ Researchers re-ran the model after removing coaching-related stress which improved fit indices to meet a priori criteria. Table 6 demonstrates adequate fit statistics of the modified model. The RMSEA point estimate for the final model (Model 2) was .045, SRMR was .07, and the CFI was .96 (χ² = 1869.7; p < .01).

Table 6.

Structural model fit indices (N=339).

Model	χ2	p	df	RMSEA	SRMR	CFI
Model 1	2,113.11	.00	1,159	.049	.10	.95
Model 2	1,869.67	.00	1,112	.045	.07	.96

Note. χ² = chi-square value; DF = degrees of freedom; RMSEA = root mean square error of approximation; SRMR = standardized root mean squared residual; CFI = comparative fit index.

Controlling for the effects of general stress, results indicated that burnout (p < .01) directly and negatively predicted need-supportive behaviors. In other words, coaches who reported higher levels of burnout also reported less use of need-supportive behaviors. Conversely, burnout (p < .01), directly and positively predicted use of need-thwarting behaviors.

General stress (p = .33) did not significantly predict reported use of need-supportive behaviors. However, reported levels of general stress in the last month (p < .01) did negatively predict need-thwarting behaviors. Coaches in this study who reported higher levels of general stress were less likely to report use of need-thwarting behaviors. The standardized parameter estimates and variances for general stress and burnout are reported in Figure 2.

Figure 2.

Final model.

Discussion

In examining relationships among stress, burnout, and coaching behaviors, researchers hypothesized that lower levels of stress and burnout among youth sport coaches would relate to increased reporting of need-supportive behaviors and less reporting of need-thwarting behaviors. Results provided some support for the hypotheses. In particular, findings indicated burnout strongly predicted both need-supportive and need-thwarting behaviors in the model. In other words, coaches in the study who reported lower levels of burnout also reported lower use of need-thwarting behaviors and higher use of need-supportive behaviors. This relationship was consistent with previous literature that suggested coaches reporting less burnout use more positive coaching behaviors such as showing empathy and providing positive feedback/praise.^63,64 Results extend these findings to suggest that burnout may be an important factor across a broad range of positive and negative coaching behaviors, including ones that are known to promote motivation and athlete outcomes. More specifically, this study explored behaviors aligned with SDT like providing freedom in choices, supporting athletes’ decisions, encouraging goal setting and achievement, and taking the time to get to know athletes.

Interestingly, the hypothesis that stress directly related to need-supportive and need-thwarting behaviors was less supported. Results demonstrated that coaching-related stress was not related to coaching behaviors, potentially suggesting pressures and challenges associated with “the game” and all that goes into coaching do not necessarily influence coaching behaviors. While this result may have been influenced by using a single-item scale, previous sport scholarship suggested using single-item scales to assess coach behaviors could be sufficient.^86–88 On the other hand, single-item instruments could be insufficient for capturing broader concepts and reduce the ability to reach desired power and variability in responses.^81,86,89 Certainly, this finding suggests that additional research and the development and validation of sport-specific stress measures may be warranted.

More broadly, results related to general stress suggested negative correlations with need-supportive behaviors, yet the structural model indicated general stress only directly, negatively predicted need-thwarting behaviors. Prior research has demonstrated similar inconsistencies in results. For example, Iachini ⁴⁴ found stress did not significantly predict autonomy-supportive coaching behaviors, while other scholars suggested stress negatively predicted autonomy-supportive behaviors and positively predicted autonomy-thwarting behaviors.^17,63 One reason this pattern exists may be related to theoretical conceptualizations of need-supportive and need-thwarting behaviors. Within SDT, these two concepts are often posited as opposite ends of a behavioral spectrum with more favorability placed on the use of need-supportive rather than need-thwarting behaviors.^17,20,21 However, instrumentation treats need-supportive and need-thwarting behaviors as independent concepts represented by unique items.^40,78 If these two concepts are considered independent of one another, it may be more reasonable to assume independent effects of predictors on each type of behavior. In other words, it would not necessarily be correct to assume that if one variable positively predicts need-supportive behaviors that it must also negatively predict need-thwarting behaviors.

Another potential explanation for these results may lie in the nature of the relationship between stress and burnout. Prior sport literature examines coach stress and burnout using CASM.^48,49 Aligned with this conceptualizations, burnout is seen as an outcome of prolonged exposure to stress.^48,49,64,90 In other words, burnout results from prolonged exposure to stress and is experienced at a higher level of intensity. Given burnout may be a more enduring experience of stress, traditional appraisal and coping mechanisms may not be able to mitigate the effects of burnout. ⁶⁴ In this way, coaches who experience burnout are at increased risk to use maladaptive coaching strategies such as seeking control and communicating less as a response to stressors, indicating burnout as a primary predictor of coaching behaviors rather than stress. While it is likely that higher levels of stress also would be associated with higher levels of burnout and need-thwarting behaviors, it is possible that the duration and intensity of burnout more strongly affects behavior. Future research should explore these dimensions of both stress and burnout, while practitioners may consider proactive approaches to improve coping strategies for navigating stressors to ultimately reduce burnout.

Given these relationships, sport organizations should consider policies and practices to support coaches in coping with general stressors which may lead to burnout. Entities such as the United States Olympic and Paralympic Committee ⁵⁵ have efforts underway already as identified in its Quality Coaching Framework. General practices may include creating environments that promote adaptive coping, support, and less stigmatized environments for wellness. Specifically, sport entities may consider how to improve work-life balance for coaches, insert autonomous decision-making over teams, provide opportunities for professional development, increase access to resources, and increase peer support.^41,47 By providing for coaches’ physical and psychological needs, coaches may have increased supports when experiencing stress and assets to help navigate stressors. In turn, coaches’ risks for prolonged stress leading to burnout could be mitigated.

In addition to creating environments that promote, rather than hinder, coach well-being, sport practitioners can consider the use of one of the most effective prevention tools: education. Prior coaching literature suggested coaches lack knowledge about stress, stress management, and general well-being. ⁹¹ In preparing coaches for their roles, sport organizations require coaches complete trainings with information and tools to support athletes’ physical and emotional health and safety, but these trainings often have little to no focus on how coaches can support themselves. ⁹² Furthermore, current coach training has been criticized for limited information and a lack of meaningful and informative engagement^92,93 even though coaches express interest in practical education and mentorship.^93,94 Perhaps the limited exposure to information about managing stress and preventing burnout contributes to current trends of stress and burnout in the coaching profession. ⁶¹ This may be especially important as not only can burnout impact how coaches interact with athletes but also contributes to the current concerns in recruiting and maintaining quality coaches. ⁶¹

Ultimately, in supporting coaches, sport practitioners, researchers and other stakeholders can help promote motivation and ultimately better outcomes for athletes participating in sports. Prior research emphasizes the importance of coach-athlete relationships and need-supportive behaviors for promoting motivation, enjoyment, and engagement.^3,14,15,26 Therefore, by supporting coaches to prevent experiences of burnout, athletes’ outcomes may be indirectly, positively affected. Future research may continue to extend these ideas by exploring the relationships among burnout, behaviors, and athletes’ outcomes (e.g., vitality, motivation, engagement, etc.).

Limitations and future directions

In addition to potential measurement concerns, study findings should be interpreted with attention to several limitations. First, this study was cross-sectional in design, meaning that data were collected at one time via self-report surveys. Therefore, casual assumptions about relationships among stress, burnout, and need-supportive and -thwarting behaviors cannot be made. Further, self-report measures increased the risk of bias via social desirability. Past research highlighted coaches’ propensities to more favorably perceive their own behaviors as compared to observational data and athletes’ reports of coaching behaviors.^95–98 This calls into question coaches’, including those in this sample, effectiveness in assessing their own behaviors. Aligned with SDT, Iachini and colleagues ²⁸ found that coaches struggled to identify and define coaching strategies aligned with SDT, illuminating potential concerns of coaches’ abilities to accurately report these types of behaviors in this study. To address these limitations, future research might employ longitudinal designs, observational designs, mixed methods, or reports from additional stakeholders to assess predictors, coaching behaviors, and other outcomes. This may look like the addition of athlete reports or observations of coaches across a season.

A second limitation lied in the sample characteristics. In particular, the sample for the study represented a largely homogenous population and majority White, male, team sport coaches. The sample also was limited to youth sport coaches, most of whom coached in a school-based setting. Recruitment methods, including snowball sampling, may have influenced the elevation of a singular population. Not only does this limit the generalizability of results but could also lead to bias in results only representing the experiences limited populations of coaches. This lack of variability in responses to key variables also limits the generalizability of findings. More specifically, coaches reported similar levels of stress across the sample, yet small individual differences did exist. For example, BIPOC coaches reported lower levels of stress as compared to White coaches in the sample, with many coming from school settings coaching female teams. This points to a potential future area of inquiry by also exploring relationships among coaches’ demographic characteristics (e.g., competitive level, race/ethnicity), other experiences (e.g., organizational support, professional development), and coaching behaviors to better understand how various experiences and identities affect coaching behaviors.

A final limitation potentially contributing to inconsistencies in results may have lied in insufficient measurement tools. In this study, only one item was used to assess coaching-related stress, while general stress was measured broadly without considering contextual factors. Iachini ⁴⁴ provided a similar critique in her study, suggesting contradictory results may have been the result of measuring general, rather than sport-specific stress. Measurement tools such as the PSS-10 are designed to capture stress among general populations without capturing unique pressures different professions may face. Therefore, sport-specific stress may be a more accurate representation of stress that would influence coaching behaviors, yet no measurement tool exists to measure coaching-related stress at this time. ⁴⁴ Future research may consider developing and validating coach-specific measures of stress considering the unique pressures coaches face in funding, professional development, and resources. ⁵⁵ Understanding the factors contributing to general stress among coaches might also shed light on other interventions needed to improve coaching behaviors.

Limitations within this area of research, as highlighted in this study, point to several future directions for consideration. Specifically, future research could include more nuanced examinations of how and when coach experiences disrupt need-supportive behaviors and extend models to include athlete outcomes. For example, future researchers may consider examinations of relationships among sub-scales, such as the three domains of burnout, need-supportive behaviors, and need-thwarting behaviors. Related, researchers might explore which factors of burnout (e.g., source, intensity, and/or duration) predict coaching behaviors. Adding these mechanisms may better help sport researchers and practitioners understand which facets of burnout most directly affect coaching behaviors.

Finally, an important consideration for future research is the expansion of models to predictors, coaching behaviors, and athlete outcomes across multiple sport settings. While this study and previous research imply a path of influence from coaches’ experiences and identities on coaching behaviors and ultimately on athlete outcomes, this relationship has not been confirmed by researchers. This type of inquiry would greatly benefit this area of research given the implications for improving sport settings to better promote athlete outcomes.

Conclusion

Overall, this study provides support that higher levels of burnout among coaches can correlate with less need-supportive behaviors and more need-thwarting behaviors. These findings are important considering the influence of need-supportive and need-thwarting behaviors on athletes’ outcomes. When coaches use need-supportive behaviors, research suggests that athletes report higher confidence, less burnout, increased motivation, more engagement, and prolonged participation.^14,22,24,26 Therefore, opportunities exist to continue to explore and support experiences of burnout to both prevent and intervene to support coaches. These can include general practices like creating environments that promote wellness, increasing access to resources, and providing social support.^41,47,55 Targeted interventions also can increase coaches’ knowledge about stress and coping. ⁹¹ Overall, findings indicate coaches also need support in relation to dealing with general stress and preventing burnout to be better equipped to support athletes on and off the field.