Should team coaches care about justice? Perceived justice mediates the relation between coaches’ autonomy support and athletes’ satisfaction and self-rated progression

Abstract

In contrast to the many studies in the business setting, few sport scientists have studied the predictors of athletes’ perception of justice. The present research aimed to determine the role of the interpersonal style of coaches (autonomy support) in enhancing perceived justice and its impact on satisfaction and performance. More precisely, two independent studies examined whether perceived justice mediated the relation between autonomy support on the one hand and athletes’ satisfaction and their self-rated progression on the other hand. In Study 1, female team dancers (N = 145) completed questionnaires assessing autonomy support, justice, and satisfaction. In Study 2, male handball players (N = 102) completed the same questionnaire extended with self-rated progression. Structural equation modeling revealed a positive link between athletes’ perceived autonomy support of the coach and perceived justice. Subsequently, procedural-interactional justice predicted higher levels of satisfaction and self-rated progression, while a significant indirect effect was found from autonomy support to satisfaction. Measurement and path invariance tests showed that the measurement scales and structural models were invariant across both samples. These findings signify the importance of autonomy support and perceived justice to optimize team athletes’ satisfaction and consequently their progression.

Keywords

Dance,equity theory,handball,leadership,organisational behaviour,self-determination theory

Organizational justice was originally conceptualized “to describe and explain the role of justice as a consideration in the workplace.”¹ Due to its crucial influence on employees’ work attitudes and performance, it has become the focus of a vast amount of research in the field of applied psychology.² However, research on organizational justice in the sport context has been limited almost exclusively to intercollegiate and high school sports in the United States.^3,4 Despite the demand for justice research in a club sport context,³ very few studies have been exploring the impact of justice on athletes’ attitudes and behavior.^5,6 These studies demonstrated that athletes’ perceived justice of the coach increased team identification and team cohesion and decreased social loafing in elite sport teams.⁶ Considering the impact of team cohesion and social loafing on performance, it can be assumed that perceived justice of the coach plays an important role in optimizing team athletes’ performance.^7,8

However, research remained silent regarding the role of coaches’ leadership in enhancing athletes’ perceived justice. Even in the business context, leadership and justice theories have evolved largely independently.⁹ In a recent meta-analysis, Karam and colleagues⁹ underscored the need for greater integration of both literatures with a focus on (a) articulating the specific behaviors of a fair leader and (b) examining the impact of a fair leader on employee outcomes. Therefore, the main aims of this study were to examine (a) the role of coaches’ leadership style in predicting athletes’ perceived justice and (b) the role of athletes’ perceived justice as the mediating mechanism in the relation between coaches’ leadership style and athletes’ satisfaction and progression.

Sport scientists have mostly focused on the consequences of two dimensions of organizational justice (i.e., procedural and distributive). Procedural justice refers to the fairness of the processes by which outcomes were determined.² An example of the application of a fair procedure in sport teams is the use of objective information and criteria (e.g., scouting data) to select the starting team. By contrast, distributive justice focuses on the fairness of outcome distributions.² A typical outcome in a team sport context is the playing time of an athlete. The third dimension of justice, interactional justice, has so far been underexposed in the sport setting.³ Interactional justice pertains to the human side of organizational practices. It refers to the interpersonal treatment people receive when procedures are implemented and the explanations about why procedures were used in a specific manner or why outcomes were distributed in a certain fashion.² For example, while procedural justice focuses on the actual use of objective criteria, interactional justice in sport teams highlights the fact that all players are respectfully informed which criteria will be used to select the starting team. Although procedural and interactional justice are theoretically and conceptually considered as distinct dimensions, psychometric analyses have shown that both concepts are highly correlated in sport settings.^3,10 Kim and Andrew^3,10 were the first to examine the independent effects of the three justice dimensions in a sport setting. However, due to a lack of discriminant validity they had to abandon the theoretical three-factor model of organizational justice and combined the concepts of procedural and interactional justice into one factor.^3,10 These results indicate that in sport settings it is not yet confirmed whether the organizational justice is best represented by a three-factor or a two-factor model.

Autonomy supportive coaching and perceived justice

Despite the fact that perceived justice has been linked to performance-related variables within sport teams,^4–6 sport scientists have not yet studied which coach behaviors predict athletes’ perception of justice. In contrast, scientists in the organizational psychology stated that “perhaps the most natural connection can be made between justice and leadership.”¹¹ A recent meta-analysis in the business setting confirmed this statement.⁸ It identified leader behavior as an important source of justice and explicated the specific behaviors that leaders utilize to inform subordinates’ justice perceptions.⁹

In sport settings, autonomy support is often applied to study the influence of leader behavior (i.e., coach behavior). It is a key concept of the self-determination theory (SDT)¹² which is, one of the most dominant theories in the current field of sport psychology. Autonomy-supportive coaches: (a) offer choices within certain limits and listen to the input of athletes; (b) provide relevant information or a rationale for requested tasks, and rules; (c) acknowledge athletes’ feelings and perspective, (d) provide opportunities for initiative taking, and (e) transmit non-controlling competence feedback.¹³ The impact of autonomy support has been exhaustively examined in team sports,^13–15 however sport scientists have not yet tested whether autonomy support predicts the different dimensions of athletes’ perceived justice of the coach. Nevertheless, the dominant theories that describe why employees are concerned with justice provide support for the assumption that coaches’ autonomy support is positively related to all three dimensions of athletes’ perceived justice.⁹

First, the relational model of procedural justice argues that individuals care about procedural justice because of the relational messages communicated through fair processes.^16–18 According to this model, fair procedures matter not only because of control or voice, but also because they reaffirm group values and relational status in the decision-making process.^19,20 Based on people’s concerns for the relational messages expressed by fair processes,⁹ autonomy support is likely to predict perceptions of procedural justice. Autonomy supportive coaches consult athletes about matters that affect them which express standing to employees as well as fulfills their urge to be involved in the decision-making process. Furthermore, autonomy supportive coaches also acknowledge athletes’ feelings and perspective which affirms relational status.

Second, distributive justice is based in the equity theory.²¹ According to this theory distributive justice is measured by comparing the ratios of contributions and benefits to the contributions and benefits of referent others. Based on the above described definition, coaches’ autonomy support is likely to predict athletes’ perception of distributive justice. Autonomy supportive leaders provide rationale and communicate why decisions were made, which conveys information about expectations and standards. Further, they emphasize an open and comprehensive communication climate and promote intellectual stimulation which allows employees to more fully understand decisions. This reduces the possibility that misunderstandings about the outcomes occur and enhances employees’ justice perceptions.⁸

Third, interactional justice focuses on the extent to which people are treated with respect and reflects the comprehensiveness and truthfulness of explanations.² It has been described as encounter-based, as the interpersonal interactions between leaders and subordinates occur frequently and offers employees numerous opportunities to assess the fairness of information provided.²² In line with the fairness heuristic theory,²³ which suggests that subordinates make quick assessments of the fairness of their leaders based on initial justice encounters, it has been argued that interactional justice is more salient than other forms of justice. Taking into account that autonomy supportive coaches acknowledge athletes’ feelings and offer them relevant information, it is highly assumable that autonomy support is positively related to the interactional dimension of justice.

Perceived justice as a mediating mechanism

Besides the lack of knowledge on which coach behaviors predict athletes’ perception of justice, a second important gap in both the leadership and the justice literature regards the simultaneous examination of the effects of leadership style and justice on subordinate outcomes. That is, previous studies have examined the independent relations of leadership style and justice perceptions with subordinate outcomes, but they neglected the effects of these concepts when considered together. Karam et al. argue explicitly for greater integration of leadership and justice theories, because the simultaneous examination of both concepts would provide a more comprehensive view on the effects of a leader’s decisions and behaviors.⁹ Importantly, van Knippenberg et al.²⁴ even stated that a more full-blown contribution to research in leadership requires research focusing on fairness as an explanatory (mediating) mechanism. That is, it is clear that leadership is an important determinant of a wide range of followers’ outcomes. Yet, leaders’ behaviors provide followers with informational cues about their leaders’ justice which, in turn, functions as a mechanism that affects followers’ functioning.²⁵ Accordingly, research has shown some evidence to support the statement that organizational justice mediates the relationship between leadership on the one hand and subordinate-related outcomes such as satisfaction, commitment, and organizational citizenship behavior on the other hand.^26,27 However, scientists have only just begun to explore the potential of fairness as a mediating mechanism. And until today, no study has focused on fairness as the mediating mechanism explaining the effects of coach behaviors on athlete-related outcomes.

Research aims

The first main aim of the present study is to examine whether coaches’ leadership style and, more exactly an autonomy supportive coaching style, predicts athletes’ justice perceptions. The second main aim of the current study is to examine the mediating role of perceived justice in the relation between autonomy support and athlete outcomes. More specifically, Study 1 examined whether the relation between autonomy support and athletes’ satisfaction was mediated by athletes’ perception of justice. Study 2 aimed to replicate Study 1 and in addition analyzed whether perceived justice mediated the relation between coaches’ autonomy support and athletes’ performance.

In addition, pooling the data of both samples made it possible to set two side aims. First, to demonstrate the psychometric properties of the scales and to indicate their value and applicability for future sport scientists (i.e., taking the ambiguity of the organizational justice concept in the sport setting into account), a more robust psychometric analysis was performed on the pooled data. That is, measurement invariance was tested across both samples, which included athletes from different sports and gender. This procedure was conducted to demonstrate the psychometric properties of the scales within these specific studies. Second, to examine whether the relations of the structural models were similar across both samples, path model invariance was tested for the corresponding parts of the model of Study 1 and Study 2 (i.e., autonomy, justice and satisfaction with the coach).

Study 1

Supervisor satisfaction (i.e., satisfaction with the working method of a superior) has shown to play a crucial role in the optimal functioning of employees and workgroups. For example, employees who are satisfied with their supervisor showed stronger commitment to their company and less intentions to leave, as compared to employees that were unsatisfied.²⁸ Furthermore, multiple sport scientists argued that satisfaction plays an important role in athletes’ performance.^29,30

Several studies in the business setting have demonstrated a positive link between the different dimensions of organizational justice and employees’ supervisor satisfaction.⁷ The effects of procedural and interactional justice on satisfaction can be explained by the group-value model.³¹ This model assumes that fair interpersonal treatment bolsters peoples’ self-identity or self-worth and consequently reaffirms attachments to the relationship with supervisors. In addition, distributive justice also affects one’s feelings of respect and self-esteem based on the information it provides about one’s position in the group. Research indicated that one whose self-identity, self-worth, and perceived respect has been increased through fair treatments and rewards, is more likely to derive feelings of satisfaction.³²

Furthermore, the target similarity model proposed by Lavelle, Rupp, and Brockner³³ may help explain the relation between justice and satisfaction. This theory suggests that the outcomes of the different dimensions of justice should be considered via the source typically responsible for altering these individual justice perceptions.³³ In a team sport setting the head coach is an important source of distributive, procedural and interactional justice.^5,6 As a result, it is likely that all three justice dimensions are important predictors of athletes’ satisfaction with the working method of their coach.

In addition, previous studies revealed a positive relation between managers’ autonomy support and employees’ satisfaction.³⁴ Consequently, a positive relation between coaches’ autonomy support and athletes’ satisfaction was expected. Previous research in the business context showed that subordinates’ organizational justice perceptions explained the effects of supervisors behavior on subordinates’ job satisfaction.^26,27 From a theoretical point of view these results can be explained by the fact that supervisor behaviors provides information about the employee’s relationship with the management. Consequently, perceived (un)fairness resulting from supportive/abusive supervisor behavior is likely to translate into job (dis)satifaction.^26,27 In line with these findings in the business setting, it can be expected that athletes’ perceived justice of the coach mediates the relation between coaches’ autonomy support and athletes’ satisfaction in a team sport context. Based on the above described literature, the following hypotheses were formulated:
Hypothesis 1: Athletes’ perceived autonomy support of the coach is positively related to the different dimensions of athletes’ perceived justice of the coach.

Hypothesis 2a: The three different dimensions of athletes’ perceived justice of the coach are positively related to athletes’ satisfaction with the coach.

Hypothesis 2b: The positive relation between perceived autonomy support of the coach and athletes’ satisfaction with the coach is mediated by athletes’ perceived justice.

Method

Participants

To compile the research sample we contacted the coaches of 13 Belgian competitive dance teams. All coaches agreed to participate in our study. The dance teams consisted of 15 dancers each, except for two teams which consisted of 16 dancers. In total 152 of the 197 dancers filled out our questionnaires, which implies a total response rate of 77.2%. The vast majority of the research group consisted of female dancers. More specifically, 145 female dancers compared to only seven male dancers. To enable the possibility to replicate the results across gender and because the sample of study 2 only consisted of male athletes, the seven male respondents were excluded from this sample. It should be noted that the response rates for two teams were lower (i.e., 33.3% and 38.5%) than the response rates of the other 11 teams (i.e., equal or above 60%). The data of those two teams were collected during the first week of December. In this period more athletes than usual skipped the training because of the high study workload (104 dancers were high school students, 18 were college/university students, 23 did not study anymore).

The participants were members of disco and street dance formations which share the characteristics of other sport teams. First, like other sport teams, the dancers are interdependent and have to cooperate to achieve the team goals. Second, in order to optimize the team performance group training sessions are held under supervision of a trainer/coach. The coaches of the teams that participated in our study were no part of the team (i.e., they did not dance with the team). Finally, the teams participate weekly in a dance contest of one of the two Belgian dance federations (i.e., the Federation of Disco, Show and Formation [FDSF] or the Dansliga). The dancers were on average 18.42 years old (SD = 8.22). They had been members of their respective teams for on average 5.97 years (SD = 2.87), had on average 8.49 years of experience (SD = 3.65), and trained on average 1.29 times a week (SD = .50).

Procedures

The ethical standards of the American Psychological Association (APA) and the Social and Societal Ethics Committee of the KU Leuven were followed in the conduct of the study. The head coaches of 13 teams were informed by telephone about the research and all coaches allowed their teams to take part in the study. Next, an appointment with the team was scheduled on a training day. During this appointment a trained research assistant verbally informed the athletes about the objectives of our study and invited them to participate. The accurate timing of this briefing depended on the coach’s preference (i.e., during or after the training session). The athletes who agreed to participate received a paper-and-pencil questionnaire after the training session. They completed the questionnaire in the presence of the research assistant, who clarified ambiguities and answered possible questions. Prior to completing the questionnaire, it was stated that participation was voluntary and that the anonymity would prevail at all times.

Measures

The Dutch questionnaire gathered information about autonomy support (8 items), perceived justice (12 items), and satisfaction with the coach (5 items). Participants had to answer all items on a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree). The psychometric properties of all scales were tested within this study and can be found in the results section. Invariance tests of the factor structures across studies were conducted in Study 2.

Autonomy support of the coach (8 items)

The subscale autonomy support (eight items) of the Teacher as Social Context Questionnaire was used.³⁵ The Dutch version has already been used in previous research in the physical education context.³⁶ The items were slightly adapted to fit the specific team sport context (e.g., replaced ‘PE teacher’ by ‘coach’). An example item is: “My coach encourages me to ask questions.”

Perceived justice (12 items)

Justice perceptions were assessed with 12 items. More particular, three subscales with each four items: distributive, procedural, and interactional justice. To measure distributive and procedural justice eight previously back-translated Dutch items were used.^5,6 The validity and reliability of the Dutch distributive and procedural justice items has already been supported.^5,6 In the current study, the distributive justice items were slightly modified to fit the dance context. More specifically, the ‘position’ in the dance formation replaced the ‘play minutes’ as outcome. This position is a vital outcome in dance teams (i.e., positions in the front are favored above positions in the middle, or at the back of the formation). Furthermore, to measure interactional justice four items from a business oriented justice measurement were adapted and back-translated.³⁷ The original measurement of interactional justice consists of two subscales, (a) informational justice, and (b) interpersonal justice. However, taking into account the strong overlap between the items of the autonomy support scale and the four items of the interpersonal justice scale, this subscale was omitted. Example items of the justice measurement used in the current study are: “My position on the stage is a true reflection of my contribution to the team” (i.e., distributive justice), “My coach is consistent in making decisions” (i.e., procedural justice) and “I’d understand the reasons behind the team formation” (i.e., interactional justice).

Satisfaction with the coach (5 items)

A five-item Dutch scale based on the training and instruction subscale of the Athlete Satisfaction Questionnaire was developed.³⁸ The existing three items were slightly adapted and two extra items were added to measure how satisfied the athletes were with the working method and the approach of their coach. An example of the satisfaction items is: “I am satisfied with the training approach I receive from the coach”.

Data analysis

First, descriptive statistics, intercorrelations and the internal consistencies (Cronbach’s alpha) of all subscales were calculated. Second, the factorial structure of the measurement scales (i.e., autonomy support, perceived justice, and satisfaction) was tested using a confirmatory factor analysis (CFA). Four absolute fit indices were used to evaluate overall model fit were: normed chi-square (χ²/df), comparative fit index (CFI), Tucker-Lewis index (TLI), standardized root mean square residual (SRMR), and root mean square error of approximation (RMSEA). The normed chi-square statistic (χ²/df) was considered to indicate a good fit when its value was below 3:1.³⁹ Furthermore, a CFI and TLI value close to .95, a SRMR value close to .08 and a cutoff value close to .06 for the RMSEA show a good fit of the model.⁴⁰ Third, structural equation modelling (SEM) was conducted to test the hypothesized model. The single-indicator approach was used, because it corrects for measurement error in sample sizes that are comparable to the sample size of our study. Scale means were used as indicators of the latent variables and each indicator’s factor loading and error variance was fixed following the procedure of Brown.⁴¹ See the paper of Ng, Ntoumanis, and Thøgersen-Ntoumani for a similar approach.⁴² Mplus version 7.4 was used for all statistical analyses.

Results

Descriptive statistics, correlations and scale reliabilities

Means, standard deviations, and correlations for the variables of Study 1 are provided in Table 1. Scale reliabilities (Cronbach’s alphas) are provided on the diagonal.

Table 1.
Means, standard deviations, correlations and Cronbach’s alphas for the variables used in Study 1.

Variable M SD 1 2 3 4

1. Autonomy support (8 items) 3.62 .61 (.80)

2. Distributive justice (4 items) 3.55 .74 .29* (.86)

3. Procedural-Interactional justice (8 items) 3.54 .61 .55* .47* (.83)

4. Satisfaction with the coach (5 items) 4.11 .70 .51* .35* .72* (.90)

Note. All scales used a 5-point Likert-type scale.

p < .001.

CFA of autonomy, justice, and satisfaction

Autonomy support of the coach

The CFA on the one-factor model for autonomy support provided acceptable fit to the data (chi square (χ²) = 36.71, df = 20, p = .01; CFI = .94; TLI = .92; SRMR = .05; RMSEA = .08) and factor loadings (.32 to .71) were significant (p < .001).

Perceived justice

The three dimensions are most often treated as distinct aspects of justice, however especially procedural and interactional justice frequently showed to be highly correlated. In the current study the CFA on the proposed three-factor model for athletes’ perceived justice of the coach provided a good fit to the data (χ² = 92.54, df = 51, p < . 001; CFI = .94; TLI = .92; SRMR = .06; RMSEA = .08). All factor loadings were significant (p < .001) and ranged from .46 to .85. Furthermore, the three-factor CFA showed moderate correlations between distributive justice on the one hand and procedural justice and interactional justice on the other hand (r = .50, p < .001 and r = .58, p < .001 respectively). However, the correlation between procedural and interactional justice was very high (r = .96, p < .001).

Due to the high correlation between both factors and aligning with previous empirical studies,⁸ the procedural and interactional justice items were merged into one subscale and a CFA on the two-factor model of justice (i.e., distributive justice and procedural-interactional justice) was performed. The two-factor model provided a good fit to the data (χ² = 94.37, df = 53, p < .001; CFI = .94; TLI = .92; SRMR = .06; RMSEA = .08) and all factor loadings (ranging from .45 to .85) were significant (p < .001). In addition, the two-factor model showed a moderate correlation between distributive justice and procedural-interactional justice (r = .54, p < .001). Based on these findings, the two-factor model of perceived justice was used in the main analyses.

Satisfaction with the coach

The CFA on the proposed one-factor model for athletes’ satisfaction with the coach provided good fit to the data (χ² = 4.89, df = 4, p = .30; CFI = 1.00; TLI = 1.00; SRMR = .02; RMSEA = .04) and all factor loadings (.65 to .89) were significant (p < .001).

Main analyses: SEM of autonomy, justice, and satisfaction

The hypothesized model provided a good fit to the data (χ² = 0.37, df = 1, p = .54; CFI = 1.00; TLI = 1.00; SRMR = .01; RMSEA = .00). The standardized regression path coefficients and the proportions explained variance were illustrated in Figure 1. To provide a more robust assessment of the parameter estimates, a bootstrapping procedure of 5000 iterations was performed.⁴³ The bootstrapped estimates had approximately the same values as the estimates achieved with the maximum likelihood method and are shown in Table 2.

Figure 1.
The structural equation model of autonomy support, perceived justice and satisfaction with the coach with standardized coefficients. Proportions explained variance (italics) and the covariance between distributive and procedural-interactional justice are depicted. The dashed line represents the direct effect when testing mediation with the bootstrap approach advanced by Hayes et al. (2011). Note.* p ≤.001.

Table 2.
Path coefficients and indirect/direct effects for the hypothesized model.

Path coefficients
Effect

Bootstrapped standardized coefficient (SE) Bootstrapped 95% CI Bootstrapped standardized estimate (SE) Bootstrapped 95% CI

From AS to DJ −0.35 (0.096) −0.16 to 0.54

From DJ to SAT −0.11 (0.095) −0.32 to 0.06

From AS to PIJ −0.70 (0.073) −0.55 to 0.83

From PIJ to SAT −0.90 (0.073) −0.76 to 1.04

Indirect effect

AS → DJ → SAT −0.04 (0.040) −0.12 to 0.04

AS → PIJ → SAT −0.71 (0.146) −0.43 to 1.00

Direct effect

AS → SAT 0.09 (0.14) −0.19 to 0.36

Note.* AS = autonomy support; DJ = distributive justice; PIJ = procedural-interactional justice SAT = satisfaction.

The approach advanced by Hayes, Preacher and Myers⁴⁴ was followed to examine whether perceived justice mediated the relation between autonomy support and satisfaction. A bootstrapping procedure was applied with 5000 bootstrap resamples to determine beta coefficients and the confidence intervals of the indirect and direct effects (see Table 2). In partial support of Hypothesis 2b, the results revealed a significant indirect effect from autonomy support to satisfaction through procedural-interactional justice (β = .71, 95% CI [0.43–1.00]), while the indirect effect through distributive justice on satisfaction was non-significant (β = −.04, 95% CI [−0.12–0.04]). Finally, the direct effect from autonomy support on athletes’ satisfaction with the coach was non-significant (β = .09, 95% CI [−0.19–0.36]).

Discussion

In contrast to the many studies in the business setting, few sport scientists have studied the predictors of athletes’ perception of justice. Therefore, the first main aim of Study 1 was to examine whether autonomy support of the coach was related to the different dimensions of athletes’ perceived justice. In line with Hypothesis 1, autonomy support of the coach showed a positive and substantial link to athletes’ perception of justice of their coach. More specifically, autonomy support explained 12% of the variance in athletes’ perceived distributive justice and 49% of the variance of perceived procedural-interactional justice.

The second main aim was to check whether the different subcomponents of justice mediate the relation between autonomy support and satisfaction. As a first step to test the mediation hypothesis, our model showed a positive and strong relation between athletes’ perception of procedural-interactional justice and their satisfaction with the coach. This result is consistent with Hypothesis 2a and previous research in the business setting.² However, in contrast to Hypothesis 2a athletes’ perceived distributive justice was not related to their satisfaction with the coach. While this contradicts theoretical expectations, it could possibly be declared by the fact that athletes see the ‘unfair’ distribution of outcomes not as the sole responsibility of the coach, but more as something inherently in team sports. As a result, athletes could be more focused on the coach as source of fair procedures and the provision of reasonable explanations. Consistent with the target similarity model, this would lead to less strong relations between distributive justice and satisfaction with the coach approach.³³

In the second step to test the mediational hypothesis both forms of justice were included in the full research model. This led to a non-significant direct relation between autonomy support and athletes’ satisfaction. Furthermore, partially confirming Hypothesis 2b, only procedural-interactional justice showed to mediate the relation between autonomy support and satisfaction with the coach. Together, autonomy support and justice explained 70% of the variance in athletes’ satisfaction with their coach.

Study 2

Study 1 demonstrated that coaches who provide more autonomy support were perceived as more fair, which in turn was strongly related to athletes’ satisfaction. The female formation dancers in Study 1 shared much of the characteristics of other team athletes, however it is appropriate to test whether the findings can be transferred to male athletes and other team sports. Therefore, Study 2 examined the same interrelations in a research sample of competitive male handball players (see Hypothesis 1, 2a and 2b). In addition to the relations examined in Study 1, the specific microcosm of competitive sport teams provides a unique opportunity to examine the relations between justice on the one hand and the progression of athletes’ performance on the other hand.

The justice-performance relation is explained by several theoretical frameworks in the justice literature. Equity theory argued that when employees perceive distributive injustice, they will alter his or her quality or quantity of work to restore justice.²¹ In addition, researchers have relied on the social exchange theory to explain the relations between procedural and interactional justice and employees’ performance. Social exchange theory states that organizations are arenas for long-term, mutual social transactions between the employees and the organization.^45,46 Taken into account the results of Study 1 and given that the main source of athletes’ perceived procedural and interactional justice is the treatment by their coach, procedural and interactional justice seem more personal in nature than distributive justice. As such, both procedural and interactional justice can be perceived as an outcome an athlete receives, which in turn would influence the athlete’s outcome/input ratio.⁴⁷ In other words, when coaches treat athletes fairly, athletes reciprocate by better performance.

Unfortunately, only a few studies have tried to empirically test the effects of the different dimensions of justice on performance and the ones who did have been reporting contradictory findings. Walumbwa, Cropanzano, and Hartnell showed the strongest correlation between procedural justice and learning behavior and performance.⁴⁸ In contrast, Wang, Liao, Xia and Chang reported that interactional justice had the strongest impact on employees’ performance.⁴⁹ To complete the ambiguity, Williams found that distributive and interactional justice were linked to performance, while no significant link was found with procedural justice.⁵⁰ Cohen-Charash and Spector performed a meta-analyses in the business setting to shed light on the impact of the different justice dimension on employees’ work performance.² They reported that procedural justice is strongly related to work performance, while distributive and interactional justice were hardly related. Based on the results of this meta-analysis, Study 2 will test whether procedural justice is the only justice dimension that predicts athletes’ progression of performance in a team sport context.

In addition, the relation between coaches’ autonomy support and athletes’ optimal functioning has been repeatedly investigated. Adie, Duda and Ntoumanis demonstrated that a coach-created autonomy supportive environment fosters the basic need satisfaction and the occurrence of optimal functioning among male and female adult athletes.¹⁴ Furthermore, Gillet, Vallerand, Amoura, and Baldes indicated that coaches' autonomy support facilitates self-determined motivation and sport performance.⁵¹ In line with previous research in the business setting^24,26,27 and the results of Study 1, it can be expected that the hypothesized positive relation between coaches’ autonomy supportive behavior and performance can be explained by higher levels of athletes’ perceived (procedural) justice. Consequently, two extra hypotheses were formulated for Study 2.
Hypothesis 3a: Ahletes’ perceived procedural justice of the coach is positively related to athletes’ self-rated progression of performance.

Hypothesis 3b: The relation between perceived autonomy support of the coach and athletes’ self-rated progression of performance is mediated by athletes’ perceived procedural justice.
In addition, after pooling the data of both samples, measurement invariance tests were performed across both samples. Besides testing the psychometric properties of the measurement scales within both samples independently, measurement invariance tests were performed across both samples, which included athletes from different sports and gender. This procedure was conducted. These tests contribute to a more robust analysis of the psychometric properties of the scales. Furthermore, path invariance was tested to examine whether the different corresponding relations of the hypothesized structural model were similar across the samples of Study 1 and Study 2, thereby providing empirical evidence for the robustness of this study’s findings.

Method

Participants

The head coaches of eleven Belgian handball teams were contacted. All coaches agreed to participate in our study. Four handball teams consisted of 17 players, six handball teams consisted of 18 players and one team consisted of 21 players. In total 102 of the 197 male handball players filled out our questionnaires, which implies a total response rate of 51.78%. However, it should be noted that the response rate for one team was significantly lower (i.e., 19.05%) than the response rate of the other 10 teams (i.e., equal or above 40%). The lower response rate of this team could be explained by the fact that only half of the athletes were native Dutch speakers.

The participants competed in different leagues of the Belgian handball competition, organized by the Belgische Handbal Bond (BHB) and the Vlaamse Handbal Vereniging (VHV). Three teams played in the second league (n = 26), four teams played in the third league (n = 42), three teams played in the fourth league (n = 29), and one team played in a regional league (n = 5). The athletes were on average 23.41 years old (SD = 4.63), played handball for on average 13.61 years (SD = 5.15), had been members of their teams for on average 10.71 years (SD = 5.49), and trained on average 2.38 times a week (SD = .76).

Procedures

The ethical standards of the APA and the Social and Societal Ethics Committee of the KU Leuven were followed in the conduct of the study. To recruit the research sample the presidents and administrators of 23 Flemish handball teams were informed by e-mail about the research. If they did not respond within one week, they were contacted by telephone. The coaches and athletes of the 11 teams who agreed to participate were informed by e-mail about the aims and the design of the study and received a link to the web-based questionnaire. One week after the first e-mail, a reminder e-mail with the link to the web-based questionnaire was sent to the athletes who had not filled in the questionnaire. The questionnaire was introduced by a statement to the athletes’ that their participation was voluntary and that anonymity of their answers would prevail at all times. Furthermore, the importance of responding independently and honestly was emphasized.

Measures

The same Dutch scales as in Study 1 were used to measure coaches’ autonomy support (8 items), athletes’ perceived justice of the coach (12 items), and satisfaction with the coach (5 items). In addition, athletes rated the progression of their performance with three items. Results of the psychometric analyses of all scales of Study 2 and the results of the invariance tests of the factor structures across the samples can be found in the results section of Study 2.

Autonomy support of the coach (8 items)

The same eight items as in Study 1 were used to measure athletes’ perceived autonomy support of their coach.

Perceived justice (12 items)

The same 12 items as in Study 1 were used, except for a marginally adaptation of the distributive justice items. More precisely, ‘play minutes’ replaced ‘position’ in the dance formation as outcome to fit the specific handball context.

Satisfaction with the coach (5 items)

The same five items as in Study 1 were used to measure athletes’ satisfaction with the coach.

Athletes’ self-rated progression (3 items)

This construct measures athletes’ perception of their technical, tactical, and physical progression during the season (e.g., Rate your technical/tactical/physical progression since the beginning of the season”). Athletes scored the three Dutch items on a 7-point Likert scale (strong decline = 1; strong progression = 7).

Data analysis

Similar to Study 1, the internal consistencies of all subscales were calculated using Cronbach’s alpha, and correlations and descriptive statistics were examined. Second, the factorial structure of the different scales (i.e., autonomy support, perceived justice, satisfaction, and progression) was tested using a CFA. Third, after conducting CFAs on the data of Study 2, measurement invariance of the constructs that were also measured in Study 1 (i.e., autonomy support, justice and satisfaction) was tested. For this purpose, the data of Study 1 and Study 2 were pooled and the procedure of Byrne^52,53 to test for measurement invariance was followed. In addition to conducting CFAs in the separate samples, this includes testing a series of four increasingly restrictive nested models.⁵⁴ First, a configural invariance model was tested in which the factor model is fitted for both samples simultaneously. Second, a model was estimated in which the factor loadings were constrained to be equal across both samples. Third, a model was tested in which factor loadings and variances were constrained and fourth, a final model constrained factor loadings, variances and covariances between factors, which was only applicable to the construct of justice. To test invariance, ΔCFI between the less and the more constrained model was calculated and a ΔCFI of −0.01 was used as criterion to confirm or reject measurement invariance.⁵⁵

As a fourth step in our analyses, progression of performance was added as a new dependent variable to the structural model in addition to satisfaction. Structural equation modelling (SEM) was conducted to test whether justice mediated the relations between autonomy support on the one hand, and satisfaction and progression of performance on the other hand. Analogous to Study 1, a single-indicator structural equation model was performed.⁴² Finally, when measurement invariance of the constructs was established, path invariance of the part of the hypothesized structural model that was similar across both studies was tested across the samples of Study 1 and Study 2. Testing for path invariance includes (a) estimating a configural invariance model on the pooled data (i.e., estimating the model simultaneously for both samples), and (b) testing a model in which the regression paths are constrained to be equal across samples. Next, the differences in fit between both models were calculated and a ΔCFI of −0.01 was used as criterion to confirm or reject path invariance.⁵⁵ Mplus version 7.4 was used for all statistical analyses.

Results

Descriptive statistics, correlations and scale reliabilities

Means, standard deviations, and correlations for the variables of Study 2 are provided in Table 3. Scale reliabilities (Cronbach’s alphas) are provided on the diagonal.

Table 3.
Means, standard deviations, correlations and Cronbach’s alphas for the variables used in Study 2.

Variable M SD 1 2 3 4 5

1. Autonomy support (8 items) 3.37 .56 (.75)

2. Distributive justice (4 items) 3.67 .75 .41* (.88)

3. Procedural-Interactional justice (8 items) 3.59 .58 .40* .43* (.84)

4. Satisfaction with the coach (5 items) 3.74 .77 .41* .33* .75* (.92)

5. Progression of performance (3 items) 4.76 .89 .06* .16* .29* .34* (.75)

Note. All scales used a 5-point Likert scale, except progression of performance used a 7-point Likert scale.

p < .05.

CFA of autonomy, justice, satisfaction, and progression

Autonomy support of the coach

The CFA on the eight-item one-factor model for autonomy support provided acceptable fit to the data (χ² = 16.26, df = 18, p = .58; CFI = 1; TLI = 1.02; SRMR = .05; RMSEA = .00) and all factor loadings (.22 to .78) were significant. The rather low factor loading of .22 could possibly be explained by the negative formulation of this item (i.e., My coach doesn’t really listen to my opinion).

Perceived justice

The three-factor model of perceived justice provided a good fit to the data (χ² = 65.08, df = 48, p = .05; CFI = .97; TLI = .96; SRMR = .06; RMSEA = .06; factor loadings were significant (p < .05) and ranged from .28 to .82). In line with Study 1, the results showed moderate correlations between distributive justice on the one hand and procedural and interactional justice on the other hand (r = .48, p < .001 and r = .48, p < .001 respectively), while the correlation between procedural and interactional justice was very high (r = 1, p < .001).

Consequently, the procedural and interactional justice subscales were merged into one procedural-interactional justice subscale. The CFA of the two-factor model provided good fit to the data (χ² = 62.67, df = 50, p = .11; CFI = .98; TLI = .97; SRMR = .06; RMSEA = .05; factor loadings were significant (p < .05) and ranged from .31 to .92). The two-factor model showed a moderate correlation between distributive justice and procedural-interactional justice (r = .47, p < .001). Analogous to Study 1, the main analyses were performed with the two-factor model of perceived justice.

Satisfaction with the coach

CFA on the proposed one-factor model for athletes’ satisfaction with the coach provided good fit to the data (χ² = 1.48, df = 5, p = .92; CFI = 1.00; TLI = 1.02; SRMR = .01; RMSEA = .00) and all factor loadings (.66 to .90) were significant.

Athletes’ self-rated progression

The CFA on the proposed one-factor model for athletes’ progression provided good fit to the data (χ² = 1.90, df = 1, p = .17; CFI = .99; TLI = .96; SRMR = .04; RMSEA = .01) and all factor loadings (.73 to .78) were significant.

Measurement invariance of autonomy, justice and satisfaction

The configural invariance models of autonomy support of the coach (χ² = 61.92, df = 40, p = .02; CFI = .94; TLI = .92; SRMR = .05; RMSEA = .07), perceived justice (χ² = 165.53, df = 106, p < .001; CFI = .94; TLI = .93; SRMR = .06; RMSEA = .07), and satisfaction with the coach (χ² = 14.38, df = 10, p = .16; CFI = .99; TLI = .99; SRMR = .02; RMSEA = .06) showed a good fit to the data. The comparison between the increasingly restrictive models, indicated by ΔCFI, can be found in Table 4. These results confirmed measurement invariance for autonomy support and satisfaction with the coach. For perceived justice, the ΔCFI threshold of −0.01 was slightly exceeded when constraining all factor loadings, thereby indicating no full invariance. In this case, it is recommended to investigate which parameters cause the non-invariance. If only a small portion of loadings are not invariant, group comparisons can still be made.^55,56 Further examination revealed that freeing only one factor loading already ensured a satisfactory ΔCFI (see Table 4). The following more restricted nested models also included this free factor loading, and the ΔCFI was always below the threshold, thereby confirming partial invariance for perceived justice.

Table 4.
CFI and ΔCFI for the measurement invariance tests of the constructs used in both Study 1 and Study 2.

CFI/ΔCFI

Autonomy support Perceived justice Satisfaction with coach

Model 1 0.938 0.940 0.992

Model 2 0.928/0.010 0.921/−0.0190.933/−0.007 0.992/0.00

Model 3 0.928/0.00 0.935/0.002 0.992/0.00

Model 4 0.936/0.001

Note. Model 1 = configural invariance model (baseline model); Model 2 = factor loadings constrained; Model 3 = factor loadings and factor variances constrained; Model 4 = factor loadings, factor variances and factor covariances constrained.

CFI/ ΔCFI after freeing one item.

Main analyses: SEM of autonomy, justice, satisfaction, and progression

The hypothesized model provided an acceptable fit to the data (χ² = 3.16, df = 2, p = .21; CFI = .99; TLI = .96; SRMR = .02; RMSEA = .08). The standardized regression path coefficients and the proportions explained variance are illustrated in Figure 2. To provide a more robust assessment of the parameter estimates, a bootstrapping procedure of 5000 iterations was performed.⁴³ The bootstrapped estimates had approximately the same values as the estimates achieved with the maximum likelihood method and are shown in Table 5.

Figure 2.
The structural equation model of autonomy support, perceived justice, satisfaction with the coach, and self-rated progression of performance with standardized coefficients. Proportions explained variance (italics) and the covariance between distributive and procedural-interactional justice are depicted. Dashed lines represent the direct effects when testing mediation with the bootstrap approach advanced by Hayes et al. (2011). Note.* *p <.05, **p <.001.

Table 5.
Path coefficients and indirect/direct effects for the hypothesized model.

Path coefficients
Effects

Bootstrapped standardized coefficient (SE) Bootstrapped 95% CI Bootstrapped standardized estimate (SE) Bootstrapped 95% CI

Model 1

From AS to DJ −0.50 (0.092) −0.47 to 1.05

From DJ to S −0.08 (0.094) −0.30 to 0.09

From DJ to P −0.02 (0.174) −0.37 to 0.39

From AS to PIJ −0.53 (0.112) −0.31 to 0.93

From PIJ to S −0.90 (0.072) −1.01 to 1.59

From PIJ to P −0.34 (0.137) −0.13 to 0.92

Indirect effect:

AS → DJ → S −0.09 (0.089) −0.30 to 0.04

AS → DJ → P −0.06 (0.176) −0.31 to 0.39

AS → PIJ → S −0.65 (0.234) −0.31 to 1.21

AS → PIJ → P −0.33 (0.203) −0.08 to 0.85

AS → S −0.13 (0.147) −0.18 to 0.40

AS → P −0.17 (0.230) −0.63 to 0.27

Note. AS = autonomy support; DJ = distributive justice; PIJ = procedural-interactional justice; S = satisfaction; P = progression.

In order to test whether justice mediated the relation between autonomy support on the one hand and satisfaction and progression on the other hand, the same approach as in Study 1 was used.⁴⁴ A bootstrapping procedure was applied with 5000 bootstrap resamples to determine the beta coefficients and the confidence intervals of the indirect and direct effects between autonomy support and satisfaction/progression (see Table 5). Concerning Hypothesis 2 b, the results revealed only a significant indirect effect from autonomy support to satisfaction through procedural-interactional justice (β = .65, 95% CI [0.31–1.21]), while the indirect effect through distributive justice was non-significant (β = −.09, 95% CI [−.030–0.04]). Furthermore, the direct effect of autonomy support on satisfaction was non- significant (β = .13, 95% CI [−0.18–0.40]). Regarding Hypothesis 3 b, indirect effects from autonomy support to progression were calculated. The results revealed a significant indirect effect from autonomy support to progression through procedural-interactional justice (β = .33, 95% CI [0.08–0.85]), and a non-significant indirect effect through distributive justice (β = .06, CI [−0.31–0.39]). In addition, the direct effect of autonomy support on progression was non-significant (β = −.17, CI [−0.63–0.27]).

Path invariance: Structural model containing autonomy, justice and satisfaction

The configural invariance model of the hypothesized structural model showed a good fit to the data (χ² = 1.24, df = 2, p = .54; CFI = 1.00; TLI = 1.02; SRMR = .01; RMSEA = .00). Importantly, the ΔCFI between the configural invariance model and the model constraining all regression paths was equal to .002, thereby confirming path invariance across both samples.

Discussion

Consistent with Study 1, the results of Study 2 showed that autonomy support of the coach was positively and substantially associated with competitive male handball players’ perceptions of justice of their coach. More specifically, autonomy support of the coach explained 25% of the variance in athletes’ distributive justice and 29% of the variance in athletes’ procedural-interactional justice. Furthermore, in line with the results of Study 1, our model showed a strong positive relation between athletes’ perception of procedural-interactional justice and their satisfaction, while athletes’ perceived distributive justice was not significantly related to their satisfaction with the coach. Together, autonomy support and justice explained 75% of the variance in athletes’ satisfaction with the coach. Additionally, and replicating the results of Study 1, the direct relation between autonomy support and athletes’ satisfaction was non-significant. To conclude, our results indicated that in male handball teams autonomy support was strongly linked to both forms of perceived justice, but only procedural-interactional justice mediated the positive relation between autonomy support and athletes’ satisfaction with the coach.

In addition, Study 2 examined whether perceived justice and autonomy support were also related to the self-reported progression of athletes’ performance. Partially confirming Hypothesis 3a, our results showed that athletes’ perceived procedural-interactional justice was positively linked to their progression, while distributive justice was not significantly related to progression.^33,34 In line with the results of Cohen-Charash and Spector’s meta-analysis,² these results confirm the social-exchange predictions^42,43 regarding procedural and interactional justice. However, they contradict the equity theory, which assumed that athletes would alter their quality or quantity of work to restore justice.²¹

Regarding Hypothesis 3b, Study 2 checked whether the relation between autonomy support and progression was mediated by athletes’ perceived procedural justice. Therefore, the indirect effects from perceived autonomy support to athletes’ self-reported progression were calculated. In total the model explained 12% of the variance in self-rated progression. Based on the significant indirect effect from autonomy support through procedural-interactional justice on athletes’ progression and the non-significant direct path, it can be concluded that only procedural-interactional justice mediates the positive relation between autonomy support and male handball players’ progression.

General discussion

Athletes’ perceived justice of the coach has been shown to be positively associated with athletes’ team identification, cohesion, and consequently their level of social loafing in sport teams.^5,6 Taking into account the impact of these outcomes on team performance, two studies in different team sport settings were conducted to examine whether autonomy support of the coach predicts athletes’ perception of justice, and their satisfaction with the coach and their degree of progression.

To our knowledge, this is one of the first studies to explore and reveal an autonomy supportive coaching style as an antecedent of perceived justice in sport teams. In support of Hypothesis 1, the results indicated that coaches’ autonomy support is strongly related to both perceived procedural-interactional and distributive justice of team athletes. The strong link between autonomy support and procedural-interactional justice seems a logical consequence of the fact that autonomy supportive coaches consult athletes about matters that affect them, acknowledge athletes’ feelings and perspective, and offer them relevant information.¹³ These behaviors express standing to employees, fulfill their needs for representativeness in the decision-making process and affirm relational status. As a result, an autonomy-supportive coaching style positively affects athletes’ concerns for the relational messages expressed by fair procedures. Furthermore, autonomy supportive coaches provide rationale about why decisions were made, emphasize an open and comprehensive communication climate, and promote intellectual stimulation.^12,13 Consequently, athletes will more fully understand decisions, which reduces the possibility that misunderstandings about the outcomes (e.g., playing time, starting team) lead to negative perceptions of distributive justice.⁸ In support of this explanation, research showed that an autonomy supportive environment increases athletes’ tactical knowledge.⁵⁷ This increased knowledge reduces the probability of athletes misinterpreting decisions of the coach and could partially clarify why autonomy support enhances athletes’ perception of distributive justice. For example, athletes with well-developed tactical knowledge, are more likely to understand the tactical motives behind their substitution and as a result perceive their received playing minutes as more fair.

Furthermore, our studies showed that team athletes’ procedural-interactional justice is strongly linked to their satisfaction with the coach. These findings partially supported Hypothesis 2a and the group-value model.³¹ This model assumes that fair interpersonal treatment strengthens peoples’ self-worth and enhances the quality of the relationship with supervisors.

However, in contrast to Hypothesis 2a and a previous meta-analysis in the business context,² the current results indicated that distributive justice was not significantly related to athletes’ satisfaction. The use of fair procedures and the provision of reasonable explanations seem to be a more important predictor of supervisor satisfaction than the fair distribution of outcomes in team sport settings. This could be a consequence of the fact that an ‘unfair’ distribution of the outcomes is an inherent characteristic of team sports. Not all dancers can be on the front row, and some handball players will be on the bench for most of the time. As a result, athletes are probably more focused on the coach as source of fair procedures and the provision of reasonable explanations and less as a source of fair distributions of the outcomes. Consistent with the target similarity model, this appeared to lead to less strong relations between distributive justice and satisfaction with the coach approach.³³

Third, in line with Hypothesis 2b and previous research our results showed that autonomy support is positively related to athletes’ satisfaction.³⁴ More specifically, our models demonstrated that procedural-interactional justice mediated the relation between autonomy support and athletes’ satisfaction. Confirming previous research in the business setting,^26,27 the relation between autonomy support and satisfaction can thus be explained by its positive impact on athletes’ perception of justice. These results can be explained by the fact that autonomy supportive coach behaviors suggest to athletes that they are highly valued within the team, and reduce the probability of misunderstandings, thereby enhancing perceptions of justice. Consequently, perceived fairness resulting from supportive behavior is likely to translate into satisfaction with the working method of the coach.

Finally, partly confirming Hypothesis 3a, the results demonstrated a significant positive link between athletes’ perceived procedural-interactional justice and their self-rated progression of performance, while the relation between distributive justice and progression was not significant. As mentioned before, these results contradict the equity theory.²¹ However, they are consistent with a meta-analysis in the business setting that showed a strong relation between procedural justice and performance in field studies, while distributive justice was hardly related to performance.² This link between withholding performance and perceived injustice has previously been described by Cohen-Charash and Spector.² The current results support their statement that subordinates (athletes) do not automatically withhold performance when outcomes are unfairly distributed.² However, when they perceive a lack of justice in the procedures and the information provided by the supervisor (coach), they withhold performance as a legitimate way of restoring equity.²

In support of Hypothesis 3b and confirming previous research, our results showed that autonomy support is positively related to athletes’ progression of performance.⁵¹ More specifically, the model of Study 2 demonstrated that procedural-interactional justice mediated the relation between autonomy support and athletes’ progression. In addition to the fact that autonomy support fosters basic need satisfaction and self-determined motivation and in turn facilitates sport performance,⁵¹ the current findings indicate that autonomy support also predicts athletes’ perception of procedural-interactional justice of their coach, which in turn positively predicts their progression. However, as this was the first study to examine these relations, more research is necessary to get a complete picture of the potential of perceived justice as the mediating mechanism explaining the relation between coach behavior and outcomes like athletes’ progression.

This study also thoroughly examined the properties of the sport-specific scales of autonomy support, justice and satisfaction. The analyses of both studies independently supported the psychometric properties of the Dutch autonomy and satisfaction scales. Moreover, by pooling the data of both studies, the full measurement invariance of the Dutch autonomy and satisfaction measurement scales was proven across gender and sports. Concerning the Dutch sport specific justice scale, both Study 1 and Study 2 showed high correlations between procedural and interactional justice (i.e., respectively r = .96 and r = 1). Therefore, both concepts were combined into one procedural-interactional factor. The justice scale showed to be partial invariant (i.e., only one item was not invariant) across both samples. In conclusion, these findings confirm that the scales measure the same construct across the female dancers and the male handball players and provide evidence for its usability in different samples in sport settings.

The results concerning the combination of the procedural and interactional justice dimension into one factor are not as divergent as it seems at first sight. Although interactional justice is most often viewed and treated as a distinct aspect of justice,² several studies have been showing high intercorrelations between procedural and interactional justice and consequently combined both dimensions.^3,10 Furthermore, the specific characteristics of a team sport context could have had an influence. For instance, in business settings, the organization is viewed as the source of decisions and procedures, while the manager transfers the message to the employees.^24,58 Consequently, the manager is unlikely to be the target of judgments concerning procedural justice. In this respect, a team sport setting differs from a business setting. Sport coaches are usually in charge of both the procedures that were followed to achieve outcomes (i.e., procedural justice) as well as the communication about these procedures (i.e., interactional justice). This double responsibility of the coach could partly explain the high correlations between athletes’ perceptions of both justice concepts. Future research has to study whether it is simply impossible to distinguish the procedural and interactional justice dimensions in a wide range of sport settings or that sport scientists need to develop more sophisticated sport-specific justice measurements in order to examine the independent effects of the three justice dimensions in a sport setting.

Finally, path invariance tests showed the invariance of the corresponding parts of both structural models. The different relations in the model show no differences between both samples. This implies that the main conclusions of the current study apply for two different samples that contain different gender and a different type of sport.

Practical implications

From a practical point of view, elite dance and handball teams are focused on performance and winning. Considering this focus, our models are not only interesting for sport scientists but also for coaches because they can be used to keep athletes satisfied with the working method of the coach and to optimize their performance. The invariance of the structural models indicates that the guidelines below apply to coaches of both female dance teams and male handball teams. Future research should confirm whether these findings are transferrable to sport teams in general. Based on the strong link between autonomy support and athletes’ perception of justice, some strategies for coaches to provide autonomy support are suggested: (a) provide athletes input and some choice (within specific and clear limits), (b) provide information or a rationale for tasks, decisions, and rules, and (c) acknowledge athletes’ perspective and listen to their concerns.^13,59 Concerning the provision of choice, it should be clarified that offering athletes’ some choice and freedom is clearly not enough to create an autonomy supportive climate.⁶⁰ A clear goal structure and explicit limits are indispensable. If these conditions are met the responsibility for the learning process can be shared between the coach and the athletes.¹³ Regarding the provision of a truthful rationale, coaches can proactively clarify the selection of the dance formation based on transparent principles (e.g., commitment on training). In addition, they can objectively and with respect for their athletes’ disappointment discuss a poor performance and turn it into new challenging and shared goals. Finally, coaches who acknowledge athletes’ perspective treat them as individuals with specific needs and feelings, and not as mere pawns that should be directed. They show genuine interest on and off the field, listen to their concerns and empathize with their point of view. As a result, providing autonomy support requires considerable interpersonal skills of the coach, such as active listening, and persuasive communication.

Limitations and future research

As with any research, the current studies had limitations. First, the data are cross-sectional in nature, limiting causal inferences. Furthermore, given that coaching is a dynamical process that constantly fluctuates during the season, coach behavior will be very situation depended. As a result, longitudinal studies are recommended to examine its relation with perceived justice in greater depth. Longitudinal designs could provide clarity on the dynamical relation between satisfaction and progression, which is not possible within cross-sectional studies. In addition, future research should aim for a more refined knowledge on which specific determinants of autonomy support can enhance perceived justice in specific circumstances. Moreover, it is important to extend the coach behavior beyond autonomy support. As a starting point for this exploration, future research can be inspired by the theoretical model of Delrue et al. that offers a refined insight on different aspects of autonomy support (e.g., participation, attuning) and includes other dimensions of need support (e.g., competence support or structure, controlling coach behavior, and chaotic coach behavior).⁶¹

Second, a self-rated measurement of athletes’ progression has been used, because it is very hard to objectively determine individual progression in interactive sports (e.g., handball). In interactive team sports, individual progression depend strongly upon the performances of teammates and the quality of the opponent. For example, the attacking efficiency of the line player in handball teams is strongly determined by the quality of the assists of his backcourt players and the strength of the defense. Although such considerations underscore the difficulty of creating an objective measure of individual progression in interactive team sports, it is nevertheless important that future studies try to develop (sport-specific) objective measures, in order to rule out the biases inherent to self-report scales.

Third, as the same participants at the same time filled out self-reported questionnaires, common method bias should be mentioned as a possible limitation. A part of the variance might be attributable to the measurement method rather than to the constructs the measures represent.⁵⁹ To avoid this bias, future research could for example, assess the dependent variables on a different time point than the independent variables (e.g., measure players’ satisfaction on a different time point) or use different rating sources (e.g., coaches report the progression measurement for all the players, independent observations of the autonomy supportive behavior of the coach,…)⁶².

Fourth, although invariance tests were performed, future work should replicate these findings in other samples and provide more specific comparisons between different gender or type of sport. That is, the sample of Study 1 contains athletes from a different gender and a different sport than the sample of Study 2. Future studies could compare gender differences within the same type of sport or differences between types of sport for athletes from the same gender to enable more precise conclusions on gender or sport-invariance of these findings.

To conclude, the current study was one of the first to examine whether an autonomy supportive leadership style of the coach is linked to team athletes’ perceived justice of the coach. Our results showed that coaches who provide autonomy support are perceived as more just, which in turn predicts higher levels of athletes’ satisfaction with the coach and more self-reported progression. The consistency of the relations, as shown across gender and in different team sports, strengthens the reliability of the findings. Obviously, these results are also interesting for team sport coaches as they confirm that autonomy support is strongly linked to athletes’ perception of justice, which in turn plays a crucial role in the optimization of team athletes’ functioning.

	Variable	M	SD	1	2	3	4
1.	Autonomy support (8 items)	3.62	.61	(.80)
2.	Distributive justice (4 items)	3.55	.74	.29*	(.86)
3.	Procedural-Interactional justice (8 items)	3.54	.61	.55*	.47*	(.83)
4.	Satisfaction with the coach (5 items)	4.11	.70	.51*	.35*	.72*	(.90)

	Path coefficients	Effect
From AS to DJ	−0.35 (0.096)	−0.16 to 0.54
From DJ to SAT	−0.11 (0.095)	−0.32 to 0.06
From AS to PIJ	−0.70 (0.073)	−0.55 to 0.83
From PIJ to SAT	−0.90 (0.073)	−0.76 to 1.04
Indirect effect
AS → DJ → SAT			−0.04 (0.040)	−0.12 to 0.04
AS → PIJ → SAT			−0.71 (0.146)	−0.43 to 1.00
Direct effect
AS → SAT			0.09 (0.14)	−0.19 to 0.36

	Variable	M	SD	1	2	3	4	5
1.	Autonomy support (8 items)	3.37	.56	(.75)
2.	Distributive justice (4 items)	3.67	.75	.41*	(.88)
3.	Procedural-Interactional justice (8 items)	3.59	.58	.40*	.43*	(.84)
4.	Satisfaction with the coach (5 items)	3.74	.77	.41*	.33*	.75*	(.92)
5.	Progression of performance (3 items)	4.76	.89	.06*	.16*	.29*	.34*	(.75)

	CFI/ΔCFI
Model 1	0.938	0.940	0.992
Model 2	0.928/0.010	0.921/−0.0190.933/−0.007*	0.992/0.00
Model 3	0.928/0.00	0.935/0.002	0.992/0.00
Model 4		0.936/0.001

	Path coefficients	Effects
Model 1
From AS to DJ	−0.50 (0.092)	−0.47 to 1.05
From DJ to S	−0.08 (0.094)	−0.30 to 0.09
From DJ to P	−0.02 (0.174)	−0.37 to 0.39
From AS to PIJ	−0.53 (0.112)	−0.31 to 0.93
From PIJ to S	−0.90 (0.072)	−1.01 to 1.59
From PIJ to P	−0.34 (0.137)	−0.13 to 0.92
Indirect effect:
AS → DJ → S			−0.09 (0.089)	−0.30 to 0.04
AS → DJ → P			−0.06 (0.176)	−0.31 to 0.39
AS → PIJ → S			−0.65 (0.234)	−0.31 to 1.21
AS → PIJ → P			−0.33 (0.203)	−0.08 to 0.85
AS → S			−0.13 (0.147)	−0.18 to 0.40
AS → P			−0.17 (0.230)	−0.63 to 0.27

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.

ORCID iD

Stef Van Puyenbroeck

References

Greenberg

Organizational justice: yesterday, today, and tomorrow. J Manag 1990; 16: 399–432.

Cohen-Charash

Spector

PE.

The role of justice in organizations: a meta-analysis. Organ Behav Hum Decis Process 2001; 86: 278–321.

Kim

Andrew

DPS.

Relationships between organizational justice and coaches’ attitudinal outcomes in intercollegiate sports. Int J Sports Sci Coa 2015; 10: 305–326.

Whisenant

Organizational justice and commitment in interscholastic sports. Sport Educ Soc 2005; 10: 343–357.

De Backer

Boen

Ceux

, et al. Do perceived justice and need support of the coach predict team identification and cohesion? Testing their relative importance among top volleyball and handball players in Belgium and Norway. Psychol Sport Exerc 2011; 12: 192–201.

De Backer

Boen

De Cuyper

, et al. A team fares well with a fair coach: predictors of social loafing in interactive female sport teams. Scand J Med Sci Sports 2015; 25: 897–908.

Bray

Whaley

DE.

Team cohesion, effort, and objective individual performance of high school basketball players. Sport Psychol 2001; 15: 260–275.

Jordan

Gillentine

Hunt

BP.

The influence of fairness: the application of organizational justice in a team sport setting. Int Sports J 2004; 8: 139–149.

Karam

Davison

, et al. Illuminating the ‘face’ of justice: a meta-analytic examination of leadership and organizational justice. J Manag Stud 2019; 56: 134–171.

10.

Kim

Andrew

DPS.

Organizational justice in intercollegiate athletics: perceptions of coaches. Sport Manag Rev 2013; 16: 200–210.

11.

Colquitt

Greenberg

Organizational justice: a fair assessment of the state of the literature. In: Greenberg

(ed) Organizational behavior: the state of the science. 2nd ed. Mahwah: Lawrence Erlbaum Associates, 2003, pp. 165–210.

12.

Deci

Ryan

RM.

The “what” and “why” of goal pursuits: human needs and the self-determination of behavior. Psychol Inq 2000; 11: 227–268.

13.

Mageau

Vallerand

RJ.

The coach-athlete relationship: a motivational model. J Sports Sci 2003; 21: 883–904.

14.

Adie

Duda

Ntoumanis

Autonomy support, basic need satisfaction and the optimal functioning of adult male and female sport participants: a test of basic needs theory. Motiv Emot 2008; 32: 189–199.

15.

Adie

Duda

Ntoumanis

Perceived coach-autonomy support, basic need satisfaction and the well- and ill-being of elite youth soccer players: a longitudinal investigation. Psychol Sport Exerc 2012; 13: 51–59.

16.

Greenberg

Lind

EA.

The pursuit of organizational justice: from conceptualization to implication to application. In Cooper

Locke

(eds) I/O psychology: what we know about theory and practice. Oxford: Blackwell, 2000, pp. 72–105.

17.

Thibaut

Walker

Procedural justice: a psychological analysis. Hillsdale: Lawrence Erlbaum Associates, 1975.

18.

Blader

Tyler

TR.

Relational models of procedural justice. In: Cropanzano

Ambrose

(eds) The Oxford handbook of justice in the workplace. New York: Oxford University Press, 2015, pp. 351–370.

19.

Tyler

Blader

SL.

Cooperation in groups: procedural justice, social identity, and behavioral engagement. New York: Psychology Press, 2000.

20.

Lind

Kanfer

Earley

PC.

Voice, control, and procedural justice: Instrumental and noninstrumental concerns in fairness judgments. J Pers Soc Psychol 1990; 59: 952–959.

21.

Adams

JS.

Inequity in social exchange. In Berkowitz

(ed) Advances in experimental social psychology. New York: Academic Press, 1965, pp. 267–299.

22.

Holtz

Harold

CM.

Fair today, fair tomorrow? A longitudinal investigation of overall justice perceptions. J Appl Psychol 2009; 94: 1185–1199.

23.

Lind

EA.

Fairness heuristic theory: justice judgments as pivotal cognitions in organizational relations. In Greenberg

(ed) Advances in organizational justice. Stanford: Stanford University Press, 2001, pp. 56–88.

24.

van Knippenberg

De Cremer

van Knippenberg

Leadership and fairness: the state of the art. Eur J Work Organ Psychol 2007; 16: 113–140.

25.

Salancik

Pfeffer

A social information processing approach to job attitudes and task design. Adm Sci Q 1978; 23: 224–253.

26.

Tepper

BJ.

Consequences of abusive supervision. Acad Manag J 2000; 43: 178–190.

27.

Zellar

Tepper

Duffy

MK.

Abusive supervision and subordinates’ organizational citizenship behavior. J Appl Psychol 2002; 87: 1068–1076.

28.

DeConinck

Stilwell

CD.

Incorporating organizational justice, role states, pay satisfaction and supervisor satisfaction in a model of turnover intentions. J Bus Res 2004; 57: 225–231.

29.

Chelladurai

Riemer

HA.

A classification of facets of athlete satisfaction. J Sport Manag 1997; 11: 133–159.

30.

Lorimer

Coaches’ satisfaction with their athletic partnerships. Int J Coach Sci 2009; 3: 57–66.

31.

Lind

Tyler

TR.

The social psychology of procedural justice. 1st ed. New York: Plenum, 1988, p. 267.

32.

Hendrix

Robbins

Miller

, et al. Effects of procedural and distributive justice on factors predictive of turnover. J Soc Behav Pers 1998; 13: 611–632.

33.

Lavelle

Rupp

Brockner

Taking a multifoci approach to the study of justice, social exchange, and citizenship behavior: the target similarity model. J Manag 2007; 33: 841–866.

34.

Gillet

Gagné

Sauvagère

, et al. The role of supervisor autonomy support, organizational support, and autonomous and controlled motivation in predicting employees’ satisfaction and turnover intentions. Eur J Work Organ Psychol 2013; 22: 450–460.

35.

Belmont

Skinner

Wellborn

, et al. Teacher as social context: a measure of student perceptions of teacher provision of involvement, structure, and autonomy support. Technical report no. 102. Rochester: University of Rochester, 1988.

36.

Sierens

Vansteenkiste

Goossens

, et al. The synergistic relationship of perceived autonomy support and structure in the prediction of self-regulated learning. Br J Educ Psychol 2009; 79: 57–68.

37.

Colquitt

JA.

On the dimensionality of organizational justice: construct validation of a measure. J Appl Psychol 2001; 86: 386–400.

38.

Riemer

Chelladurai

Development of the athlete satisfaction questionnaire (ASQ). J Sport Exerc Psychol 1998; 20: 127–156.

39.

Kline

RB.

Principles and practice of structural equation modeling. 3rd ed. New York: The Guilford Press, 2011, p. 428.

40.

Bentler

PM.

Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Model 1999; 6: 1–55.

41.

Brown

TA.

Confirmatory factor analysis for applied research. 2nd ed. New York: The Guildford Press, 2015, p. 462.

42.

Ntoumanis

Thøgersen-Ntoumani

Autonomy support and control in weight management: what important others do and say matters. BR J Health Psychol 2014; 19: 540–552.

43.

Hayes

. Beyond Baron and Kenny: statistical mediation analysis in the new millennium. Commun Monogr 2009; 76: 408–420.

44.

Hayes

Preacher

Myers

TA.

Mediation and the estimation of indirect effects in political communication research. In: Bucy

Holbert

(eds) The sourcebook for political communication research: methods, measures, and analytical techniques. New York: Routledge, 2011, pp. 434–465.

45.

Cropanzano

Prehar

Chen

PY.

Using social exchange theory to distinguish procedural from interactional justice. Group Organ Manag 2002; 27: 324–351.

46.

Wayne

Shore

Liden

RC.

Perceived organizational support and leader member exchange: a social exchange perspective. Acad Manag J 1997; 40: 82–111.

47.

Greenberg

The social side of fairness: interpersonal and informational classes of organizational justice. In: Cropanzano

(ed) Justice in the workplace: approaching fairness in human resource management. Hillsdale: Lawrence Erlbaum Associates, 1993, pp. 79–103.

48.

Walumbwa

Cropanzano

Hartnell

CA.

Organizational justice, voluntary learning behavior, and job performance: a test of the mediating effects of identification and leader-member exchange. J Organiz Behav 2009; 30: 1103–1126.

49.

Wang

Liao

Xia

, et al. The impact of organizational justice on work performance: mediating effects of organizational commitment and leader-member exchange. Int J Manpower 2010; 31: 660–677.

50.

Williams

The effects of distributive and procedural justice on performance. The J Psychol 1999; 133: 183–193.

51.

Gillet

Vallerand

Amoura

, et al. Influence of coaches’ autonomy support on athletes' motivation and sport performance: a test of the hierarchical model of intrinsic and extrinsic motivation. Psychol Sport Exerc 2010; 11: 155–161.

52.

Byrne

BM.

Testing for multigroup invariance using AMOS graphics: a road less traveled. Struct Equ Model 2004; 11: 272–300.

53.

Byrne

BM.

Structural equation modeling with AMOS. Basic concepts, applications, and programming. New York: Routledge Taylor & Francis Group, 2010.

54.

Van Puyenbroeck

Stouten

Vande Broek

Coaching is teamwork! the role of need‐supportive coaching and the motivational climate in stimulating proactivity in volleyball teams. Scand J Med Sci Sports 2018; 28: 319–328.

55.

Cheung

Rensvold

RB.

Evaluating goodness-of-fit indexes for testing measurement invariance. Struct Equ Model 2002; 9: 233–255.

56.

Putnick

Bornstein

MH.

Measurement invariance conventions and reporting: the state of the art and future directions for psychological research. Dev Rev 2016; 41: 71–90.

57.

Vande Broek

Boen

Claessens

, et al. Comparison of three instructional approaches to enhance tactical knowledge in volleyball among university students. J Teach Phys Educ 2011; 30: 375–392.

58.

McFarlin

Sweeney

PD.

Distributive and procedural justice as predictors of satisfaction with personal and organizational outcomes. AMJ 1992; 35: 626–637. doi.org/10.5465/256489.

59.

Reinboth

Duda

Ntoumanis

Dimensions of coaching behavior, need satisfaction, and the psychological and physical welfare of young athletes. Motiv Emot 2004; 28: 297–313.

60.

De Muynck

Soenens

Waterschoot

, et al. Towards a more refined insight in the critical motivating features of choice: an experimental study among recreational rope skippers. Psychol Sport Exerc 2019; 45: 101561.

61.

Delrue

Reynders

Vande Broek

, et al. Adopting a helicopter-perspective towards motivating and demotivating coaching: a circumplex approach. Psychol Sport Exerc 2019; 40: 110–126.

62.

Podsakoff

MacKenzie

Lee

, et al. Common method bias in behavioral research: a critical review of the literature and recommended remedies. J Appl Psychol 2003; 88: 879–903.

	Path coefficients		Effect
	Bootstrapped standardized coefficient (SE)	Bootstrapped 95% CI	Bootstrapped standardized estimate (SE)	Bootstrapped 95% CI
From AS to DJ	−0.35 (0.096)	−0.16 to 0.54
From DJ to SAT	−0.11 (0.095)	−0.32 to 0.06
From AS to PIJ	−0.70 (0.073)	−0.55 to 0.83
From PIJ to SAT	−0.90 (0.073)	−0.76 to 1.04
Indirect effect
AS → DJ → SAT			−0.04 (0.040)	−0.12 to 0.04
AS → PIJ → SAT			−0.71 (0.146)	−0.43 to 1.00
Direct effect
AS → SAT			0.09 (0.14)	−0.19 to 0.36

	CFI/ΔCFI
	Autonomy support	Perceived justice	Satisfaction with coach
Model 1	0.938	0.940	0.992
Model 2	0.928/0.010	0.921/−0.0190.933/−0.007*	0.992/0.00
Model 3	0.928/0.00	0.935/0.002	0.992/0.00
Model 4		0.936/0.001