The role of leadership in group cohesion: insights from sport club volunteering

Abstract

Purpose

This study aims to address the challenge many organizational leaders face in fostering workforce cohesion. Focusing on the context of sports club volunteering, this study investigates how leaders can enhance group cohesion among volunteers. The study findings provide valuable insights applicable across various work settings.

Design/methodology/approach

Through a multilevel design, this study examined the role of autonomy–supportive and structuring leadership in shaping social and task cohesion within volunteer teams at the group level. In total, 557 volunteers nested within 52 nonprofit sports clubs situated in the Flemish region of Belgium were involved in this study, providing a robust foundation for our analysis.

Findings

This research revealed that regular volunteers form “true groups”, exhibiting substantial between-group variance in social and task cohesion and a strong within-group consensus. Additionally, the findings underscored the significance of autonomy–supportive leadership in fostering cohesion, demonstrating a positive relation with social and task cohesion at the group level.

Originality/value

This study demonstrates that volunteer teams, like work teams in the for-profit sector, actively engage in interpersonal exchanges within their organization, which help shape their collective sense of unity and alignment with common objectives. Leaders can enhance these interactions by fostering an autonomy–supportive environment where members feel empowered to share their perspectives and ideas. Additionally, the findings suggest that the nature of the task and the specific context can influence which leadership style is most effective, with the provision of structure also playing a role. With these insights, leaders in diverse organizational settings can effectively nurture the development of cohesive groups.

Citation

De Clerck, T., Van Doren, N. and De Bock, T. (2024), "The role of leadership in group cohesion: insights from sport club volunteering", Organization Management Journal , Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/OMJ-01-2024-2061

Introduction

In many organizations, leaders encounter challenges in effectively managing and retaining their workforce (Forner, Jones, Berry, & Eidenfalk, 2020, Forner et al., 2023). A key factor influencing workforce effectiveness and commitment is group cohesion (Doherty & Carron, 2003; Castaño, Watts, & Tekleab, 2013; Grossman, Nolan, Rosch, Mazer, & Salas, 2022). Group cohesion within an organization is characterized by the social bonds among workers and their collective dedication to organizational goals. When a group is highly cohesive, its members are more likely to thrive, leading to enhanced performance and higher retention rates (Doherty & Carron, 2003; Castaño et al., 2013; Grossman et al., 2022).

The central aim of this study is to examine the leadership styles that foster group cohesion. For this purpose, we focus on the specific context of sports club volunteering where board members assume leadership roles. These leaders rely heavily on volunteers for critical functions such as operational support (e.g. administration or maintenance) and coaching, which are performed with minimal or no financial compensation (De Clerck, Willem, Aelterman, & Haerens, 2021a). To investigate how sports club leaders can foster cohesion among these volunteers, we adopt a comprehensive, multi-level approach, offering valuable insights applicable to organizational settings beyond sports clubs.

Before exploring the relation between leadership and group cohesion, we aim to establish a thorough understanding of the concept of cohesion and its multifaceted dimensions.

Literature review

Group cohesion: a multi-dimensional construct

Group cohesion is widely recognized as a dynamic process in which members stick together and remain united in achieving their instrumental goals while fulfilling their emotional needs (Doherty & Carron, 2003). It is a multidimensional construct, with social and task cohesion representing its most distinctive dimensions (Castaño et al., 2013; Grossman et al., 2022; Horsham, Abrams, Davies, & Lalot, 2024). Social cohesion refers to the degree to which group members appreciate and enjoy each other’s company, fostering strong emotional connections and a deep sense of belonging to the group (Horsham et al., 2024). Task cohesion on the other hand involves the level of commitment and motivation group members have toward achieving their collective goals, focusing on efficient collaboration, effective communication, and coordinated efforts (Grossman et al., 2022).

Achieving a healthy equilibrium between social and task cohesion presents its own unique challenges (Grossman et al., 2022; Picazo, Gamero, Zornoza, & Peiró, 2015). At times, groups may find themselves overly absorbed in social interactions and the maintenance of harmony, inadvertently sidelining progress on tasks. Conversely, a relentless focus on tasks within the group, devoid of attention to nurturing social bonds, can result in feelings of isolation or alienation among group members. Thus, groups that actively strive to balance social and task cohesion are poised to cultivate favorable group dynamics, enhance performance, and achieve their objectives effectively.

Research has supported the importance of both dimensions. Within the specific volunteering setting, which is central to this study, quantitative studies have demonstrated that volunteers (including those in sports clubs) experiencing high levels of social cohesion (Doherty & Carron, 2003; Horsham et al., 2024) and task cohesion (Doherty & Carron, 2003: De Clerck et al., 2021a) among their peers were more effective, satisfied and willing to stay. Similar findings have emerged from meta-analyses across other domains, including work teams in for-profit organizations, military teams, and sports teams (Castaño et al., 2013; Grossman et al., 2022). Qualitative studies, particularly conducted within the context of sports teams, have reinforced the significance of striking a healthy balance between social and task cohesion, revealing that an excessive emphasis on task cohesion can escalate performance pressure, while an overemphasis on social cohesion can induce conformity pressures (Eys & Brawley, 2018).

Leadership and cohesion

To attain an ideal equilibrium between social and task cohesion, proficient leadership is essential (Anderson & Sun, 2017). Leaders can navigate this balance by using both relation-oriented and task-oriented leadership approaches, thereby cultivating an environment that encourages positive social interaction and ensures effective task completion, respectively.

While relatively underexplored in the context of volunteer teams (see De Clerck et al., 2021a), the significance of leadership in shaping group cohesion has garnered considerable attention in studies across various domains. Research suggests that leadership is particularly relevant in shaping social cohesion within military teams (Fors Brandebo, Börjesson, & Hilmarsson, 2022), task cohesion within work teams in for-profit organizations (Van der Voet & Steijn, 2021), and both social and task cohesion in sports teams (Kim & Cruz, 2016).

However, most of these studies have not consistently examined the contingent impact of relationship- and task-oriented leadership approaches on social and task cohesion, which is essential for understanding the role of leadership in group cohesion (Anderson & Sun, 2017). Additionally, research has traditionally viewed cohesion as an individual experience, whereas it is increasingly acknowledged as a collective experience shared among group members (Forsyth, 2021). Investigating cohesion as a collective phenomenon requires a multi-level design (Forsyth, 2021). This approach is crucial to ascertain whether teams can be considered “true groups”. True groups are defined as interacting social entities that are united by a shared sense of cohesion. This is evident when there is significant variance in group cohesion at the group level. Moreover, it allows exploration of whether leadership at the group level, conceptualized as the team climate fostered by leaders through consistent behaviors toward their followers, can shape these collective perceptions of group cohesion.

Examining the role of leadership in volunteer group cohesion: a multi-level perspective using Self-determination theory

Tapping into these gaps in the existing literature on leadership and cohesion, the central aim of our study is to obtain a comprehensive, multi-level understanding of the role of leadership in group cohesion, with a particular focus on voluntary sports clubs. To achieve this, we use Self-determination theory (SDT; Deci & Ryan, 2000; Deci, Olafsen, & Ryan, 2017) to provide a deeper insight into how relationship- and task-oriented leadership styles relate to volunteers' experiences of social and task cohesion in sports clubs. SDT is a prominent meta-theory on human motivation that distinguishes between two distinctive motivating leadership styles: an autonomy–supportive (i.e. relationship-oriented) and a structuring (i.e. task-oriented) leadership style (De Clerck, Aelterman, Haerens, & Willem, 2021b: Ryan & Deci, 2017). Autonomy–supportive leaders embody an open and flexible attitude toward volunteers' interests, preferences and desires. For volunteers, such an approach is deemed particularly relevant as they are more likely to willingly invest their time and energy when they feel a sense of ownership and empowerment, which in turn enhances their motivation and well-being within the organization (Forner et al., 2020, 2023).

In addition to autonomy support, SDT highlights the significance of providing structure (Ryan & Deci, 2017). Leaders achieve this by aligning activities with volunteers’ competencies and offering assistance. Even in the flexible context of volunteering, structuring activities is essential to ensure volunteers function optimally, work effectively, and accomplish their goals.

We rely on multi-level analyses to delve into the role of autonomy–supportive and structuring leadership styles in group cohesion. More precisely, we study this relation at the group level, investigating whether leaders’ consistent display of autonomy–supportive and structuring behaviors toward volunteers relate to their shared perceptions of group cohesion (Forsyth, 2021). Our hypotheses are as follows:

First, we posit that volunteers, despite often operating within looser structures compared to work teams in the for-profit context, constitute true groups who share a collective perception of group cohesiveness (Burke et al., 2005). These shared perceptions can emerge through both formal (e.g. meetings) and informal (e.g. social gatherings) interactions within the organization (i.e. H1; Doherty & Carron, 2003).

Second, we posit that nonprofit leaders, even though they operate in environments with less hierarchical power compared to for-profit contexts, can influence collective perceptions of group cohesion (De Clerck et al., 2021a). Building on the importance of nurturing volunteers’ autonomy (Forner et al., 2020, 2023), we argue that leaders who consistently empower volunteers by providing choices, justifying decisions, and valuing their perspectives are primed to cultivate cohesion within volunteer teams. We anticipate that this approach will predominantly bolster social cohesion, as such empowered volunteers are more inclined to collaborate and work harmoniously toward team objectives (i.e. H2a).

Additionally, recognizing the necessity for nonprofit leaders to provide volunteer groups with the requisite structure (Ryan & Deci, 2017), we predict that when these leaders offer clear instructions, oversee their implementation and consistently provide positive feedback, they will significantly enhance group cohesion. We anticipate that this approach will particularly strengthen task cohesion, as volunteers operating within a clear and structured framework are likely to exhibit heightened commitment to group tasks and increased motivation to pursue shared objectives (i.e. H2b).

Method

Sample

Data for our research were collected from nonprofit sports clubs in Flanders, the Dutch-speaking part of Belgium. In Belgium, the sports sector holds a substantial position within the volunteering landscape, encompassing 18.2% of all volunteering activities (Hustinx & Dudal, 2020). This statistic translates to over 735.000 individuals, constituting 7.8% of the population of 15 years and above, actively engaging as volunteers within the Belgian sector. On a global scale, the sports sector ranks among the top five volunteering sectors, underscoring its widespread importance and impact [United Nations Volunteers Programme (UNV), 2021].

To recruit sports clubs for our study, we sent a call to participate to the sports club leaders (i.e. board members) through the Flemish Sports Federation, the umbrella federation of all Flemish sports federations. To amplify outreach across the sports sector, the Flemish Sports Federation included our invitation to participate in their newsletter. Upon their agreement to participate, board members facilitated the dissemination of an online questionnaire among the club’s volunteers. To guarantee widespread exposure within the sports clubs, this questionnaire was prominently featured in the club’s newsletter and shared across its various social media platforms. Throughout the process, board members were assured of the confidentiality of the data collected, fostering trust and commitment to the study’s objectives. Participating volunteers were dedicated to regular service, taking on various volunteering responsibilities such as coaching a team, providing administrative support, or maintaining the fields. They were asked to assess their experiences of social and task cohesion among their peers within the organization, along with providing insight into their perceptions of the board’s leadership.

In each participating sports club, a minimum of three (regular) volunteers completed the questionnaire (Ling et al., 2016). The number of respondents within the sports clubs ranged from 3 to 31 volunteers. Our study involved 557 volunteers nested within 52 sports clubs (Maas & Hox, 2005). Of these volunteers, 58% were male, which aligns with the broader demographic of regular volunteers, where men constitute 54% of the population [United Nations Volunteers Programme (UNV), 2021]. The average age of our participants was 40.77 years (SD = 13.52), with an average of 6.37 years of volunteering experience within their respective sports clubs (SD = 8.66). A minority (25%) were young adults aged between 15 and 29 years. This demographic distribution is consistent with existing research, which indicates that older adults are more inclined to engage in regular volunteering activities [United Nations Volunteers Programme (UNV), 2021].

To account for sampling error, we used a 95% confidence level with a critical t-value of 1.97 (555 df), meaning we are 95% confident that the true population parameter lies within plus or minus 1.97 standard deviations of our sample estimates (Hair, 2009).

Measures

Motivating leadership styles. To assess the board’s autonomy–supportive and structuring leadership styles, we relied on a validated situation-based questionnaire developed by De Clerck et al. (2021a, 2021b). This questionnaire describes seven specific management situations volunteers may encounter within the sports club. For each situation, volunteers were asked to rate the board’s autonomy–supportive (8 items) and/or structuring leadership style (5 items) on a seven-point Likert scale, ranging from 1 (does not describe my board at all) to 7 (does describe my board extremely well). For instance, volunteers were presented with a situation in which the board organized a meeting with volunteers to evaluate the sports club's activities. Subsequently, volunteers were asked to what extent the board creates opportunities for volunteers to provide input during the meeting (autonomy support) and clarify the purpose of the meeting so that they know what to expect (structure). Internal consistencies of the scales, as assessed by Cronbach’s alpha (α), were excellent, with values of 0.91 for autonomy support and 0.84 for structure.

Social and task cohesion. To measure the volunteers’ perceptions of social and task cohesion among peer volunteers, we used the 18-item Group Environment Questionnaire developed by Doherty and Carron (2003). The survey was designed to evaluate social and task cohesion within volunteer sports executive committees. For this study, we substituted the terms “Committee work” with “Volunteer work” and “’(Members of the) Committee” with “Volunteers”. Volunteers were explained to the respondents as peers who also regularly dedicated their time to the sports club. The social cohesion scale assessed the perceived social integration of volunteers within the club and their affinity toward the social aspects of volunteer work (9 items; α = 0.89; e.g. “There are good relationships among volunteers”; “I enjoy socializing with other volunteers”). The task cohesion scale, on the other hand, evaluated the perceived task integration of volunteers within the club and their attraction to the volunteer tasks (9 items; α = 0.92; e.g. “Volunteers are united in trying to reach goals”, “I feel a sense of accomplishment from my volunteer work”). Volunteers rated each item on a seven-point Likert scale, ranging from 1 (does not describe me/the club at all) to 7 (does describe me/the club extremely well).

Data analysis

First, descriptives of the study variables (i.e. autonomy support, structure, social cohesion and task cohesion) and (multi-level) correlations between these variables were calculated. Next, multi-level regression analyses were used to test the relation between leadership styles and volunteer group cohesion at the group level. Before conducting multi-level regression analyses at the group level, we assessed the group-level properties of the study variables among all volunteers by calculating the within-group and between-group variance for each of the study variables using ICC(1) and ICC(2) measures (Woehr, Loignon, Schmidt, Loughry, & Ohland, 2015). ICC(1) represents the proportion of between-group variance in the total variance, illustrating how well a volunteer’s perception represents the group (Burke et al., 2005). ICC(2) adjusts for group size, estimating the reliability of group means. Calculating significance using an analysis of variance (ANOVA) framework (Biemann, Cole, & Voelpel, 2012), the results showed significant variance at the group level, with ICC(1) values ranging from 0.13 (social cohesion) to 0.18 (autonomy support) and ICC(2) values ranging from 0.62 (social cohesion) to 0.70 (autonomy support). As an additional analysis, we assessed the within-group agreement (r_wg(j);Biemann et al., 2012), comparing the observed variance among volunteer team members to the expected variance under the null hypothesis of no agreement (Biemann et al., 2012; Woehr et al., 2015). In our study, values varied from 0.70 (structure) to 0.89 (task cohesion), confirming high homogeneity in volunteer team members’ perceptions of leadership and cohesion (Woehr et al., 2015). Overall, these statistics support conducting multi-level analyses at the group level and aggregating individual scores.

Multilevel regression analyses consisted of the following four steps. First, we estimated the null or intercept-only model, splitting the variance of volunteers’ perceptions of social cohesion (Model 0a) and task cohesion (Model 0b) into within- and between-variance. In the second step, to control for confounding variables in our model, we included age, gender and years of volunteering as covariates in the models predicting social cohesion (Model 1a) and task cohesion (Model 1b). Age and gender were included as they may play a role in regular volunteering [United Nations Volunteers Programme (UNV), 2021], while previous research also pointed to the role of years of volunteering in the behaviors and attitudes of volunteers (e.g. Doherty & Carron, 2003). Moreover, we examined whether volunteer role could serve as a confounding variable affecting cohesion perceptions (see De Clerck et al., 2021a). Specifically, we examined differences in perceptions of cohesion between volunteers providing operational support (e.g. administration, maintenance; n = 287) and those involved in sports-technical support (e.g. coaching; n = 270). T-tests revealed differences between these subgroups in terms of task cohesion, with volunteers providing sports-technical support experiencing higher levels of task cohesion (M = 5.43, SD = 1.03) than volunteers providing operational support (M = 5.21; SD = 1.14; t = 2.46 (554), p < 0.05). Therefore, we included volunteer role as a covariate in the model predicting task cohesion (Model 1b). It is important to note that despite these differences in average scores, there is some overlap in individual perceptions of task cohesion across these subgroups as shown by the significant ICC’s and high within-group agreement. No social cohesion differences between these subgroups were found (t(=0.33 (554), p = 0.74), so, for reasons of parsimony, the volunteer role was not included in the model predicting social cohesion (Model 1a). In the third step, to control for the role of individual perceptions of leadership, we entered (group-mean centered) individual-level autonomy support and structure as predictors of social cohesion in Model 2a and task cohesion in Model 2b. Finally, with the central aim of our study in mind, we proceeded to the fourth step. Here, we introduced (grand-mean centered) group-level autonomy support and structure as predictors of social cohesion in Model 3a and task cohesion in Model 3b. All multi-level analyses were conducted using MLWiN version 3.06 (Browne & Rasbash, 2009).

Results

Descriptives of the study variables and (multi-level) correlations are displayed in Table 1.

Table 2 shows the multi-level model predicting social cohesion. The random parts of the null model showed that the variances at both the individual and group levels differed significantly from 0. None of the covariates that were inserted in the model were significantly related to social cohesion. Individual-level autonomy support [B = 0.39, S.E. = 0.05, χ² (1) = 53.83, p < 0.001] and structure [B = 0.12, S.E. = 0.05, χ² (1) = 5.65, p < 0.05] related significantly to social cohesion. At the group level (the focus of this study), only autonomy support [B = 0.52, S.E. = 0.13, χ² (1) = 17.18, p < 0.001] related significantly to social cohesion.

Similar results were obtained for the multi-level model predicting task cohesion (Table 3). The random parts of the null model revealed that the variances at both the individual and group levels were significantly different from zero. None of the covariates that were entered into the model were significantly connected to task cohesion. At the individual level, autonomy support [B = 0.40, S.E. = 0.05, χ² (1) = 69.29, p < 0.001] and structure [B = 0.22, S.E. = 0.05, χ² (1) = 19.90, p < 0.001] related significantly to task cohesion. At the group level, only autonomy support [B = 0.69, S.E. = 0.11, χ² (1) = 38.80, p < 0.001] was significantly related to task cohesion.

Discussion

To enhance workforce effectiveness and encourage long-term commitment, leaders must cultivate group cohesion within their organization (Doherty & Carron, 2003; Castaño et al., 2013). Given that group cohesion primarily operates at the group level (Forsyth, 2021), this study provides a multi-level perspective on how leadership fosters group cohesion among sports club volunteers, providing insights relevant to all sectors involving team-based work environments.

Volunteers constitute “true groups” within nonprofit organizations

Multi-level results first supported the assumption of previous volunteer cohesion research (e.g. Doherty & Carron, 2003) which suggested, albeit without empirical confirmation, that regular volunteers constitute “true groups” within nonprofit organizations. Indeed, the significant between-group variance in both social and task cohesion, along with the high within-group agreement, showed that volunteers within an organization are interdependent in a meaningful way, demonstrating shared beliefs around group cohesiveness (Burke et al., 2005). This finding confirms our H1 that volunteer teams, like conventional work teams in the for-profit sector (Grossman et al., 2022), engage in interpersonal interactions with their peers. These interactions influence their collective sense of harmony and unity around shared goals (Forsyth, 2021). Notably, existing literature suggests that cohesive groups, whether composed of employees or volunteers, that are harmonious and united in pursuing common goals tend to be more successful, effective, satisfied and committed to staying (Castaño et al., 2013; Doherty & Carron, 2003).

Autonomy support is particularly important for fostering cohesion within volunteer groups.

Multi-level regression analyses further demonstrated that nonprofit leaders, akin to their counterparts in more hierarchical structured for-profit organizations (Anderson & Sun, 2017), can foster cohesive groups. Consistent with previous research (e.g. Kim & Cruz, 2016), we found a strong relation between nonprofit leaders’ autonomy support and social cohesion at the group level. This supports our H2a, indicating that a group climate where leaders respect the volunteers’ choices and perspectives, provide clear rationales for decisions, and give control over the development of the group strategies, stimulates volunteers’ emotional attachment to the group and a sense of harmonious collaboration (Castaño et al., 2013; Horsham et al., 2024). This also aligns with previous research in the for-profit context, highlighting autonomy support as a critical motivating leadership style, significantly contributing to employees’ well-being and optimal functioning (e.g. Deci et al., 2017).

Surprisingly, autonomy support also played a pivotal role in task cohesion, even more so than a structuring style. Thus, to enhance task cohesion, leaders should respect the volunteer group’s autonomy as they will be more likely to be committed to the group’s tasks, take initiative, demonstrate creativity, and be motivated to contribute their unique skills to achieve common goals. This finding contrasts with our H2b and the prevailing literature in the for-profit sector, which underscores the importance of the provision of structure for enhancing employee cohesion (Van der Voet & Steijn, 2021). This divergence may stem from the fact that volunteers, who freely donate their time and effort to the organization, often prioritize collaborative, self-directed, and interdependent work over an emphasis on structure. This autonomy stimulates their investment in the tasks at hand, which enhances the effectiveness of the group’s efforts (Castaño et al., 2013; Horsham et al., 2024).

Notably, also structuring leadership correlated significantly with task cohesion (see the multilevel correlation matrix – Table 1), indicating that providing clear directions and establishing well-defined roles and responsibilities contribute to some extent to cohesion among volunteers. However, this relation was no longer significant in the regression model, indicating that autonomy support had an even stronger impact on task cohesion.

Although not the primary focus of this study, multi-level findings also revealed that autonomy support and structure were related to social and task cohesion at the individual level. While individual perceptions of cohesion are important for outcomes such as satisfaction and work effort (e.g. Doherty & Carron, 2003), it is the shared, group-level perceptions that have a greater impact on overall performance and group cohesion (Castaño et al., 2013; Horsham et al., 2024). This implies that leaders who consistently use motivating practices create a more effective team environment than those who only occasionally do so.

Harnessing the bigger picture: practical implications of the study findings

The findings of our study are broadly applicable to the field of management, emphasizing the critical importance of fostering an autonomy–supportive environment. This approach not only enhances the effectiveness of volunteer groups, as demonstrated in our study but can also be applied to any collaborative setting, including employees in the for-profit sector (Deci et al., 2017). According to SDT, leaders can cultivate autonomy by employing strategies such as involving their followers in decision-making, actively listening to their viewpoints, and consistently providing clear rationales (Deci et al., 2017). When leaders implement these strategies, they fulfill followers' basic psychological needs for autonomy, competence, and relatedness, resulting in enhanced functioning and well-being. Specifically, when individuals within the organization feel supported in their autonomy, they not only experience a sense of ownership (i.e. autonomy) but also enhanced effectiveness (i.e. competence) and stronger connections to the organization (i.e. relatedness). Thus, extending autonomy support to all followers, rather than focusing solely on individualized support, is crucial. This inclusive approach prevents feelings of isolation and fosters a collective sense of purpose, ultimately enhancing group cohesion.

While creating an autonomy–supportive environment is crucial for fulfilling basic psychological needs, it is important to recognize that different groups and individuals have diverse needs and motivations (Forner et al., 2020). Some people thrive under autonomy–supportive conditions, while others may require more structure and guidance to feel competent and effective. Additionally, some may need greater recognition of their work to fulfill their need for relatedness. The nature of the task and specific context can also influence which leadership style is most effective. For creative tasks, autonomy may be more beneficial, whereas tasks requiring precision and coordination might necessitate a more structured approach. Therefore, it is essential for nonprofit leaders, as well as leaders in other organizational contexts, to be adaptive and responsive to the dynamic needs of their teams. This necessitates flexibility and the ability to use a mix of leadership styles tailored to the specific demands of tasks and the varying needs of team members. By understanding and addressing these needs, leaders can enhance members’ engagement, satisfaction, and overall group effectiveness.

Limitations and future directions

The data of our study was collected at a single point in time, suggesting the need for longitudinal research to establish causality. Future studies could also benefit from using a (multilevel) structural equation modeling approach to examine the relationships among leadership, cohesion, and outcomes. Additionally, exploring the relationship between a paid executive and the volunteer board would provide a more comprehensive understanding of how various leadership factors contribute to group functioning. Finally, we discussed the implications of our study for the broader field of management. Future research could build on these insights to further develop and refine these implications.

Conclusion

This study represents an important step toward a better understanding of the role of leadership in group cohesion. Focusing on the voluntary sports sector, the findings revealed a strong relation between autonomy support and group cohesion. This suggests that leaders should consistently rely on autonomy–supportive practices such as providing choices to team members, empathizing with their perspective, and offering explanations when choices are limited to create cohesive teams.

Table 1.

Descriptives and (multi-level) correlations between study variables

Variables	N	M	SD	1	2	3	4
1. Autonomy support	557	4.54	1.27
2. Structure	556	4.46	1.28	0.70***
3. Social cohesion	556	5.29	1.09	0.59***/0.33***	0.52***/0.28***
4. Task cohesion	556	5.31	1.09	0.69***/0.40***	0.63***/0.34***	0.75***

Notes:

Correlations between leadership (predictor) and cohesion (outcome) are subdivided into the correlations at the individual level (left) and correlations at the group level (right); *p < 0.05; **p < 0.01; ***p < 0.001

Source: Authors’ own work

Table 2.

Multi-level model predicting social cohesion (n = 556)

		Social cohesion
	Model 0a	Model 1a		Model 2a		Model 3a
	B (S.E.)	B (S.E.)	β	B (S.E.)	β	B (S.E.)	β
Fixed part
Intercept	5.30 (0.07)	5.65 (0.22)		5.43 (0.19)		5.38 (0.18)
Gender		0.01 (0.10)	0.01	0.06 (0.08)	0.06	0.08 (0.08)	0.08
Age		−0.01 (0.00)	−0.13	−0.01 (0.00)	−0.09	−0.01 (0.00)	−0.09
Years volunteers		0.01 (0.01)	0.01	0.01 (0.01)	0.01	0.01 (0.01)	0.01
Individual-level autonomy support				0.39 (0.05)***	0.50	0.39 (0.05)***	0.50
Individual-level structure				0.12 (0.05)*	0.16	0.13 (0.05)*	0.16
Group-level autonomy support						0.52 (0.13)***	0.34
Group-level structure						0.03 (0.13)	0.02
Random part
Group-level variance	0.15 (0.05)**	0.15 (0.05)**		0.19 (0.05)**		0.04 (0.02)
Individual-level variance	1.03 (0.07)***	1.02 (0.06)***		0.69 (0.04)***		0.70 (0.04)***
Deviance test model	1641.56	1634.39		1438.86		1400.05
χ2 (df)	26.95 (1)***	7.17 (3)		195.53 (2)***		38.81 (2)***

Notes:

*p < 0.05; **p < 0.01; ***p < 0.001

Source: Authors’ own work

Table 3.

Multi-level model predicting task cohesion (n = 556)

		Task cohesion
	Model 0b	Model 1b		Model 2b		Model 3b
	B (S.E.)	B (S.E.)	β	B (S.E.)	β	B (S.E.)	β
Fixed part
Intercept	5.32 (0.08)	5.69 (0.21)		5.31 (0.18)		5.29 (0.55)
Gender		0.00 (0.10)	0.00	0.01 (0.08)	0.02	0.02 (0.07)	0.02
Age		0.00 (0.00)	−0.04	0.00 (0.00)	−0.03	0.00 (0.00)	−0.03
Years volunteers		0.01 (0.01)	0.05	0.01 (0.01)	0.07	0.01 (0.01)	0.08
Volunteer role		−0.20 (0.10)	−0.20	0.04 (0.08)	0.03	0.03 (0.08)	0.03
Individual-level autonomy support				0.40 (0.05)***	0.51	0.40 (0.05)***	0.51
Individual-level structure				0.22 (0.05)***	0.27	0.22 (0.05)***	0.27
Group-level autonomy support						0.69 (0.11)***	0.45
Group-level structure						−0.01 (0.12)	0.00
Random part
Group-level variance	0.19 (0.06)***	0.19 (0.06)***		0.25 (0.06)***		0.03 (0.02)
Individual-level variance	1.03 (0.07)***	1.02 (0.06)***		0.57 (0.04)***		0.56 (0.04)***
Deviance test model	1646.01	1638.93		1345.46		1278.93
χ2 (df)	71.41 (1)***	7.08 (4)		293.47 (2)***		66.53 (2)***

Notes:

*p < 0.05; **p < 0.01; ***p < 0.001

Source: Authors’ own work