The power of believing you can
About the trainer-player relationships within soccer and the individual development and performance of players
University of Amsterdam MSc Business Administration | Leadership & Management Author: J.A.B. Bartels | 11281200 Supervisor: Claudia Buengeler Course: Master Thesis Date: 17 August, 2017 Final Version
Statement of originality
This document is written by Student Jip Bartels who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Acknowledgements
This Master Thesis was conducted as the final assignment for my MSc Business Administration with Leadership and Management as track. The Thesis is based upon my personal interests and the gathered knowledge and skills during the courses of this Master. Finalizing this Thesis was truly a journey with ups and downs. Moving to Melbourne a couple of weeks before the first deadline has unfortunately led to some delay, as my amount of available time, energy and focus was reduced. Looking back I can conclude that writing a Thesis took more time and effort than expected. However, this means that I have learned a lot and that my contentment is enormous.
Completing the Thesis would not have been possible without the outstanding help and guidance of my supervisor, Claudia Buengeler, who has shown to be a flexible, friendly, compassionate and understanding supervisor. Thank you and I wish you all the best in Kiel.
Furthermore I would like to thank my family and student colleagues for their mental support, advice and social distraction when needed.
At last I want to thank the participants of this research for filling in all the paper surveys. Their contribution made it possible to conduct this research.
Table of content
Abstract p. 6
1. Introduction p. 7
2. Theoretical Background p. 11
2.1. LMX differentiation p. 11
2.1.1. LMX differentiation perceived as variety p. 13
2.2. LMX differentiation in relation to self-efficacy p. 14
2.3. Individual LMX as moderator p. 18
2.4. Self-efficacy in relation to individual development and performance p. 21
2.5. Self-efficacy as mediator p. 23 3. Method p. 25 3.1. Sampling technique p. 25 3.1.1. Demographic characteristics p. 26 3.2. Measurement procedure p. 26 3.2.1. Additional analyses p. 27 3.3. Measures p. 27 3.3.1. LMX differentiation p. 27 3.3.2. Individual LMX p. 28 3.3.3. Self-efficacy p. 28
3.3.4. Individual Development and Performance p. 28
3.4. Control variables p. 29 4. Results p. 30 4.1. Correlations p. 30 4.2. Hypotheses results p. 31 4.2.1. Results Hypothesis 1 p. 31 4.2.2. Results Hypothesis 2 p. 32 4.2.3. Results Hypothesis 3 p. 34
5. Discussion p. 35
5.1. Findings p. 35
5.1.1. Findings Hypothesis 1 p. 36
5.1.2. Findings Hypothesis 2 p. 38
5.1.3. Findings Hypothesis 3 p. 39
5.1.4. Findings additional analyses p. 41
5.2. Practical implications p. 42 5.3. Limitations p. 43 5.4. Future research p. 44 6. Conclusion p. 45 7. References p. 47 8. Appendices p. 54 Appendix A p. 54 Appendix B p. 74 Appendix C p. 75 Appendix D p. 76 Appendix E p. 76 Appendix F p. 77 Appendix G p. 79 Appendix H p. 80
Abstract
The first aim of this research is to make a theoretical contribution to the existing literature about LMX relationships, especially in the context of soccer teams. The idea is that an individual and differentiated approach (LMX differentiation) by a trainer increases the players’ level of efficacy, whereby the relation between LMX differentiation and self-efficacy is moderated by the players’ perceived quality of the trainer-player relationship (individual LMX). Subsequently, the individual level of self-efficacy mediates for the level of development and performance of the individual player. Within this quantitative research a correlational and cross-sectional research design is used, whereby a total of 149 participants were selected by a non-probability quota sample. The results of this research supported a positive relationship between self-efficacy and individual development and performance. However, no results were obtained that supported the moderation of individual LMX or the mediating role of self-efficacy. Despite these unexpected findings, the present results contribute to the field of LMX relationships, especially in the sports context. Moreover, practical implications regarding more focus on the role of players’ self-efficacy within youth training programmes are given. In order to check the validity and generalizability of this research and to gain more knowledge about the investigated variables, future research within the field of LMX relationships should focus more on (sport) teams outside of soccer with other samples.
Key words: LMX differentiation, self-efficacy, LMX quality, individual development,
1. Introduction
“The nature of the athlete–coach relationship has an important role to play in the athlete’s development both as a performer and as a person. It is recommended that coach education programmes provide information that will assist coaches to develop effective relationships with their athletes.” (Jowett & Cockerill, 2003, p. 313).
As this quote illustrates, for athletes to have an optimal development and performance level, it is considered essential to have a trainer with whom they have an effective and high-quality relationship. Nonetheless, there is not enough focus on effective trainer-player relationships within soccer according to Van Breukelen, van Der Leeden, Wesselius, and Hoes (2012), and Vella, Oades and Crowe (2013). Both studies stated that trainers focus too much on the team performance and that the development of the individual athlete comes second. In the world of soccer it is still normal and acceptable for trainer-athlete relationships to be hierarchical and one-sided. Hereby the trainer is responsible for the decision making, often without the involvement of the individual player. This is reinforced because, in comparison with business managers, trainers have more autonomy and authority to make decisions (Høigaard, Jones & Peters, 2008). Jowett and Cockerill (2003) even stated that one of the most heard reasons for stagnant development and failing performances within sports is an ineffective and low-quality relationship between trainer and athlete.
In order to be effective, the trainer-athlete relationship should be adapted to the needs of the individual athlete (Cranmer & Myers, 2015; Høigaard et al., 2008; Jowett & Cockerill, 2003). These studies highlighted that a trainer-athlete relationship needs to be reciprocal in order to be of high-quality and create an optimal climate whereby the athletes perceive a high level of self-efficacy. In return this high level of self-efficacy enables them to learn and develop themselves. So, there should be some congruence between the actual trainer-athlete relationship and the preferred relationship by the players in order to maximize their self-efficacy and their individual development as well as performance.
Within literature, leader-member exchange (LMX) as described by Graen and Uhl-Bien (1995) will be relevant for this study. Namely, LMX focuses on the leader-follower relationships within a team and entails that a leader develops differential relationships with his or her individual followers. In order to gain more insight into the dyadic trainer-player relationships within teams such as the ones studied, LMX differentiation, as described by Boies and Howell (2006) and Henderson, Liden, Glibkowski and Chaudhry (2009), will be used to get more focus within this research. According to them, a focus on LMX
differentiation explicitly acknowledges that dyadic leader-follower relationships vary within teams. LMX differentiation is conceptualized as the amount of variability in LMX relationships perceived by team members (Boies & Howell, 2006; Henderson et al., 2009).
Applying organizational LMX theory to the context of sport teams is tenable because organizations and sport teams have similarities. For example, they are comprised of a collection of individuals (e.g. trainers and players) who share collective goals (e.g. winning and developing skills) and they have identifiable boundaries. Besides, each individual team member is assigned different roles, positions and tasks, based on skills and experience. In addition, despite these individual differences, team members must coordinate their efforts and actions to achieve team-level goals (Cranmer & Myers, 2015).
Nevertheless, the trainer-player relationships within soccer also have their differences in comparison to the common leader-follower relationships within business management context. For example, trainers have more autonomy to make decisions compared to their organizational counterparts and they more often show a uniform leadership style that is consistently used among all team players, while LMX-based relationships are more common in organizational settings (Cranmer & Myers, 2015; Høigaard et al., 2008). Furthermore, according to Cranmer (2016), most sport teams exist of unisex members, differ in interdependence between players, and success is based on the place in the league table and a direct head-to-head comparison competition within and between other teams. This results in differentiation of players by skill set and given amount of playing time (starters and nonstarters). Additionally, due to the high intensity of both cooperation and competition between and within sport teams, it is suggested that the need for high-quality trainer-athlete relationships is highly important (Cranmer, 2016).
Furthermore, LMX differentiation has become an important framework in the last decade for building individual, high-quality and effective LMX relationships within team settings (Henderson et al., 2009). This is because the biggest part of previous research about trainer-athlete relationships came short and has shown conflicting results. For example, researchers assumed that trainers’ leadership is consistent across a team (Cranmer, 2016; Cranmer & Myers, 2015). According to them, these previous studies failed to recognize and understand that trainers develop unique and differentiated relationships with individual athletes and that the athletes’ perception of this trainer-athlete relationship (individual LMX) influences the athletes’ performance.
Besides, whereas the positive relation between self-efficacy and individual performance is well known (Feltz & Lirgg, 2001; Gernigon & Delloye, 2003; Walumbwa & Hartnell, 2011),
the relations between LMX differentiation, individual LMX, self-efficacy and individual development as well as performance are complex and relatively under researched (Hu & Liden, 2013; Liao, Liu & Loi, 2010; Walumbwa & Hartnell, 2011). Specifically, Hu and Liden (2013) state that there is not much known about the psychological mechanisms through which LMX constructs affects outcomes at the individual level. As LMX differentiation has only recently arisen as a framework (Henderson et al., 2009), prior research has not yet clarified how LMX differentiation operates in relation to self-efficacy and individual performance. Moreover, as previously mentioned, trainer-athlete relationships differ from leader-follower relationships within an organizational context. This means that the use of LMX differentiation as a theoretical perspective within the context of sports also provides new insights into how dyadic LMX relationships influence the perceptions of relationships with a trainer (Cranmer & Myers, 2015; Jowett & Cockerill, 2003).
This brings us to the main aim of this research, which is to make a theoretical contribution by exploring the relationships between LMX differentiation (independent variable), Individual
LMX (moderator), Self-efficacy (mediator) and the players’ Individual Development and Performance (dependent variables) within soccer. This is captured in the research question
(see Figure 1): ‘Does LMX differentiation benefit individual players’ self-efficacy, development and performance, and does this depend on individual players’ relationship quality with their trainer?’
By examining the role of individual LMX in relation to LMX differentiation and self-efficacy, this research contributes by creating more clarity and consensus about the potential positive and negative effects of LMX differentiation on followers’ self-efficacy. For example, a potential negative effect is described by Hooper and Martin (2008), who state that LMX differentiation counters the principles of equity and fairness, which reduces the perceived quality of the individual LMX level. They found that LMX differentiation was related to a negative individual LMX and that this has a negative impact on the followers’ self-efficacy and performances.
Contrary to this potential disadvantage, LMX differentiation has its potential advantages according to Walumbwa and Hartnell (2011). Namely, they suggest that as a consequence of differentiated leadership, athletes may feel an increased sense of competence, which is incorporated into their perceived individual LMX and positively related to their self-efficacy. Self-efficacy entails the belief in your own competence to successfully perform a certain task (Bandura, 1986). The creation of self-efficacy occurs especially when leaders provide differentiated treatment, as trainers stimulate the autonomy and self-competence of their
athletes (Walumbwa & Hartnell, 2011). According to many researchers, self-efficacy has a significant positive relationship with the development and performance of individuals (Feltz & Lirgg, 2001; Gernigon & Delloye, 2003; Walumbwa & Hartnell, 2011).
These mentioned potential advantages also suggest that for training programmes of players and trainers it is practically relevant to know how self-efficacy can be stimulated. Namely, the Dutch soccer is getting behind on the rest of Europe and the world. This is because the Royal Dutch Soccer Association (KNVB) and the Dutch government have strict financial regulations. It is a vicious circle; as a result of reduced financial possibilities, Dutch clubs cannot afford to buy the best players and this results in disappointing results in sport and financially attractive European competitions, like the ‘Europa League’ and the ‘Champions League’. This means that in order to regain its leading position and compete with wealthier competitors, the KNVB and the Dutch soccer clubs must distinguish themselves by returning to its roots: an innovative youth academy that produces the best players in the world (Oppenhuisen & van Zoonen, 2006; Van Hilvoorde & Stokvis, 2013).
All in all, the theoretical and practical implications of this study will shed light on the effects of LMX differentiation, especially within soccer. This will be done by giving more insights into the potential positive and negative effects of LMX differentiation on players’ self-efficacy and their development as well as performance.
This study is organized as follows: The next part includes the theoretical background upon this research is built, which will include the presentation of the hypotheses. This will be followed-up by the method section. After the method section, the results and discussion section will be presented. The final part is the conclusion.
Figure 1. The conceptual model underlying this research. LMX Differentiation Individual LMX Self-efficacy Individual Development and Performance
2. Theoretical Background
The exploration of the relationships between LMX differentiation, individual LMX, self-efficacy and individual development as well as performance is based upon an extensive theoretical background of the relevant literature. In this part the main theoretical constructs of this research and the Hypotheses will be presented, explained and discussed.
2.1. LMX differentiation
LMX differentiation is based upon the LMX principles of Graen and Uhl-Bien (1995). According to them, LMX is based on an effective, reciprocal, dyadic and relationship-based approach between leader and follower. Hereby trust, respect, mutual obligations and individual consideration play an important role within this relationship as it is about mutual learning and building a strong bond between leader and follower. The development and quality-status of these LMX relationships is highly dependable of the personal characteristics and behaviours of both leader and follower. High-quality relationships are considered to be very positive for leaders, members and teams in general. This is because all parties profit from the gained access to the beneficial exchanges between them. For those dyadic members whereby high-quality relationship are obtained, exchanges are made on the behavioural and emotional level. Mutual respect, obligations, trust, support and loyalty grow throughout this process. Moreover, they have a long time span of reciprocation (Graen & Uhl-Bien, 1995).
As mentioned earlier, LMX differentiation refers to the extent to which LMX relationships are perceived to vary within a team. According to Henderson et al. (2009) and Le Blanc and González-Romá (2012), LMX differentiation is the major premise of the overarching LMX theory, because this theory states that leaders build different kinds of relationships with their followers. However, research on LMX has mainly focused on the individual- or dyadic level (i.e., one’s relationship quality with a supervisor), without studying leader-follower dyads within their respective workgroup context. The emerging stream of research on LMX differentiation explicitly accounts for this differentiation within a workgroup.
Within this study the idea behind LMX differentiation is that when the leader accommodates the differing individual needs of its followers, this will increase their development and performance. Members will be persuaded to engage in more activities, have more constructive and supportive attitude and take more responsibility. The mutual trust,
respect and obligation between leader and member empowers and motivates both (Henderson et al., 2009).
However, LMX differentiation can also have its disadvantages within a dynamic, high interdependent and complex environment. For example, a disadvantage of LMX differentiation can be that it takes time to create high-quality relationships and that the process is too highly dependent on the leader, as he or she has the responsibility to initiate the development of relationships. Next to this initiating role of the leader, whether LMX differentiation can be effective also depends on the team members’ attitude and abilities (Graen & Uhl-Bien, 1995).
In addition to the team members’ attitude, Hooper and Martin (2008) found attribution biases in cases of differential treatment by the leader. Namely, if oneself is the target of differential treatment, this is perceived as favourable, fair, and positive. However, when another team member is the target of this favourable treatment this is perceived as unfair and negative. Perceiving unfairness leads to lower liking and trust, with negative effects on the perceived LMX differentiation as result. These perceived inequalities may lead to in- and out-groups and consequently to a decreased perception of individual LMX quality (Hooper & Martin, 2008).
Adding to this discussion, Kauppila (2015) has a more positive view on the effects of LMX differentiation, namely he states that LMX differentiation does not always have to be ‘good’ for in-group members and ‘bad’ for out-group members, or vice versa. Kauppila (2015) provides a more nuanced view by arguing that the effects of LMX differentiation may be dependent on the contextual circumstances. Although Kauppila (2015) agrees that perceptions of inequality tend to negatively relate to LMX differentiation outcomes, he states that this does not completely explain why LMX differentiation has more negative or positive influence on individually high or low LMX members. On the one hand, he proposes that high LMX members may react negatively to LMX differentiation, as this threatens their favoured position and that they can get more obligations or a higher workload. On the other hand, he proposes that low LMX members may react positively to LMX differentiation, as they see possibilities to improve their position.
So, it can be stated that it is important that team members perceive a sense of equitable and fair treatment by their leader. LMX differentiation can have a negative influence on these perceptions, but not necessarily. Thus, LMX differentiation may have its potential advantages, depending on the circumstances and the way it is implemented. When LMX differentiation is implemented in the right way, it may lead to positive outcomes for individual members.
2.1.1. LMX differentiation perceived as variety
Considering the LMX differentiation variable, Harrison and Klein (2007) discussed the concept of diversity. According to them the conception of diversity is ambiguous and therefore they proposed that diversity is not one thing but three things. According to them diversity can be perceived as separation, disparity and variety. Hereby the content, operationalization, and likely consequences of those three things differs. This study transplants the view of variety and disparity as described by Harrison and Klein (2007) for diversity to the LMX differentiation literature, which is tenable as both concepts are closely related (Harris, Li & Kirkman, 2014).
First, separation reflects the differences or disagreements among team member based on their task- or team opinion or values. For example, it reflects the distribution of opposing attitudes and beliefs considering a team relevant issue. Differentiation as separation is mainly found in theory and research of similarity attraction (Byrne, 1971; Clore & Byrne, 1974), social identity and self-categorization (Hogg & Terry, 2000; Tajfel & Turner, 1979) and attraction selection (Schneider, Goldstein & Smith, 1995). These studies mainly posit that greater similarity (reduced separation) leads to higher levels of trust and performance (Harrison & Klein, 2007).
Second, disparity describes differences among team members in valued social assets or resources such as status, decision making authority and privileges over others. Some members of the team are superior to other team members in status and privileges. High disparity means that power, status, or other socially valued resources are unequally distributed among team members (Harrison & Klein, 2007). Differentiation as disparity is mainly found in sociological studies, whereby the focus is often on differences in status, power and payment. Within these studies disparity is often negatively related to performance (Harrison & Klein, 2007; Siegel & Hambrick, 2005).
Third, variety describes differences among team members based on categories. Hereby the primary categories are their unique sources of relevant knowledge and skills. The team member have obtained these unique assets and resources by their education, training or experiences (Harrison & Klein, 2007). Low variety means that team members are redundant, as they don’t add new unique characteristics. Likewise, teams with high variety are more likely to make effective decisions and to have higher levels of performance (Argote & Ingram, 2000; Austin, 2003). Differentiation as variety is not very common in business management research. However, partly similar to this study, McGrath, Berdahl and Arrow (1995) also used differentiation variety to refer to differences in knowledge, skill and ability.
Moreover, Hooper and Martin (2008) used LMX differentiation perceived as variety to examine what the effects are of perceived LMX variability on social harmony in teams.
As described above, differentiation as separation and as disparity are commonly related to lower levels of trust and performance. Hence, in this research differentiation will mainly be conceptualized as variety. This is because this study focuses on a positive relationship between differentiation and performances, whereby the unique knowledge, skills, abilities and personality of the team members are seen as valuable categories which can be enhanced by education and training.
Another incentive to perceive differentiation as variety in this research is because, as mentioned earlier, previous research has not yet provided what is needed to test the hypotheses of this research. In this way this research breaks with traditional differentiation research, which has focused more on differentiation as separation or disparity.
2.2. LMX differentiation in relation to self-efficacy
“Our coach–athlete relationship was different. I knew and could see it from the way the coach treated the other swimmers in the team. One day I asked my coach why he is different towards me. He said that ‘your goals are different from the other swimmers and so you have to work longer and harder’. (Jowett & Cockerill, 2003, p. 321).
As this quote illustrates, by differentiating and providing a personal approach, the trainer can support the individual athlete to gain belief in successfully performing their specialized skills and tasks. Stimulating the athletes’ personal and task related skills enhances their self-efficacy according to Walumbwa, Cropanzano and Goldman, 2011.
Nonetheless, as mentioned before, the current relationships between trainer and player within soccer are often typified as hierarchical and one-sided, whereby the trainer has the power to take all the decisions while the player has little autonomy (Høigaard et al. 2008). Hereby trainers often falsely belief that hierarchical relationships with their players will give them control and success.
As a result there is often low LMX differentiation within soccer teams. When trainers belief that one of their players will perform poorly, they most likely send messages of mistrust, highlight errors and pay less attention to their individual successes. Hereby, the trainer does not acknowledge and act upon the athletes’ personal needs, which restricts the athletes’ autonomy (Walumbwa & Hartnell, 2011). Moreover, these messages of mistrust have a negative effect on the athletes self-efficacy, as a sense of autonomy is one of the most important components of high self-efficacy beliefs (Mageau & Vallerand, 2003).
In order to better understand the relationship between LMX differentiation and self-efficacy, it is important to note that self-efficacy is not the same as self-confidence. Self-confidence is a person’s holistic belief about him- or herself, and how he or she feels about him- or herself. Self-efficacy is specifically about a person’s feelings and beliefs in successfully completing a certain domain, skill or task (Bandura, 1986). In sports, self-confidence is a general term which refers to an athlete’s certainty about his or hers ability to be successful in sport. On the other hand, self-efficacy refers to an athletes’ belief that he or she can be successful in specific tasks and skills under specific conditions. For example, a soccer player may be confident that he or she can play soccer well, but feel less efficacious about his or her ability to successfully perform a specific skill in a challenging soccer situation (Munroe-Chandler, Hall & Fishburne, 2008).
Furthermore, Bandura (1986) stated that self-efficacy entails four important components, which are enactive mastery (personal attainments), previous vicarious experiences, verbal persuasion and psychological arousal. Leaders are capable of providing these four components. Although these components influence the self-efficacy level, the individuals’ cognitive appraisal and internal processing of these components determines the final self-efficacy level. Self-self-efficacy can thus be seen as internal individual judgement of performance capabilities. This is also supported by Gernigon and Delloye (2003), who found that verbal persuasion, like constructive feedback, and prior physical or imaginal successes increases the self-efficacy of athletes.
In addition, Maddux (2009) states that self-efficacy is important because it is about the power of believing you can. Without a strong belief in your own competence, capabilities and possibilities an athlete will take less initiative and will not try to learn new skills, which will result in a stagnant development and performance curve. A trainer should enhance this sense of self-efficacy by focusing on the needs of their athletes and support them to perform well. The trainer should enable and empower the athlete to make decisions about their sport specific tasks, this will enable the athlete to perform effectively in the team (Prussia, Anderson, & Manz, 1998).
So, promoting the self-efficacy goes mainly through the four components described by Bandura (1986) and is based on mutual trust, verbal communication and a reciprocal relationship between leader and follower, which are also important components of LMX differentiation (Salanova, Lorente, Chambel & Martínez, 2011). This is also in line with Walumbwa, Bruce, Avolio and Zhu (2008), who stated that the behaviour of leaders may enhance followers’ self-efficacy through the four components of Bandura (1986).
Next to this, Høigaard et al. (2008), Jowett and Cockerill (2003), and Mageau and Vallerand (2003) state that positive competence feedback and individual consideration are determinative factors for the creation of self-competence, and thus for self-efficacy. Hereby it is important to take into account that the positive feedback includes the promotion of autonomy and competence, as these are fundamental components of self-efficacy. Namely, self-efficacious individuals incorporate this autonomy and sense of competence into their own (Van Mierlo, Rutte, Vermunt, Kompier, & Doorewaard, 2006).
In order to create a high perception of autonomy and self-competence, Mageau and Vallerand (2003) argue that the trainer carries the responsibility for displaying a form of leadership that includes the trainer acknowledging the athletes’ perspective, feelings, personal needs, and to provide them with opportunities of choice and initiative taking without the use of pressure or coercion. This individual consideration is also an important component of self-efficacy and LMX differentiation, which means that this form of leadership can be seen as a part of LMX differentiation and positively contributes to followers’ self-efficacy.
Important to note is that LMX differentiation in relation to self-efficacy and individual development and performance may also has its potential negative effects. As previously mentioned, Hooper and Martin (2008) stated that LMX differentiation counters the principles of equity and consistency and that this has a negative impact on the relationships and performances within the team. According to them, team members can be very sensitive for social comparison and differential treatment. This can be problematic because it influences their perceptions of self-concept and self-efficacy.
Moreover, LMX differentiation can be especially problematic for team members who perceive a relative low individual LMX relationship. This is because they will receive less tangible and intangible resources within their leader-follower dyad. This is called the ‘in- and out-group’ effect by Cranmer and Myers (2015). Hereby the in-group receives more trust, respect, empowerment, information, support, autonomy and positive performance appraisals than the out-group, which are important components of self-efficacy. As a result, in contrast to the out-group, only the in-group has an increased development and performance curve (Cranmer & Myers, 2015).
In order to reduce these potential disadvantages of LMX differentiation, Graen and Uhl-Bien (1995) argue that leaders should have a supportive differentiated style to stimulate followers’ self-efficacy, empowerment and sense of trust. These high-quality LMX relationships provide the followers with effective learning opportunities to develop new skills and gain confidence in their competences. This kind of differentiation makes that the
leader-follower relationships are perceived as more equitable and this increases the sense of fair treatment. According to Graen and Uhl-Bien (1995) making the offer of a high-quality relationship by the leader to all team members, the LMX process may be perceived as more equitable and fair. Whether all offers will actually result in high-quality relationships is unlikely, as creating a high-quality relationship also depends on the function, tasks, attitude and reactions of the team members.
Additionally, as previously stated by Cranmer (2016), trainer-player relationships within soccer are characterized by a high interdependence and a high intensity of cooperation and competition. Therefore there is a strong need for high-quality trainer-athlete relationships Because of the high level of interdependence and differentiation between players, LMX differentiation as leadership style can thus be an advantage in this environment (Liden, Erdogan, Wayne & Sparrowe, 2006). Namely, Liden et al. (2006) stated that teams who are characterized by high task interdependence, differentiating among members and developing individual exchange relationships can be a way to coordinate the individual contributions of the team members. According to them task interdependence refers to the degree to which team members need complementary expertise to work with each other, share tangible and intangible information and communicate with each other in order to accomplish goals. Hereby the leader will decide what kind of relationships to build with his team members, based on their individual tasks, skills and needs. In this way LMX differentiation is more likely to be perceived as an advantage than under conditions where low task interdependence is present (Liden et al., 2006).
Moreover, according to Liden et al. (2006), team members want their leader to show a differentiated leadership style. This is because team members inevitably vary in their contributions, skills, tasks, roles, needs and personality. These differences make them able to achieve (group) goals and so team members may be less motivated when a leader develops the same kind of LMX relationships with each team member. Leaders who differentiate are more likely to increase the individual performance and development of their team members, as they get a more personal approach. Hereby the leader provides each individual team member with specific resources. This approach increases their sense of self-efficacy and performance level (Liden et al., 2006).
This is supported by Van Breukelen et al. (2012), who perceived differentiation as variety. They stated that trainers have to find a balance between cooperation and competition among their team members. Trainers can achieve this by being consistent in which criteria they use for their decisions. They should use development- and performance related criteria in
applying differentiated treatment. These criteria should be based on differences in skills, talents and experience of their athletes. By contrast, trainers should not base their differentiation on athletes’ demographic or non-task-related criteria. Doing this would have a negative effect on the quality of LMX relationships. Trainers should invest in finding the best way to communicate and explain their decisions to the individual team members (Van Breukelen et al., 2012).
Thus, LMX differentiation as variety may also be perceived as more equitable in high task interdependence teams, such as soccer teams, because the differentiation is based upon performance criteria and creating the most effective coordination between varying roles, positions and tasks (Liden et al., 2006; Van Breukelen et al., 2012).
In short, the effect of LMX differentiation as leadership strategy on athletes’ self-efficacy depends to a large extent on how supportive the differentiation is perceived by those athletes. While applying the LMX differentiation, take into account the potential weaknesses and strengths. Hereby it is important to note that individual consideration and a supportive relationship with each team member is not the same as low LMX differentiation. High LMX differentiation is still possible, because individual team members do not have the same tasks, skills, personality or need for resources (Van Breukelen & Wesselius, 2007).
2.3. Individual LMX as moderator
Within trainer-athlete relationships the trainer is responsible for creating and maintaining high levels of self-efficacy, whereby the athletes’ individual differences and needs are central. This is because the followers’ individual perception of the relationship with their leaders determines their individual LMX score and followers who have a high-quality LMX level exhibit greater efficacy, as they incorporate the positive relationship into their self-concept (Walumbwa & Hartnell, 2011). This is supported by Boies and Howell (2006), who state that high LMX relationships are characterized by dimensions that are similar to self-efficacy, such as empowerment, respect, trust and mutual obligations. This suggest that a high individual LMX level positively influences the relationship between LMX differentiation and self-efficacy and vice versa.
In other words, trainers are advised to pay extra attention to the varying and unique qualities and characteristics of individual athletes in order to maximise their self-efficacy. This is supported by Jowett and Cockerill (2003, p.322) who state that “The most important
issues for the coach are to listen, to look, and to understand the performer; in addition, talk to and with the performer.”
In addition, Walumbwa and Hartnell (2011) argue that, in order to increase followers’ self-efficacy, leaders must show confidence and trust in their followers and should be encouraged to develop personal relationships with each individual follower. Hereby the followers are stimulated to challenge their leaders’ ideas and propositions and the leaders should focus on the capabilities and potency of the followers and give them the autonomy and feedback to evolve themselves. It is hereby important that the athlete understands why the trainer gives him or her a differentiated treatment, because this understanding increases the perceived quality of their dyadic relationship.
Prussia et al. (1998) suggested that trainers who focus on developing and sharing individual information with their athletes have high-quality trainer-athlete relationships. In return this has a positive influence on the athlete’s willingness to take initiative and this is an important component of self-efficacy.
This is supported by Van Breukelen and Wesselius (2007), who also highlight the importance of the perceived individual LMX quality. They state that in relation to LMX differentiation, a positive interpretation and evaluation of the LMX differentiation by team members is the most important factor in creating an effective mutual relationship. This is supported by Vella et al. (2013), who state that when athletes perceive an effective high-quality trainer-athlete relationship, this reinforces their perception of competence, social skills, autonomy and self-esteem, which are important components of their self-efficacy. As described by Mageau and Vallerand (2003), this also works the other way around, which means that a perceived low-quality relationship can negatively affect the athletes’ self-efficacy. In line with these statements Boies and Howell (2006) state that in high individual LMX quality relationships, the followers are given more challenging and developmental tasks. Whereas with low-quality relationships the followers’ performance occurs according a more formal and less challenging set of rules.
Adding to the discussion, Boies and Howell (2006) found that high LMX differentiation rather than low LMX differentiation, was positively related to self-efficacy. An explanation might be that high individual LMX minimizes the perceived inequity and competition for attention as consequence of the high LMX differentiation. A leader can accomplish this by differentiation in optima forma, which means that the leader combines high LMX differentiation with high-quality LMX relationships (Boies & Howell, 2006). Adding to this differentiation in optima forma, Van Breukelen et al. (2012) suggest that it is important for leaders to try to develop high-quality relationships with as many team members as possible.
This can be done by differentiation as variety, as the exact nature of these relationships can vary. This also makes high differentiation still possible.
Moreover, Jowett and Cockerill (2003), Schyns, Paul, Mohr and Blank (2005) and Vella et al. (2013) also found indications for a moderating role of individual LMX. Namely, Schyns et al. (2005) argued that at the individual level, LMX quality was positively related to occupational self-efficacy, as individual LMX could be related to individuals' sense of competence. In line with these statements, Jowett and Cockerill (2003) argue that when the dyadic trainer-athlete relationship is perceived as ineffective by the athlete, not only the quality of the trainer-athlete relationship reduces, also the whole psychological well-being of the athlete suffers. Consequently the athletes’ sport related development and performances decrease. These views are also supported by Vella et al. (2013) who state a high-quality trainer-athlete relationship is associated with high self-efficacy and a higher holistic development of the athlete. For example social skills, goal-setting behaviour, initiative-taking, self-efficacy and performances are positively stimulated.
In addition to this discussion, Liden et al. (2006) argue that there is indeed a positive relationship between LMX differentiation and individual performance, but that this relationship could be moderated, as the individual LMX status influences the team members’ reaction to LMX differentiation. They stated that LMX differentiation is positively related to performance, but only under certain conditions. On the one hand Liden et al. (2006) argue that under conditions of low LMX differentiation, low-quality LMX individuals feel that there is no possibility to improve their LMX status and thus feel that there is no incentive to increase their efforts. On the other hand, they argue are that under conditions of high LMX differentiation, low-quality LMX individuals might be more motivated to increase their efforts and performance, as they hope to develop a high-quality LMX relationship with their leader.
Thus, according to the literature a high level of self-efficacy can be created by LMX differentiation, whereby individual LMX plays an influential role. It is important that the leader expresses supporting differentiation, which entails confidence, trust, support, individual consideration and empowerment as these are also important dimensions of self-efficacy. Moreover, the differentiation should be perceived as fair by the team members, as this positively influences their individual LMX (Graen & Uhl-Bien, 1995; Liden et al., 2006). The relationship between LMX differentiation and self-efficacy will be positively influenced by a high individual LMX score and negatively influenced by a low individual LMX score (Liden et al., 2006; Mageau & Vallerand, 2003; Walumbwa & Hartnell, 2011).
Therefore, on the basis of the above mentioned literature, Hypothesis 1 is proposed: ‘Perceived LMX differentiation positively relates to self-efficacy when perceived individual LMX is high rather than low.’
2.4. Self-efficacy in relation to individual development and performance
This part addresses the importance of self-efficacy for the development and performance of an athlete. The relationship between self-efficacy and the individual development as well as performance has already been profoundly researched as most empirical studies have stated that self-efficacy is positively related to the development and performance of individual athletes. In 1998, Stajkovic and Luthans conducted a meta-analysis of 114 studies and found that self-efficacy was positively correlated with work-related performance. Also in the context of sports, athletes with high self-efficacy were found to have better performances than athletes with low self-efficacy (Feltz & Lirgg, 2001).
Despite the fact that the majority of research supports the assumption that high self-efficacy leads to high performance, there is also a small minority of research that states the opposite. For example, Vancouver, Thompson, Tischner and Putka (2002) state that the relationship between self-efficacy and performance is reciprocal and that no causal conclusions can be drawn. Although, there is also some evidence that it might be a two-way direction relationship whereby self-efficacy and performance both stimulate each other, Silver, Mitchell and Gist (1995) claim that this would still mean that self-efficacy is a reliable predictor and determinant of performance. This is because without self-efficacy high levels of performance are unlikely to occur. Moreover, according to the concept of self-efficacy, a successful experience will increase the expectation of future success, and a failure would decrease the expectation of future success. However, Gernigon and Delloye (2003) found that repeated successful experiences builds a resilient self-efficacy buffer, which reduces the negative effects of failure on self-efficacy.
Furthermore, Vancouver et al. (2002) also presented findings that found that high self-efficacy leads to overconfidence which resulted in less efforts and persistence than their low self-efficacy counterparts. Similar results were found by Bandura and Jourden (1991), who found that an increased level of self-efficacy resulted in lower levels of performance. Their explanation of this negative effect was that too much self-efficacy reduces the incentives to display the hard work needed to attain a high level of performance.
However, this is contradicted by Chemers, Hu and Garcia (2001) as they stated that high self-efficacious individuals are characterised by more thoughtful and skilful analytic
strategies for improving their performance compared to their less self-efficacious counterparts.
Another reason why high self-efficacy may lead to decreased performance is because once a high level of self-efficacy is reached, this individual will have an increased perception of his future success, which may lead to unrealistic goal-setting. In other words, if one has committed him- or herself too early to the goal or challenge, the change that they will make mistakes and fail increases. Contrary, when self-efficacy is low, one will be more likely to choose realistic goals, which enhances the change on a successful performance. The tendency that (too) high self-efficacy leads to more mistakes is accountable for the negative relationship between individual self-efficacy and performance. Hereby individuals do not exactly understand which factors are responsible for their performance level, and thus make the wrong attributions about how they achieve the wanted performances (Vancouver et al., 2002). To put it briefly, it seems that in at least in some circumstances the relationship between self-efficacy and performance is contrary to what the dominant self-self-efficacy literature proposes.
In contrast to Vancouver et al. (2002), Maddux (2009) found that self-efficacy plays a major role in successful problem-solving, decision-making and coping with unforeseen difficult situations. And as opposite to the statements of Bandura and Jourden (1991), another reason why self-efficacy positively relates to higher performance outcomes, is because that self-efficacy determines the individual’s choice of goals and goal-aimed behaviour (Walumbwa & Hartnell, 2011). Individuals with high self-efficacy will select the challenges and goals of which they belief that they will successfully complete. This also means that individual with low levels of self-efficacy have low aspirations and commitment considered the goals that they select. Automatically individuals with a high level of self-efficacy will attain higher levels of performance (Walumbwa & Hartnell, 2011). This is in line with the statements of Earley (1994), who proposes that self-efficacy affects performance mainly through an increase in an individual’s effort and persistence to successfully complete a certain task or skill. Accordingly, individual’s with high self-efficacy work harder and longer than individuals with low self-efficacy, which leads to higher levels of development and performance (Earley, 1994).
Thus, the main majority of research in the field of self-efficacy supports its positive relationship with individual performance, however there is small minority who state there might be some circumstances in which this relationship can be negative. Overall, athletes’ self-efficacy research has indicated that self-efficacy is a reliable predictor of sports performance (Earley, 1994; Feltz & Lirgg, 2001; Gernigon & Delloye, 2003; Walumbwa &
Hartnell, 2011). Therefore, on the basis of the above mentioned literature, Hypothesis 2 is proposed: ‘Self-efficacy positively relates to individual development and performance.’
2.5. Self-efficacy as mediator
Why might LMX differentiation and the individual LMX level influence the individual development and performance? Such a question suggest that a mediator must account for these relationships. Based on the above presented theoretical findings this research will now propose that self-efficacy mediates the relationship between LMX differentiation and individual performance and development.
To begin, self-efficacy is already found to be a fundamental predictor of performance. Self-efficacy is the psychological mechanism behind performance according to Bandura (1986). Thus, “perceived self-efficacy is a significant determinant of performance that
operates partially independently of underlying skills" (Bandura, 1986, p. 391). Since decades
the majority of studies that focused on a mediating role of self-efficacy between leadership behaviours and follower performance have indicated that self-efficacy perceptions subsequently affect performance (Walumbwa et al., 2011).
Additionally, in support of Bandura (1986) and Walumbwa et al. (2011), Boardley, Kavussanu and Ring (2008) gave support for a mediating role of self-efficacy by stating that self-efficacy is one of the most powerful psychological mechanisms to achieve sports performance goals. Trainers can enhance players’ self-efficacy by improving their technical and psychological abilities. Improving the necessary skill set of an individual player stimulates his or her self-efficacy and executing the acquired skill set successfully will in turn lead to improved performances (Boardley et al., 2008). Improving the players’ technical skills, verbal persuasion and evaluative feedback are different mechanisms to improve a players’ perception of self-efficacy. Supporting a mediating role for self-efficacy, Kirkpatrick and Locke (1996) stated self-efficacy has such a strong immediate effect on performance, that other variables such as motivation, personality and other incentives influence performance through self-efficacy.
Furthermore, Walumbwa and Hartnell (2011) state that, based on self-efficacy theory, individuals evaluate personal factors to accomplish a certain task or goal. Hereby personal factors include the previous by Bandura’s (1986) described mastery and successful experiences. Building on an effective relationship between leader and follower, these experiences stimulate efficacy. A high LMX relationship with the leader affects the self-efficacy of the follower by incorporating these positive psychological benefits into an
individual’s self-concept. As a result, an individuals’ performance is determined by his or her level of self-efficacy (Walumbwa & Hartnell, 2011).
Despite of the importance of a high-quality relationship between trainer and athlete regarding self-efficacy, Rezania and Gurney (2014) found that most of the trainers in sport only focus on the physical, technical and strategic skills of an athlete. This is because these components are easier to define and control, in contrast to the athletes’ psychological skills and their personal relationships with these athletes. This approach reduces the perceived quality of LMX differentiation and thus the individual LMX quality, as it does not acknowledge the underlying psychological mechanisms of self-efficacy. As a result, because of the fact that self-efficacy is so important for individual development and successfully performing sport related tasks, improving physical, technical and tactical skills loses their potential effects when the athletes’ self-efficacy level is (too) low.
Based on the above mentioned literature, Hypothesis 3 is proposed: ‘Self-efficacy mediates the moderated relationship of LMX differentiation on individual development and performance by individual LMX such that the relationship is positive when individual LMX is high rather than low.’
3. Method
In this section the empirical process of this research is presented. First, the sampling technique is described. This is followed by the measurement procedure. After this the scales and control variables are described.
3.1. Sampling technique
In total a number of 105 players and 44 trainers participated in the research with a participation rate of 100%. Hereby one player was excluded because of missing values. The high participation rate can be explained because not all team members were required to participate in this research, as the focus was on dyadic relationships between trainer and player. Additionally, the participating players and their trainers were selected and personally approached by the researcher before or after training sessions. Hereby the researcher used his networks to conduct the survey questionnaires. Moreover, a non-probability quota sample was used, so that only trainers and players who are part of the same team and are actively practicing soccer had a chance of being selected (Saunders, Lewis & Thornhill, 2009). In this way it was also secured that the right amount of dyadic trainer-player relationships was obtained. Moreover, by personally selecting and approaching the participants it was secured that the right target group participated: young (12 - 18 years) or adolescent (18 - 25 years) soccer players that were active within a Dutch soccer (youth) academy.
The inclusion of the trainers’ and players’ perspective is necessary in order to contribute to a more holistic view of the trainer-player relationship. As the research is about the dyadic relationship between trainers and players, players and their trainer who are member of the same team, were selected. Per team each player that participated in the research was randomly selected. Per trainer at least one player has participated in the research in order to examine their dyadic relationship. There are no trainers who participated without at least one of their players also participating and vice versa.
An information session took place before administering the questionnaires (Appendix B and C), whereby aside from information on the survey procedure, participants are ensured that anonymity is guaranteed, participation is voluntary, and data will be treated strictly confidential.
3.1.1. Demographic characteristics
All participants were male and 82.4% was higher educated (VWO, HBO or WO). Most participants played part-time (81.1%) and the average age was 16.9 years (SD = 4.17), ranging between 11 and 27. Most of them had the Dutch nationality (93.4%), while 72.6% had a Dutch ethnicity. Remaining nationalities were German (1.9%), Moroccan (1.9%), Surinam (1.9%), Portuguese (0.9%) and Indonesian (0.9%). Remaining ethnicities were Moroccan (6.6%), German (3.8%), Indonesian (2.8%), English (1.9%), Antilles (1.9%), Belgium (0.9%), Turkish (0.9%), Portuguese (0.9%), Spanish (0.9%), American (0.9%), Armenia (0.9%), Colombian (0.9%), Ghana’s (0.9%), Cape Verdant (0.9%) and Tunisian (0.9%).
Additionally, 44 male trainers were included in the sample with a participation rate of 100%. Out of the trainers 86.3% were higher educated. Again most participants played part-time (88.1%). The average age was 31.3 years (SD = 10.0), ranging between 18 and 66. Most of them had the Dutch nationality (90.9%) and ethnicity (86.4%). Remaining nationalities were Turkish (2.3%), Eritrean (2.3%), Antilles (2.3%) and South-African (2.3%). Remaining ethnicities were Moroccan (4.5%), Belgium (2.3%), Burkina-Faso (2.3%), Egyptian (2.3%) and Italian (2.3%).
3.2. Measurement procedure
Within this quantitative research design a correlational research design is used, whereby the statistical package for the social sciences (SPSS) is used to analyse the data. SPSS is used because it is commonly accepted as a programme to analyse social and business data (Saunders, 2011). This includes that the research is aimed at exploring the type of relationship between variables. Furthermore, the research is based on a cross-sectional survey design as this is the most feasible way to test hypotheses and create external generalizability (Saunders et al., 2009).
Two paper-based surveys were provided (Appendix A): one for the coaches and one for the players. For 27 trainers only one player participated in the survey, whereby 17 trainers were connected with 2 to 5 players. The average amount of players per trainer was 1.28. The reason why the amount of players per trainer differs is because the process of conducting paper-based surveys on multiple teams with multiple players per trainer took too much time. Moreover, the focus of the research lies on the unique dyadic relationship between trainer and player and that is why in a later stadium the decision was made to select and link only one player per team to the trainer.
3.2.1. Additional analyses
As not all trainers were exclusively linked to a unique player, the analysis was split in two. The primary analyses were run for all participants (N = 105) to preserve statistical power. This however means that independence was assumed among the members of a team. Whereas several team members were nested under the same trainer in 74.3% of the sample, analyses were also run on the basis of only one randomly selected player per trainer (in total 44 dyads).
Moreover, additional analyses were run using measures of differentiation as disparity to see whether the results differed fundamentally from those obtained following a conceptualization of differentiation as variety. These additional analyses for N = 44 and other differentiation measures were run in order to establish robustness of the findings and are reported in the Appendices F, G and H.
3.3. Measures
In this part the used measurement scales will be presented below. Starting with independent variable LMX differentiation and followed by moderator individual LMX, mediator
self-efficacy and dependent variable individual development and performance. Several scales from
the literature were used to measure the different variables in the model and their reliability was checked to establish whether items showed consistency. The variables
LMX-differentiation, individual LMX and self-efficacy were measured based upon the individual
players’ perspective. The individual development and performance variable was measured based upon the individuals’ and the trainers’ perspective. All scales are created by calculating the mean of the items.
For measuring these main constructs, self-constructed and existing questionnaires are used and adapted according to this research needs. As the respondents are Dutch, the questionnaires are translated into Dutch, based on back-translation, in order to assure the best understanding. These translations are checked by a professional psychologist with a high English command and approved by the thesis supervisor. All used measurement scales succeeded Cronbach’s alpha of .70 and can thus be considered as reliable.
3.3.1. LMX differentiation
To start, LMX differentiation is measured based upon a 5 item (1 - 5)1 questionnaire from Vidyarthi, Liden, Anand, Erdogan and Ghosh (2010) with a focus on differentiation perceived
1
as variety. Cronbach’s alpha is 0.812 and an example item is: “In comparison to others in my team, I receive more support from my trainer.” Differentiation by Vidyarthi et al. (2010) is used for the primary analyses because it perceives Differentiation as variety, has a high Cronbach’s alpha of 0.81 and has already been published before.3
3.3.2. Individual LMX
Individual LMX is measured based upon a unpublished Dutch 6 item (1-5 scale) questionnaire
provided by professor Van Breukelen (2017) which measures the perceived quality of the dyadic relationship from the players’ perspective. Cronbach’s alpha is 0.83 and an example item is: “Hoe is de band met uw coach?”
3.3.3. Self-efficacy
Self-efficacy is measured based up a 4 item (1-7 scale) questionnaire from Hannah,
Schaubroeck & Peng (2016). Cronbach’s alpha is 0.81 and an example item is: “How would you evaluate your competency as a soccer player?”
3.3.4. Individual Development and Performance
In order to have a broader and deeper perspective Individual Development and Performance is measured from the trainers’ and players’ perspective with combined measure scales. Hereby it is assumed that the measures are consistently positively related and manifest the same construct.
Considering the measurement of Development two scales are combined. One scale is from a trainers’ perspective and is a 3 item (1 - 7) questionnaire from Gong, Huang & Farh (2009), which focuses on individual development. The Cronbach’s alpha is 0.72 and an example item is: “My players make significant contributions to the overall performance.” The other scale is from a players’ perspective and is a 4 item (1 - 4) questionnaire from Sullivan, LaForge-MacKenzie & Marini (2015), which also focuses on the individual development. Cronbach’s alpha is 0.86 and an example item is: “I have improved: technical skills.” These two scales combined have a Cronbach’s alpha of .82, which makes it a reliable measure.
Also considering the measurement of Performance two scales are combined. One is from a trainers’ perspective and is a 3 item (1 - 7) questionnaire from Hannah, Schaubroeck & Peng (2016), which focuses on individual performance. Cronbach’s alpha is 0.91 and an
2 All presented Cronbach’s alpha are measured specifically in this study. 3
example item is: “This player exhibits strong basic skills.” The other scale is from a players’ perspective and is provided by Van Breukelen (2017). This Dutch 2 item (1 - 5) questionnaire again focuses on players’ individual performance and has an Cronbach’s alpha of 0.62. An example item is: “Is dit een goed sportseizoen als u kijkt naar uw eigen prestaties?” Together these two scales have a Cronbach’s alpha of .83, which makes it a reliable measure.
3.4. Control variables
Next to the main research constructs, the research also included control variables such as age
player in years, age trainer in years, team size, tenure in months (ratio variables), and ambition player (nominal variable). These control variables were included in order to rule out
alternative or confounding explanations.
Age trainer was added as control variable because leaders who are older are considered to
have more experience, and thus might better be able to develop high quality relationships with its individual team members. Moreover, older leaders might be less inclined to focus on their own success and have more personal attention for their younger team members. Consequently, this could influence the players’ development and performance (Zacher, Rosing, Henning & Frese, 2011). The control variables ambition player and age player were added because, people in their early- to mid-career are expected to have more ambition to achieve career success than people who are approaching their retirement. This increases their level of performance according to the findings of Erdogan, Kraimer and Liden (2004) and Flaherty and Pappas (2002). Furthermore, team size was added as control variable to check if the total number of team members had an influence on the perceived individual LMX and on the individual development as well as performance of the player. Namely, team size might influence the interaction and dynamics among team members (Hooper & Martin, 2008; Wheelan, 2009). At last, tenure was added because of its potential effects on team members' self-efficacy and the relationships with their leader. Namely, tenure captures the experience of team members working within a particular team with a particular trainer, which might influence their mutual relationships, and consequently their individual performance (Erdogan & Liden, 2002; Hooper & Martin, 2008; Liao et al., 2010). Moreover, González‐Romá, Fortes‐Ferreira and Peiro ( 2009) suggested a positive linear relationship between tenure and performance, because greater tenure inclines that team members have overcome initial problems which were caused by novelty and a lack of experience. Accordingly, greater tenure stimulates coordination and communication between team members mutually and between team members and their leader. Thus, this can influence their self-efficacy and performance.
4. Results
In this part the analysis of the results is described. In the beginning the correlation results are presented, which are subsequently followed up by the hypotheses results4.
4.1. Correlations
Table 1 shows the means, the standard deviations, and the correlations between the study variables, whether significant or insignificant. The table includes the main variables LMX differentiation (Differentiation), individual LMX, self-efficacy, individual performance and
development, as well as the control variables team size, tenure, age player, age trainer and ambition player. In line with the expectations, Performance and Development are highly
correlated in a positive direction (r = .76). It can also be seen that Differentiation is significantly related to both Performance (r = .25) and Development (r = 0.25). Moreover,
individual LMX has a significant relation to both Performance (r = .71) and Development (r =
.60). As expected, Self-efficacy has a high positive relationship with Performance (r = .59) and Development (r = .65). Considering the control variables only Ambition Player is significant related to Performance (r = 0.28), but not to Development (r = 0.15). Ambition
Player also significantly correlates negatively with Age Player, which means that players lose
ambition while they grow older.
Table 1 :Means, standard deviations and Spearman correlations for N=1055
# Variables M SD 1 2 3 4 5 6 7 8 9 10 11 1 Team Size 15.85 2.65 - 2 Tenure 13.77 9.73 -.09 - 3 Age Player 16.91 4.19 .18 -.03 - 4 Age Trainer 30.42 9.65 -.02 .39** .38** - 5 Ambition Player*** 0.52 0.50 -.10 .11 -.40** .24** - 6 Differentiation 2.74 0.52 .03 .08 .22** .08 -.06 - 7 Individual LMX 3.43 0.79 .04 .05 -.04 -.16 .15 .29** - 8 Self-efficacy 4.22 0.79 .12 .13 -.11 -.07 .18 .24* .54** .53** - 9 Performance 4.34 1.01 .06 .18 -.04 -.07 .28** .25* .71** .68** .59** - 10 Development 4.10 0.78 -.01 .09 -.01 -.02 .15 .25** .60** .65** .65** .76** -
5. Note. *p <.05, **p <.01. ***Dichotomous variable
4 The results of the additional analyses for N = 44 and the other measures of LMX differentiation are reported in
the Appendices F, G and H.
5
Before we go into the multiple regression models that were used to test the hypotheses, it is good to note that the assumptions underlying the regression were met. Apart from the fact that no multicollinearity can be found, as within all regressions VIF was structurally lower than four, residuals are approximately normally distributed and no serious heteroscedasticity shows (see appendix D and E for the respective figures with respect to Table 1). This means that no assumptions underlying the regression are violated, which validates the used statistical approach.
4.2. Hypotheses results
All variables were standardized with exception of the dummy variable, which is Ambition
Player. The hierarchical regression exists of four steps. In the first step the control variables
are entered, while in the second step the independent variables are entered. In the third step, the moderation between the independent variables is tested and in the fourth step the mediation variable is included.
4.2.1. Hypothesis 1
In Table 2 the regressions results of Self-efficacy as dependent variable (DV) and
Differentiation as independent variable (IV) are presented.
Table 2: Regression results with Self-efficacy as DV and Differentiation as IV (N=105) 6
Step 1 Step 2 Step 3
Variables B SE (B) B SE (B) B SE (B) Step 1: Controls Team Size Tenure Age Player Age Trainer Ambition Player Step 2: Main effects Differentiation Individual LMX Step 3: Moderation Differentiation * Individual LMX .158 .158 -.044 -.078 .282 .099 .109 .114 .118 .215 .137 .084 -.116 .031 .153 .112 .487** .085 .094 .099 .102 .185 .089 .089 .149 .090 -.118 .022 .123 .139 .470** -.081 .086 .094 .099 .102 .187 .094 .091 .087 R² in % Adjusted R² in % F-value 7.5% 2.9% 1.613 34.2% 29.5% 7.206** 34.8% 29.4% 6.407** **Significant at the 0.01 level (2-tailed) * Significant at the 0.05 level (2-tailed)
6
To start, Hypothesis 1 ‘Perceived LMX differentiation positively relates with self-efficacy when perceived individual LMX is high rather than low’was tested. The moderation was not significant. As can be seen in step 3 of Table 2, the interaction term has a coefficient of -.08 (p > 0.05). This means that there is no evidence to support Hypothesis H1 proposing a moderating role for Individual LMX.
4.2.2. Hypothesis 2
Table 3 shows the regressions results of Performance as dependent variable and
Differentiation as independent variable. When you look at the regression for all participants
all four models are significant. The maximum variance in Performance explained is 56.4% in the final model.
Table 3: Regression result with Performance as DV and Differentiation as IV (N=105) 7
Performance
Step 1 Step 2 Step 3 Step 4
Variables B SE (B) B SE (B) B SE (B) B SE (B) Step 1: Controls Team Size Tenure Age Player Age Trainer Ambition Player Step 2: Main effects Differentiation Individual LMX Step 3: Moderation Differentiation * Individual LMX Step 4: Mediation Self-efficacy .086 .201 .108 -.115 .564** .097 .107 .111 .115 .210 .055 .110 .037 .029 .391* .045 .666** .070 .077 .082 .084 .152 .073 .073 .054 .110 .037 .030 .394* .042 .668** .009 .101 .214 .131 .166 .297 .146 .118 .137 .015 .086 .068 .024 .362* .005 .545** .031 .263** .069 .074 .079 .081 .148 .075 .081 .069 .080 R² in % Adjusted R² in % F-value 12.2% 7.8% 2.76* 55.7% 52.5% 17.4** 55.7% 52.0% 15.1** 60.2% 56.4% 15.9** **Significant at the 0.01 level (2-tailed) * Significant at the 0.05 level (2-tailed)
7
