یک تحقیق در سطح تیم از رابطه بین تمایز، و تعهد و عملکرد تبادل رهبر ــ عضو (LMX)

Although the differential treatment of team members by their leader is at the heart of Leader–Member Exchange (LMX) theory, empirical studies exploring the role of within-team LMX differentiation in relation to team outcomes are still relatively scarce. This study among 269 Dutch secondary school teachers from 33 different teams tested the hypotheses that the relationship between LMX differentiation and team commitment, and team performance is moderated by LMX-quality median. Moreover, we hypothesized that team members' perceived dissimilarity regarding work values and orientations would be positively related to within-team LMX differentiation. Teachers completed questionnaires on LMX-quality, dissimilarity, and team commitment, whereas team performance was rated by school principals. Results indeed showed that LMX differentiation is positively related to both outcome variables in teams with a low LMX-quality median only. As expected, more perceived dissimilarity between team members was related to more within-team variability in LMX-scores. These results contribute to knowledge on hypothetical antecedents and consequences of LMX differentiation at the team level.

Leader Member Exchange (LMX)-theory contends that leaders develop differential types of relationships with each of their subordinates through a series of work-related exchanges (Graen and Cashman, 1975 and Graen and Scandura, 1987). The so-called quality of these leader–member exchange (LMX) relationships is assumed to reflect the extent to which leader and subordinate mutually exchange resources and support. Whereas low quality LMX-relationships are based on exchanges directly specified by the employment contract, high quality LMX relationships include exchanges of both material and non-material goods beyond what is required by the formal employment contract. In the latter type of relationships, LMX-partners generally show high levels of mutual respect, trust, affection and obligation (Dansereau et al., 1975, Graen, 1976, Graen and Schiemann, 1978, Graen and Uhl-Bien, 1995 and Liden et al., 1993). The development of differential relationships among leaders and the subordinates who report directly to them in their work groups is referred to as LMX differentiation (Liden, Erdogan, Wayne, & Sparrowe, 2006, in: Henderson, Liden, Glibowski, & Chaudry, 2009). High differentiation within a work-group or team suggests that there is a broad range in (the perceptions of) overall LMX-quality, whereas low differentiation refers to a context in which the range of overall LMX quality within a work-group or team is small (Henderson et al., 2009). Given its explicit focus on the quality of the exchange relationship that develops between a leader and each follower, the vast majority of empirical research on the LMX-theory has concentrated on testing relationships between LMX-quality and work-related outcomes at the dyadic (i.e., leader–member) level. Over the past decades, many studies have supported the positive effects of high LMX-quality on individual employee outcomes such as job satisfaction, job involvement, in-role performance, organizational commitment, organizational citizenship behavior and decreased propensity to quit (see e.g., Duarte et al., 1993, Epitropaki and Martin, 1999, Gerstner and Day, 1997, Graen and Uhl-Bien, 1995, Ilies et al., 2007, LePine et al., 2002 and Vecchio, 1982). However, the prevailing individual level focus in research on LMX is surprising, given that the LMX model was originally advanced to account for how leaders' differential treatment of multiple subordinates in a work group influences activity within the group (Henderson et al., 2009). Though individual-level perceptions of LMX-quality are reflective of interpersonal social exchange behaviors and motives in the leader–member dyad, they do not capture how within-group patterns in LMX-quality may create a social context that influences employee attitudes and behaviors (Henderson et al., 2009 and Mayer and Piccolo, 2006). As Day and Schyns (2010) note, the different relationships a leader has with each follower can influence other relationships in the collective, and thus affect the results of leadership at the work group level. LMX differentiation can be considered the core element of the LMX-model (Henderson et al., 2009). However, empirical research on the role of LMX differentiation at the (work) group or team level is still scarce and its results are inconsistent. Recently, two studies tested the interaction effect between LMX differentiation and mean or median LMX-quality at the team level (Boies and Howell, 2006 and Liden et al., 2006). Whereas Liden et al. (2006) found a significant interaction effect on team performance that showed the expected sign, Boies and Howell (2006) found a significant interaction effect on team potency and team conflict that showed a sign contrary to what was expected. So, the role of LMX differentiation at the team level is still far from clear, and more research on this construct is needed in order to advance LMX-theory. This research will contribute to moving this theory into its fourth stage of development as identified by Graen and Uhl-Bien (1995), in which leader–member relationships are studied within their respective work unit. Therefore, the main goal of the present study is to examine the relationship between LMX differentiation and two relevant team outcomes (team performance and affective team commitment). As we will show later, we hypothesize that this relationship is moderated by LMX-quality median. Taking into account the central role that LMX differentiation has within LMX-theory, it is also theoretically important to identify its antecedents, so that its nomological network becomes clearer. Thus, we will also examine whether perceived dissimilarity among team members is related to LMX differentiation. Besides being important from a theoretical perspective, as outlined above, an increased understanding of the role of LMX-differentiation is also relevant from a practical point of view. Nowadays many organizations are organizing work around teams as the basic unit of their structure. Thus, the examination of leadership effects at the group level might be particularly relevant for organizations where leaders lead teams rather than single individuals, and where the work is organized around whole units (Boies & Howell, 2006). Gaining more detailed insight into the circumstances under which LMX differentiation is either functional or dysfunctional, may be useful to design strategies for managing teams based on LMX propositions. For instance, if LMX differentiation is found to be functional when the general (median) level of LMX-quality within a team is low, team managers should learn to build differentiated leader–member relationships with specific team members.

A still unresolved issue in LMX research is whether leaders should or should not differentiate among their subordinates. The performance of a leader who attempts to develop generalized reciprocity with each team member might suffer from an overinvestment in redundant contacts (Brass, 1995, in: Sparrowe & Liden, 1997). On the other hand, if a leader differentiates severely among team members, some of them are likely to become disenfranchised (Yukl & Van Fleet, 1992, in: Sparrowe & Liden, 1997). So, there is a strong need for further research on the correlates and influences of the LMX differentiation process (Gerstner and Day, 1997, Graen and Uhl-Bien, 1995 and Henderson et al., 2009) as well as on context factors influencing the relationship between LMX differentiation and team outcomes. Considering this, the present study among 33 governance teams of secondary school teachers tested the potential moderating role of teams' LMX-quality median in the relationship between teams' LMX differentiation and team performance and affective team commitment, respectively. Moreover, the relationship of team members' perceived dissimilarity with LMX differentiation was also investigated. The results of hierarchical multiple regression analyses showed that LMX differentiation was positively related to team performance and affective team commitment in teams with a low LMX-quality median yet unrelated to both outcomes in teams with a high LMX-quality median. These results provided empirical evidence supporting Hypothesis 1a and Hypothesis 1b, respectively. Taking into account the scarcity of studies investigating the relationship of LMX differentiation with team outcomes, and the inconsistent results obtained by these studies, the results reported here for two distinct team outcomes represent clear empirical evidence for conditions under which LMX differentiation may contribute positively to team functioning. We also observed a positive relationship between dissimilarity in work values among team members and LMX differentiation. This result supported Hypothesis 2. Considering the very little empirical attention that has been devoted to the antecedents of LMX differentiation to date (Henderson et al., 2009, p. 519), this finding represents a promising initial step in this line of research. 5.1. Implications for theory and research Our study has a number of implications for future theoretical development and research. First, together with recent empirical studies on LMX (e.g., Cogliser and Schriesheim, 2000 and Henderson et al., 2009), our study supports the idea that LMX theory can and should be investigated at the team level. In so doing, future research and theoretical development will contribute to moving this theory into its fourth, team-level stage of development (Graen & Uhl-Bien, 1995), and developing meso-level models of leadership (House, Rousseau, & Thomas-Hunt, 1995). Second, in the present study we have shown that LMX differentiation, in combination with LMX median, “creates a group-level context” (Henderson et al., 2009, p. 519), which contributes to shaping important team outcomes (i.e., team performance and affective team commitment). These results, together with those obtained by Liden et al. (2006), stress the central role that LMX differentiation has in LMX theory, and clearly show that its relationship with the aforementioned outcomes depends on LMX-quality median. Future theory and research on the influence of LMX quality on team outcomes should consider both LMX constructs. Third, in addition, LMX differentiation may also provide the framework to shaping individual level outcomes for team members. Members of work teams seem to evaluate the quality of their LMX-relationships not only in the absolute sense (low vs. high), but also with reference to their perception of other members' relationships with the leader (Henderson et al., 2009). Thus, they make social comparisons between their own and their fellow team members' LMX relationships. As such, the pattern of diverse LMX relationships that develop within the work team provides the parameters and framework for valid social comparisons (Hogg et al., 2005). Future research on LMX theory should pay attention to the role of social comparisons focused on LMX quality in the explanation of individual team members' outcomes. Other studies have shown the usefulness of social comparisons in work settings to understand employees' outcomes (e.g. Buunk, Zurriaga, González-Romá, & Subirats, 2003). Moreover, results of previous empirical studies (Sias, 1996 and Sias and Jablin, 1995) have demonstrated that if followers perceive differences among peers such as on effort, loyalty, and task performance, they may actually be upset if the leader does not differentiate. So, future studies could also take into account team members' perceptions of their peers' functioning. Fourth, our findings highlight the importance of dispersion constructs in research on work teams. LMX differentiation is a dispersion construct because it refers to the variability of a lower-level property (i. e., the quality of the leader–member exchange relationship) within a group or unit. Dispersion constructs are relatively rare in the organizational literature, but recent research on different topics (e.g., team climate, team mental models, team affect) has shown that dispersion constructs (climate strength — i. e., within-team agreement in climate perceptions, mental model sharedness, and within-team affective homogeneity, respectively) can help us to improve our understanding about team processes and outcomes (Barsade et al., 2000, González-Romá et al., 2002 and Mathieu et al., 2005). The results reported here stress the importance of a dispersion construct (LMX differentiation) for future development of LMX theory. Fifth, our study has contributed to identifying a potential antecedent of LMX differentiation (i.e., dissimilarity among team members regarding work values). This is important from a theoretical point of view because recent theoretical models on the antecedents of LMX differentiation (e. g., Henderson et al., 2009) did not include this variable as one of the potential antecedents. The results reported here suggest that it should be incorporated into these models. Thus, our study contributes to improving our understanding of the nomological network of LMX differentiation. 5.2. Practical implications Practically speaking, our findings emphasize the importance of context factors as moderators of the effects of within-team differentiation in LMX-relationships. As Fig. 1 and Fig. 2 show, the highest scores on team performance as well as affective team commitment are found in teams where a low level of LMX differentiation is combined with a high LMX median. This finding is in line with the prescriptive perspective on LMX (Graen & Uhl-Bien, 1995), which recommends that leaders work to establish a good relationship with every follower. According to this perspective, high consensus (low differentiation) along with positive relationship quality (high average LMX quality) is a desired and effective way of leading and achieving so-called LMX-excellence (Schyns & Day, 2010). However, this could pose problems for leaders who do not have the resources to establish a good relationship quality with all followers, e.g. because of strict organizational rules or a large span of control (Schyns & Day, 2010). According to the findings of our study, in these cases (i.e., teams with a low LMX-quality median) leaders may actually enhance performance and affective commitment by differentiating in their LMX-relationships with their subordinates. They can do so by giving some of their followers support and acknowledgment, allowing them to participate in decision making, delegating on them important team tasks, talking to (and listening) them about their concerns and expectations (Graen et al., 1982 and Yukl, 2009). The theoretical rationale underlying Hypothesis 2, and the results obtained on the relationship between dissimilarity among team members regarding work values, on the one hand, and LMX differentiation, on the other, suggest that by staffing team members with similar work values LMX differentiation will decrease because it will be easier for leaders to meet team members' needs and desires. Staffing team members with work values that fit the team and organization values is a common practice that should yield homogeneous teams regarding work values. The practical implications of this staffing strategy must be assessed considering the ability of the leader to develop high-quality relationships with most of his/her followers (and consequently, creating a high LMX-quality median). If the leader is able to do this, team performance and affective team commitment will be high (as Fig. 1 and Fig. 2 show); however, if the leader is not able to do it these two team outcomes will be lower. These two scenarios highlight how selection practices may affect LMX differentiation and the important role that leaders' skills have in this context. 5.3. Limitations Our study shows some limitations. First, its cross-sectional design does not enable to establish the direction of causality. For example, instead of being an antecedent of LMX differentiation, dissimilarity might as well be a consequence of the extent to which a team leader treats his subordinates unequally. Longitudinal studies are needed to further clarify the relationship between LMX differentiation and its potential antecedents. Second, the sample of teams used in our study was composed of only one type of team. This limits the generalizability of our results. It would be interesting to see if the results of this study can be generalized to the functioning of teams outside the school setting. Third, we used a sample of 33 teams. This small sample size means that we had low statistical power for hypothesis testing. However, the results obtained supported all the study hypotheses. Fourth, most of the study variables come from the same source (the team members). Therefore, common method variance might have contributed to the observed relationships. However, the fact that some of the relationships were also observed when the criterion variable (team performance) came from a different source (school principals) suggests that common method variance was not a problem in the present study. This study was designed in order to contribute to the literature on LMX differentiation by showing that the relationships of LMX differentiation with team outcomes are contingent upon context factors, and that perceived dissimilarity among team members regarding work values and orientations is related to LMX differentiation. According to the results obtained, we can state that the two expected contributions were achieved.