You don’t actually want to get closer to the star: How LMX leads to workplace ostracism

Abstract

High-quality leader-member exchange (LMX) is commonly seen as beneficial to employees. However, this is not always the case in the eyes of other members of the same team. Based on social comparison theory, we propose that members who have high-quality LMX relationships with team leaders might face workplace ostracism through being envied by other members of the same team. Further, we hypothesize that this indirect influence is mitigated by the high-quality LMX member’s agreeableness. Based on data from 196 employees, we found that though ostensibly LMX quality directly led to less workplace ostracism, it had a positive effect on workplace ostracism through being envied by other team members, and agreeableness buffered this indirect positive effect. Theoretical and managerial implications are discussed.

Leader-member exchange (LMX) theory suggests that leaders establish and maintain exchange relationships with followers (Dienesch & Liden, 1986). For decades, as a high-quality LMX relationship consists of mutual respect, trust, and loyalty between the leader and follower (Graen & Uhl-Bien, 1995), prior research has widely investigated its applausive outcomes on the part of followers, such as improved job satisfaction and performance (Gerstner & Day, 1997). Yet, the LMX theory itself purports that leaders have limited time, energy and resources, and thus they are unlikely to form a high-quality exchange relationship with each of their followers (Dienesch & Liden, 1986). This differentiation of LMX relationships within one team is not only a choice, but also a necessity.

Despite the detrimental effects of LMX differentiation that have been documented in one recent literature review (Henderson et al., 2009), high-quality LMX relationships are still recognized as beneficial to the follower in the leader-follower dyadic pair. As the star worker literature would suggest, we argue that a follower who has a high-quality LMX relationship with the leader could enjoy benefits on the one hand, but that there may be some downsides of this focal relationship as well (Kim et al., 2010; Vidyarthi et al., 2010).

Employees are often nested within work groups, and consequently the differentiation nature of LMX relationships is likely to trigger the social comparison process of employees (Vidyarthi et al., 2010). Drawing on social comparison theory (Festinger, 1954), therefore, we develop a model that depicts how a high-quality LMX relationship could lead this focal follower to workplace ostracism, defined as the extent to which an individual perceives that he or she is ignored or excluded by others (Ferris et al., 2008), from other team members who have relatively lower-quality exchange relationships with the same leader. As previous studies have shown that social comparison behaviors could effectively elicit certain emotional responses (Greenberg et al., 2007; Buunk et al., 1990), we argue that the envy emotion of other team members (i.e., being envied) could serve as the mediating mechanism through which high-quality LMX relationships lead to workplace ostracism by other team members of the same team. We then take a step forward to examine the buffering role of agreeableness of followers, since the target person’s personality effectively determines the extent to which one translates emotion into behavior (Tai et al., 2012). Figure 1 depicts our overall research model.

Fig. 1

Theoretical Model

Our study makes several important contributions to the LMX, workplace ostracism, and emotion literature. First, while prior studies on LMX encourage followers to establish and maintain high-quality exchange relationships with their leaders to earn extra benefits for themselves (Gerstner & Day, 1997), ours suggest that these benefits come at a price of relational crisis with other team members. Second, although workplace ostracism as a commonly experienced social phenomenon has attracted growing scholarly attention, most aim to examine its consequences, and thus few studies have explored its antecedents (Robinson et al., 2013). Our study is among the first to investigate how leader-follower interaction may result in ostracism. Third, our study shows that interpersonal comparison within a team could arouse emotional responses among team members. Given that extant research has identified many intra-person emotion inducers at work (Thompson et al., 2016), we identify an interpersonal emotion of being envied inducer (i.e., LMX), adding a significant piece of knowledge to the emotion literature.

Theory and hypothesis

LMX and workplace ostracism

LMX theory suggests that leaders are usually unable to keep high exchange relationships with all the employees within the team due to limited time and resources (Dansereau et al., 1975). As a result, leaders have an inclination to treat followers differently (Dulebohn et al., 2012). In most circumstances, leaders would establish high-quality LMX relationships with only a few key members, and develop mediocre LMX relationships with other members (Dansereau et al., 1975. With “in-group” members, leaders build a socio-emotional exchange relationship consisting of promise, support and trust. However, with “out-group” members, leaders set up a transactional relationship based on work contracts (Dienesch & Liden, 1986; Graen & Uhl-Bien, 1995). These high- and low-quality LMX relationships interact with each other within a team.

In another vein, social comparison theory purports that people have the inborn inclination to evaluate their own abilities and opinions accurately, and the evaluations depend on comparisons with others (Festinger, 1954). By comparing oneself with others, social comparison serves as a fundamental need to fulfill one’s quest to discover the reality of themselves based on self-relevant information (Buunk & Gibbons, 2007). Given that social comparison exists in all kinds of organizations (Greenberg et al., 2007), and that employees whom the leader favors appear to be observed by others (Duchon et al., 1986), it is possible for employees to initiate comparisons of LMX relationships, considering the incremental information for evaluating work circumstances that the comparison of one’s own LMX relationship with those of others in the same work group provides (Vidyarthi et al., 2010).

Integrating the above two theoretical perspectives, we could infer that followers within a team would inevitably compare their LMX relationships with each other. During this constant comparison process, followers who realize that their LMX relationships are worse than others will feel their mental balance impacted (Kim et al., 2010). Though the formation and maintenance of the LMX relationship is bilateral, the leader is predominant and more importantly, the relationship is relatively stable once established (Maslyn & Uhl-Bien, 2001). Therefore, followers are unlikely to improve the quality of an existing LMX relationship on their own.

There is consistent evidence that both the assimilation and contrast effect can be encountered in many organizational situations (Greenberg et al., 2007; Buunk & Gibbons, 2007). Specifically, assimilation seems more likely if targets and standards are close, whereas contrast appears to occur when they are not (Mussweiler et al., 2004). As discussed above, high LMX is mostly not reachable for “out-group” members. Consequently, they only have an unfavorable self-evaluation comparing themselves to others who are better off, considering the fact that comparison with an upward target will lowered mood and/or self-evaluation (Buunk & Gibbons, 2007). Moreover, since everybody has the instinct to maintain a positive self-evaluation (Tesser, 1988), followers who have relatively worse LMX relationships might suffer drastically because they cannot reach the targets of the upward social comparison (Collins, 1996). Since people’s behaviors are affected by the results of comparison with targets, we would argue that followers with lower-quality LMX relationships tend to take action to restore their positive self-evaluations (Vidyarthi et al., 2010).

One major action that followers with low-quality LMX relationships take is to ostracize the followers with high-quality LMX relationships. In fact, as a universal phenomenon that exists in all types of organizations (e.g., Ferris et al., 2008), workplace ostracism has received burgeoning scholarly attention for the past decade. As a kind of negative interpersonal experience, it can bring emotional exhaustion and depression, weaken self-regulatory resources (Ferris et al., 2015; Wu et al., 2012), reduce job satisfaction, job performance and organizational citizenship behaviors, and raise deviant behaviors and turnover intention (Ferris et al., 2008; Ferris et al., 2015). Previous studies suggest that the causes of workplace exclusion can be divided into two categories: punitive tendency and protective tendency (Robinson et al., 2013). Engaging in workplace ostracism behavior could possibly restore or maintain positive self-evaluation and balance mental states for followers in lower-quality LMX relationships (Khan et al., 2014). For example, ostracizing the “in-group” team members can be seen as a form of punishment from “out-group” members. At the same time, excluding the “in-group” members can reduce contact among followers with differentiated LMX relationships. In this way, followers with lower-quality LMX relationships are able to avoid the negative experiences coming from upward social comparison. Therefore, we propose:

Hypothesis 1: LMX is positively related to perceived workplace ostracism.

The mediating role of being envied

Envy, which is an affective manifestation of lowered self-evaluation, may act as a kind of negative state reacting to falling short of others in attaining desirable outcomes following unfavorable upward comparison (Schaubroeck & Lam, 2004; Kim & Glomb, 2014). Individuals’ emotion of envy is conceptualized in three related ways. The first one refers to a dispositional characteristic which is stable over a period of time. The second one usually associates with a specific individual as referent and has an episodic nature. The last one is a general emotion occurring in a specific circumstance (Duffy et al., 2012; Khan et al., 2014). According to prior definitions, envy can be either benign without hostile emotion or malicious with hostile emotion (Parrott, 1991; van de Ven et al., 2009). However, more researchers are convinced that hostile emotion is indispensable for envy (Smith & Kim, 2007). In this paper, we conceptualize being envied as an episodic individual emotion that has a hostile component fitting with our theory background.

A shared view on envy is that it is a kind of negative emotion that surfaces when one lacks and desires others’ superior qualities, achievements, or possessions (Parrott & Smith, 1993). Extant literature has uncovered three prerequisites for being envied by others. First, the envied target has the better things. Second, the things that one lacks are vital to his or her self-concept. Third, the envious person perceives that the target is similar to him or her to some extent. (Cohen-Charash & Mueller, 2007).

In work teams, as the social comparison theory would suggest, team members always have numerous opportunities to observe what benefit they enjoy and what they lack relative to their colleagues (Duffy et al., 2012). Given the fact that leaders treat followers differently is universal and followers have an instinct for social comparison, it is inevitable for followers to perceive the differences between their and others’ LMX relationships (Kim et al., 2010). Since high-quality LMX relationships always embody a mass of valuable organizational resources (Dienesch & Liden, 1986; Graen & Uhl-Bien, 1995), a member’s destiny in the organization hinges largely on the quality of his or her relationship with the leader (Sparrowe & Liden, 2005). Thus, the significance of LMX should be self-evident. Given that employees work in the same team and are followers of the same leader, they are ideally equal in the organizational hierarchy. Hence, members of the same team have enough similarity to compare with each other. Since followers with lower-quality LMX relationships will not improve their relationship with their leaders on their own, the upward social comparison will potentially cause negative emotions, i.e., envy emotion (Buunk et al., 1990).

Regarded as a strong emotion, envy is hard to hide or change once it arises (Parrott and Smith, 1993). Consequently, the emotion of envy is likely to compel people to do something to discharge the negative experience so that people can maintain mental balance (Tai et al., 2012). Followers who are envious tend to take measures to reduce the gap between him/her and the envied target. Two ways to reach that goal are improving the self and/or depressing the other (Cohen-Charash, 2009). As we discussed before, it is almost impossible for followers to change the LMX relationship only on their own (Maslyn & Uhl-Bien, 2001). As a result, when followers feel envious toward others who have higher-quality LMX relationships, they are more likely to adopt destructive interpersonal behaviors to undermine the target. Research has shown that envy has been deemed to have the power to profoundly affect interpersonal relationships, creating conflict that can manifest as violence and aggression (Vecchio, 2000; Thompson et al., 2016). Previous research has supported the notion that an envious person may harm the envied person, which negatively alters the envied person’s better outcomes (Kim & Glomb, 2014).

Nevertheless, in most social contexts, people are encouraged to appreciate the merits of others, and will be considered immoral or even disgust others if they openly show their envy. The workplace is not an exception. Norms in the workplace always oppose expressing envy openly and few people want to be labeled envious (Smith & Kim, 2007). Consequently, followers do not intend to use public aggressive behaviors to hurt the envied targets. Instead, they are inclined to undermine targets more covertly (Tai et al., 2012), such as lessening organizational citizenship behaviors (OCBs) (Kim et al., 2010), reducing contacts, decreasing the availability of information (Cohen-Charash, 2009), and so on. Since these behaviors are always seen as voluntary rather than compulsory, employees can adopt these behaviors with less risk (Kim et al., 2010).

Compared with other negative interpersonal behaviors in the workplace, ostracism is regarded as more socially acceptable, less psychologically harmful, and less likely to be prohibited (O’Reilly et al., 2015). Thus followers who are envious of other team members tend to engage in ostracizing behavior more easily. Based on the above discussion, we propose the following hypothesis.

Hypothesis 2: Being envied by other team members mediates the relationship between LMX and workplace ostracism.

The moderating role of agreeableness

Although research has shown that envy can cause serious counterproductive work behavior (Cohen-Charash and Mueller, 2007), it does not mean that envy will create permanent destructive consequences. In fact, under particular conditions, episodic envy can even lead to constructive results (Cohen-Charash, 2009).

Current studies indicate that interpersonal impressions (e.g., friendliness and altruism) underlie one of the key dimensions of people making inferences about others, which shape people’s emotions and behaviors (Fiske et al., 2007; Cuddy et al., 2011). For example, an employee’s evaluation of the warmth of the envied target has direct influence on how that employee behaves in response to feelings of envy (Tai et al., 2012). Notably, good qualities such as friendliness, helpfulness, sincerity, trustworthiness and morality are more likely to affect interpersonal behaviors after the experience of envy (Tai et al., 2012). People who are perceived as having those good qualities could elicit positive emotions and behavior from others (Fiske et al., 2007).

Among a variety of personalities, agreeableness reflects the very good quality of interaction “along a continuum from compassion to antagonism”, including trust, straightforwardness, altruism, compliance, and modesty-mindedness as six facets (Costa et al., 1991). We propose that employees who have high agreeableness are more likely to interact in a friendly way with others in daily work. They are always compassionate and therefore, they may be more welcomed in interpersonal relationships (Graziano et al., 1996). Even if these followers are envied because of their high-quality LMX relationships, the impression of agreeableness can act as a buffer. Given that people who relay cues of disagreeableness are especially likely to be ostracized by others (Hales et al., 2016), we contend that envied employees with high levels of agreeableness are less likely to be ostracized by others.

Hypothesis 3: Agreeableness moderates the relationship between being envied by other team members and workplace ostracism, such that for a follower with a high level of agreeableness, the positive relationship between being envied by other team members and perceived workplace ostracism is weaker.

Given that we have argued that LMX positively influences workplace ostracism via the emotion of envy from other team members (i.e., Hypothesis 2), and that agreeableness moderates the relationship between being envied by other team members and workplace ostracism (i.e., Hypothesis 3), we expect that the indirect effect of LMX on workplace ostracism varies as a function of agreeableness. Specifically, this indirect effect via being envied by other team members is stronger for employees low in agreeableness. Thus, we propose the following moderated mediation hypothesis:

Hypothesis 4: Agreeableness moderates the indirect relationship of LMX with workplace ostracism via being envied by other team members, such that this indirect effect is stronger for employees with low agreeableness.

Method

Participants and procedures

We collected data from a battery manufacturing company located in Southern China. With the help of the CEO, we administrated a paper and pencil survey to all 216 employees from 55 teams during work time. After screening invalid data such as responses containing unidentifiable or wrong name and teams with less than 3 members, our final sample consisted of 196 employees from 49 teams, yielding an employee response rate of 90.74%. Team size ranged from 3 to 6 members.

Of all the 196 respondents, females account for 33.02%, and males account for 66.98%. The average age is 26.89 years. 44.9% of them hold a high school degree, 33.2% of them have a bachelor’s degree, 7.1% of them have a master’s degree, only 1 employee has a doctorate, and the rest (14.3%) have other education levels such as junior high school. On average, the employees have worked under the supervision of the current team leader for 2.44 years.

Measures

We measured all the major study variables with well-validated scales adopted from prior studies. As the study was conducted in China, we followed the back-translation procedure to translate the scales from English into Chinese. Unless noted, responses to survey questions were anchored on 5-point Likert scales (1 = strongly disagree, 5 = strongly agree).

LMX

We used a 7-items scale developed by Graen and Uhl-Bien (1995) to measure the quality of LMX. Respondents were asked to report their perceptions of their relationship with their team leader, with one sample item as “Regardless of how much formal authority he/she has built into his/her position, what are the chances that your leader would use his/her power to help you solve problems in your work?” The Cronbach’s alpha was 0.86.

Being envied

Being envied by other team members was measured using in-degree envy centrality as described by Kim and Glomb (2014). We adopted the network measure scale and each respondent was asked to evaluate his/her perceived envy toward each one of the other team members with one item: “I envy this person’s leader-member exchange relationship. For example, (a) it is so frustrating to see this person succeed so easily; (b) feelings of envy toward the LMX of this person constantly torment me; (c) I generally feel inferior to this person’s success on LMX; or (d) this person’s success on LMX makes me resent this person.” We then assessed being envied for each respondent by averaging the envy ratings of all other team members.

Workplace ostracism

We used a 10-item scale developed by Ferris et al. (2008) to measure subjective perceptions of workplace ostracism. An example item is “Others ignored you at work.” Cronbach’s alpha was 0.91.

Agreeableness

We used a 6-item scale developed by Costa et al. (1991) to measure agreeableness. An example item is “I think I am altruistic.” Cronbach’s alpha was 0.87.

Control variables

Beside demographic variables including gender, age, and education, we controlled for the effect of employee tenure of working with the current team leader.

Descriptive statistics and hypothesis testing

Table 1 presents the means, standard deviations, and correlations for the study variables. We applied regression analysis to test the moderation hypothesis (Aiken et al., 1991). To further gauge the mediation and conditional indirect effects, we performed regression and bootstrapping analysis via SPSS Process Macro as suggested by Preacher and Hayes (2008), and Preacher et al. (2007).

Table 1 Mean, Standard Deviation, and Correlations

Hypothesis 1 predicts that LMX is positively related to workplace ostracism. Contrary to our prediction, the regression coefficient is significant but negative (β = −0.28, p < 0.001; Model 3 of Table 2). Thus Hypothesis 1 is not supported. Hypothesis 2 postulates that being envied mediates the relationship of LMX with ostracism. Regression results showed that (a) LMX is positively related to being envied (β = 0.20, p < 0.01; Model 2 of Table 2), (b) LMX is negatively related to ostracism per Hypothesis 1, and (c) when entering the mediator into the regression equation (Model 4 of Table 2), the R-squared changed by 0.23 (p < 0.001), and the coefficients of being envied is significant and positive (β = 0.50, p < 0.001; Model 4 of Table 2). Notably, the coefficient of LMX changed from −0.28 (p < 0.001, Model 3 of Table 2) to −0.39 (p < 0.001, Model 4 of Table 2), which is not consistent with the classic mediation effect test that the coefficient of an independent variable should be smaller when the mediator is entered into the equation (Baron & Kenny, 1986). However, according to Zhao et al. (2010), this result implies a typical competitive mediation effect where the directions of direct and indirect effects are opposite. To further gauge this mediation effect, we then ran a bootstrapping analysis (5000 bootstrap samples) and results show that the proposed mediation effect is significant (B = 0.11, SE = 0.06, 95% CI: [0.02 0.24]), thus Hypothesis 2 is supported.

Table 2 Regression Analysis Results of the Mediating Effect of Being Envied

Hypothesis 3 contends that agreeableness moderates the relationship between being envied and workplace ostracism. We centered being envied and agreeableness, and multiplied them to generate an interaction term. When entering the interaction term into the regression model, the coefficient of the interaction term is significant (β = −0.18, p < 0.01; Model 4 of Table 3). Thus Hypothesis 3 is supported. Figure 2 plots the moderation effect.

Table 3 Regression Analysis Results of the Moderating Effect of Agreeableness

Fig. 2

Moderation Effect of Agreeableness

Hypothesis 4 predicts the conditional indirect effect of agreeableness on the relationship between LMX and ostracism via being envied. Bootstrapping analysis (5000 bootstrap samples) shows that at high level of agreeableness (+1 SD), the indirect effect is weaker (B = 0.07, SE = 0.04, 95% CI: [0.01 0.18]) than that when agreeableness is low (B = 0.13, SE = 0.05, 95% CI: [0.03 0.24]), supporting Hypothesis 4.

Discussion

Based on social comparison theory, this research theorized and empirically tested the effect of LMX on workplace ostracism, and further investigated the mediating effect of being envied by other team members and the moderating role of agreeableness. Results show that the differentiated relationships developed between leaders and followers could induce the social comparison process in followers. Those who have higher-quality LMX relationships are more likely to be envied by other team members, which in turn leads to workplace ostracism. We further found that this indirect influence from LMX to ostracism via being envied is buffered by a follower’s agreeableness, where this influence turns weaker when the follower has a higher level of agreeableness. Our study generates several interesting implications both theoretically and practically.

Theoretical implications

The first contribution of the current study is theorizing LMX as the antecedent of workplace ostracism. Numerous studies have investigated the consequences of workplace ostracism, but relatively few studies explored the antecedent factors (Robinson et al., 2013). Only a handful of studies have focused on the individual characteristics of the ostracized people. Our study set out from an interpersonal perspective to see how the unequal distribution of LMX relationships triggered ostracism behaviors, adding a significant piece of knowledge to the ostracism literature.

Second, this study contributes to the LMX literature. Recognizing the differentiated distribution of LMX relationships within a team, previous studies have uncovered the dark side of LMX differentiation (Henderson, et al., 2009). However, a high-quality LMX relationship itself is still associated with many positive outcomes (Gerstner & Day, 1997; Dulebohn et al., 2012). Although there are several studies showing the negative consequences of high-quality LMX relationships such as stress (Harris & Kacmar, 2006), ours is among the first to show the behavioral downsides, providing an important perspective in understanding the social consequences of LMX.

Third, we also contribute to the envy research. Existing literature has explored many intra-person inducers of the emotion of envy such as Machiavellianism (Vecchio, 2005). Our study provides new empirical evidence that interpersonal comparison could trigger the emotion of envy as well. This is especially important for the study of the roles of emotion in the interpersonal interaction process, where emotions are experienced more intensely. More importantly, we show that personalities (e.g., agreeableness) could function as a crucial boundary condition for turning emotions (e.g., envy) into behaviors (e.g., ostracize others), providing a more comprehensive understanding of emotions in the workplace.

Practical implications

Given that workplace ostracism prevails in daily work life, our study has several implications for management practice as well. Because employees with high-quality LMX relationships are likely to face the risk of being envied, they are more likely to be ostracized in the workplace. Therefore, on the one hand, leaders need to be aware of the potential negative consequences of an overtly good relationship with a follower in the first place. Thus leaders should take corresponding actions to keep differences in LMX relationships within an acceptable range. These behaviors include intentionally observing the disparities among LMX relationships, and consciously communicating with out-group members when the difference becomes significant.

On the other hand, employees themselves should certainly pursue high-quality LMX relationships, but stay alert to the changes in their interpersonal relationships at the same time. Although agreeableness is not easy to change within a certain period of time, we still encourage employees to take actions to avoid being ostracized by other team members. For example, those who are “in-group” members of the team leader should be cooperative and open to other team members, and conduct extra-role behaviors such as helping.

Limitations and future research

In addition to the findings, there are also some limitations to be noted. First, despite a multi-source design for data collection, our data is still cross-sectional. Thus the findings cannot infer any causal influence. Thus, we strongly encourage future studies to adopt a lab experiment method to further gauge the causality among our major study variables. In addition, given that individual envy emotions might shift from time to time, Experience Sampling Method (ESM) might be more accurate in investigating the dynamic within-person processes in which team members may respond differently in terms of envy intensity and frequency toward the ones who have high-quality LMX relationships (Wheeler & Reis, 1991). Therefore, we call for studies to adopt an ESM approach to further examine the dynamic part of the emotion of envy.

Second, we measured the emotion of envy in other team members using only one item due to the network measure design suggested by previous studies (Kim & Glomb, 2014). This operationalization of envy largely reduced raters’ fatigue as they had to rate several team members, but potentially led to decreased reliability. We expect that future research could manage to measure envy through a better method.

Third, contrary to our prediction, LMX is negatively associated with workplace ostracism. Specifically, the results show a negative direct effect of LMX on workplace ostracism and a positive mediation effect of LMX on workplace ostracism via being envied by team members. This implies that there are other mediators in the LMX-ostracism relationship that could account for the negative paths through which LMX negatively impacts workplace ostracism. For example, perceived power status may explain the negative association. Employees who have high-quality relationships with leaders may be seen as representing a higher power status in the team, and other team members are likely to help rather than undermine them to get reciprocal returns. We call for future research to theoretically and empirically examine other potential mediators.

Finally, there are other potential moderators except for agreeableness. For example, warmth and competence are crucial in interpersonal judgment (Fiske et al., 2007; Cuddy et al., 2011) and thus could possibly moderate the relationship between LMX and ostracism. Moreover, leader characteristics may influence the way employees perceive LMX relationships. For example, a leader with a high power distance orientation may reduce the feelings of envy of other team members toward the star followers, because team members are likely to attribute the unequal distribution of LMX relationships as a result of leader behaviors. We encourage more examinations on other moderators.