Experiencing workplace ostracism with loss of engagement

Purpose –Drawing on social exchange theory and a cultural perspective, this study examines the relationship between workplace ostracism and job engagement by focusing on the mediating role of felt obligation and the moderating role of collectivism. Design/methodology/approach –A two-wave survey was conducted over four months in a private service business in China. The participants comprised 108 Chinese employees. Findings – The results indicate that workplace ostracism has a negative relationship with job engagement through a reduced sense of felt obligation. Collectivism strengthens the main effect of workplace ostracism on felt obligation and its indirect effect on job engagement via felt obligation. Research limitations/implications – This study contributes to understanding of the internal mechanism of the workplace ostracism–job engagement model by identifying the mediating role of felt obligation. It also emphasizes that collectivist cultures can enhance the effects of workplace ostracism. However, the generalizability of our findings may be limited due to this cultural factor. Practical implications –Our findings show thatworkplace ostracismplays a significant role in reducing job engagement. Therefore, it is essential to reduce the incidence of ostracism in the workplace. Originality/value – By addressing the previously unexplored mechanism that mediates the relationship between workplace ostracism and job engagement, this study provides new directions for research on workplace ostracism and job engagement.


Introduction
Job engagement has been defined as a stable and pervasive affective-cognitive state characterized by vigor, dedication and absorption (Schaufeli et al., 2002). Basit (2017) indicated that helping employees to fully engage at work and maximize their potential can benefit organizations. However, research has suggested that a large proportion of employees cannot maintain a high level of engagement. For example, the Global Findings Report (Towers Watson, 2016) pointed out that only 37% of all employees were highly engaged in their work; the remaining 63% were either unsupported, detached or disengaged. It is therefore essential to understand the factors that hinder employees' job engagement. Many studies have examined the antecedents of job engagement, such as individual differences (e.g. Rich et al., 2010;Shuck et al., 2011), job characteristics (e.g. Saks, 2006), affective state (e.g. Bledow, 2011), perceived leadership (e.g. Carasco-Saul, 2015), perceived social support (e.g. Othman and Nasurdin, 2013) and working conditions (e.g. Basit, 2017;Chen et al., 2013).
However, the effects of relational experiences at work on job engagement have not been explored (Green et al., 2017). Ostracism is a relational and universal phenomenon experienced in personal life (Nezlek et al., 2012(Nezlek et al., , 2015 and in organizations (Ferris et al., 2008;Williams, 2007). Workplace ostracism has been defined as "the extent to which an individual perceives that he or she is ignored or excluded by others" (Ferris et al., 2008(Ferris et al., , p. 1348. In a survey conducted by Fox and Stallworth (2005), 66% of the employees had been ignored or excluded at work. Leung et al. (2011) further proposed that experiencing workplace ostracism can reduce job engagement, but its mediating and moderating mechanisms have not been explored. Thus, we focus on how and when workplace ostracism influences employees' job engagement.
We take a social exchange perspective in this study and argue that felt obligation has an important mediating effect on the relationship between workplace ostracism and job engagement. Exchange-based obligations are generated through a series of interactions between individuals in a state of reciprocal interdependence (Cropanzano and Mitchell, 2005;Eisenberger et al., 2001;Emerson, 1976;Saks, 2006). A sense of obligation can also increase the likelihood of exchanging resources for purposive action (Adler and Kwon, 2000). However, if individuals receive "silent treatment" or a "cold shoulder" (i.e. they are ostracized) instead of demonstrations of trust, loyalty and mutual commitment (Cropanzano and Mitchell, 2005), relationships become unbalanced. Feelings of personal obligation, gratitude and trust thus become less likely (Blau, 1964), resulting in disengagement from work (Green et al., 2017).
Furthermore, the phenomenon of ostracism should be considered in the specific social context, in particular the cultural setting (Mao et al., 2018). Indeed, Over and Uskul (2016) indicated that cultural characteristics can be vital cues through which individuals understand ostracism and determine their responses. Therefore, we focus on collectivism as a cultural force, which is characterized by interdependence, security, duty, group harmony, interpersonal relationships and norms that prioritize group issues or goals (Triandis, 1994(Triandis, , 2001. In this regard, China is a typical representative of a collectivist culture (Oyserman and Lee, 2008). Originating from Confucianism (Chiu and Kosinski, 1995), Chinese collectivism has influenced the values, beliefs, attitudes, norms, roles and behavior of many generations (Triandis, 1995). Individuals in collectivist cultures typically build close and interdependent relationships with others and appreciate the feeling of belonging to a group (Felfe et al., 2008). If these relationships are broken (e.g. individuals are ostracized and their close relationships threatened), those with closely interdependent relationships may suffer more than those with different self-concepts (Cross et al., 2000). Further research on the moderating effect of collectivism on ostracism is needed, but some scholars have suggested that people with a higher collectivist orientation are more sensitive to workplace ostracism, because in a collectivist society, their roles as organizational members are undermined (Leung et al., 2011;Wu et al., 2016). Thus, we aim to identify the boundary conditions of the effects of workplace ostracism, and argue that its negative impact may be greater for individuals with a more collectivist orientation. Overall, we explore the relationship between workplace ostracism and job engagement via felt obligation and propose that the process varies depending on people's level of collectivist orientation.
Our study makes several theoretical contributions. First, we explore why and how workplace ostracism is related to job engagement. Our mediating approach contributes to the literature on workplace ostracism and job engagement by revealing why and how ostracism in the workplace is a barrier to job engagement. Second, we apply social exchange theory and propose that exchange-based obligations mediate the relationship between workplace ostracism and employees' job engagement. This novel approach can inform researchers and managers regarding how job engagement can be enhanced. Third, echoing Mao et al. (2018), we identify a cultural factor (i.e. collectivism) as an important boundary condition when exploring the fluctuating effects of workplace ostracism, thus contributing to the literature on workplace ostracism and cultures. Our theoretical model is presented in Figure 1.

Theory and hypotheses
Social exchange theory Social exchange theory is an influential conceptual paradigm and is useful for interpreting workplace behavior (Cropanzano and Mitchell, 2005). This theory suggests that complex social interactions generate obligations to reciprocate the beneficial actions of others (Blau, 1964;Cropanzano and Mitchell, 2005;Emerson, 1976). Cropanzano and Mitchell (2005) proposed three foundational ideas of social exchange. First, social exchange involves rules and norms, including reciprocated and negotiated rules. Reciprocity, as a form of interdependent exchange, involves bidirectional interpersonal transactions, whereby an individual supplies a benefit and the receiving party reciprocates in kind (Gergen, 1969). Negotiated rules are aimed at reaching beneficial arrangements, such as group members negotiating their tasks and responsibilities. Second, social exchange involves an exchange of resources, which can be economic or socioemotional Foa, 1974, 1980). Economic resources are tangible assets that can be exchanged to meet financial needs, whereas socioemotional resources are symbolic and particularistic resources that can be exchanged to achieve personal and social ends. Third, social exchange relationships evolve over time when individuals care about others and thus generate beneficial interpersonal connections (Cropanzano et al., 2001).
Workplace ostracism and employees' job engagement Workplace ostracism occurs when "an individual perceives that he or she is ignored or excluded by others" in the workplace (Ferris et al., 2008(Ferris et al., , p. 1348). Ostracism has been described as a form of exclusion, rejection, interpersonal mistreatment and social death, and thus involves negative and painful psychological and physical experiences (for a review, see Mao et al., 2018). Studies have indicated that workplace ostracism can have negative effects on employees' attitudes and behavior, such as threatening their sense of belonging, control, self-esteem and meaningful existence (Ferris et al., 2015;Williams, 2001). Workplace ostracism has also been shown to reduce job satisfaction, enhance turnover intention (Ferris et al., 2008), increase psychological distress (Wu et al., 2012) and inhibit organizational citizenship behavior (e.g. Wu et al., 2016), prosocial behavior (e.g. Balliet and Ferris, 2013) and creativity (Kwan et al., 2018).
Drawing on social exchange theory, we argue that workplace ostracism undermines employees' job engagement. Relational experiences at work are important for job engagement (Green et al., 2017). Engagement is linked with high-quality connections, such as perceived  Workplace ostracism social support at work (Bakker and Demerouti, 2008;Green et al., 2017;Rich et al., 2010). Workplace ostracism isolates employees who are ostracized (Fox and Stallworth, 2005) and can have severe and unpleasant effects (Ferris et al., 2008). If employees fail to establish and maintain high-quality connections with their colleagues, their interdependent relationships may be threatened or destroyed (Kwan et al., 2018). Reactions to negative interpersonal experiences include confusion, anxiety, sadness, and even anger instead of gratitude (Williams, 2007(Williams, , 2009, which threaten physical and psychological well-being (Williams and Nida, 2011). Employees' attention is distracted from work when they focus on interpersonal relationships, preventing them from fully engaging with their roles and investing cognitive, emotional and physical efforts at work. Thus, we propose the following hypothesis: H1. Workplace ostracism has a negative relationship with employees' job engagement.

Mediating role of felt obligation
Drawing on social exchange theory, we argue that workplace ostracism reduces employees' sense of felt obligation, which in turn causes them to disengage from their work. Felt obligation is based on the norm of reciprocity, through which individuals are inclined to reciprocate the benefits they receive from others (Perugini et al., 2003). Thus, employees can feel a sense of obligation when they receive economic or socioemotional benefits from their coworkers that meet their financial or psychological needs (Cropanzano and Mitchell, 2005). However, if employees face workplace ostracism, they will feel alienated, depressed, helpless and worthless (Yaakobi and Williams, 2016b), and their sense of felt obligation will disappear. Social exchange theory has strong explanatory power in terms of job engagement (Saks, 2006) and suggests that obligations are generated in reciprocal interdependent relationships. When individuals feel a sense of obligation, they tend to respond in kind with positive attitudes and behaviors (Cropanzano and Mitchell, 2005). Thus, the obligations in reciprocal work relationships can make employees feel more meaningful and psychologically available, and they will repay these feelings with a higher level of job engagement (Kahn, 1990;Saks, 2006). Conversely, when employees have a reduced sense of felt obligation, they do not feel the desire to reciprocate by increasing their job engagement. Thus, the reduction in employees' felt obligation due to workplace ostracism may lead to reduced job engagement. Therefore, we propose the following hypothesis: H2. Employees' felt obligation mediates the negative relationship between workplace ostracism and job engagement.

Moderating role of collectivism
Ostracism can be a painful experience (Eisenberger et al., 2003), and the way people interpret and respond to it may depend on specific cultural values (Yaakobi and Williams, 2016a). Thus, for employees, we consider collectivism as a moderator of the effect of workplace ostracism on felt obligation. In collectivist societies, interpersonal relationships are a key mechanism (Wasti, 2003), and creating and developing interpersonal relationships can ensure the survival and well-being of individuals (Lieberman, 2013). Highly collectivist individuals are likely to build close and interdependent relationships with others and appreciate the sense of belonging to a group (Felfe et al., 2008). However, such individuals are also sensitive to the threat of being excluded from a group (Over and Carpenter, 2009;Wesselmann et al., 2012). Ostracism implies that individuals are punished or excluded by others. Because this type of punishment threatens their social existence, ostracized people are likely to be sensitive to being excluded and to experience various negative outcomes, such as psychological pain, and may feel that their needs are threatened (Wesselmann and Williams, 2017;Williams, 2009). When highly collectivist employees perceive that they are being ignored or rejected, they are less likely to feel obligated to act. Conversely, those with a lower sense of collectivism are less likely to be sensitive to workplace ostracism because interpersonal relationships are not their primary concern (Wasti, 2003) and they focus more on other domains (e.g. professional skills) for their survival and well-being. Thus, employees with a low sense of collectivism are less responsive to workplace ostracism and thus more likely to maintain their levels of felt obligation and job engagement. Therefore, we propose the following hypothesis: H3. Collectivism moderates the negative relationship between workplace ostracism and felt obligation, such that the relationship is stronger for more collectivist employees than for less collectivist employees.
Taken together, we propose a moderated mediation model (Edwards and Lambert, 2007). Specifically, we propose that highly collectivist employees are relatively sensitive to workplace ostracism and are thus less likely to feel obligated to engage in their work. Conversely, employees with a low sense of collectivism are less affected by workplace ostracism and thus continue to engage in their work to a similar degree. Thus, we propose the following hypothesis: H4. Collectivism moderates the mediating effect of felt obligation on the relationship between workplace ostracism and job engagement, such that the effect is stronger when the level of collectivism is high rather than low.

Procedure and participants
We collected data from a newly established (less than one year) private service enterprise in the People's Republic of China. With the help of administrative staff, we invited all of the enterprise's 137 frontline employees to participate in our investigation. The targeted participants received the coded survey questionnaire in a meeting room during work hours. One of the authors collected the data and assured the participants that their responses would be kept confidential and used only for academic purposes. The completed questionnaires were returned in sealed envelopes. We gave each of the participants a notebook as a token of our appreciation for their active participation.
The study involved a two-wave survey within four months and proceeded as follows. In the first wave of the survey (T1), we asked the 137 participants to provide demographic details, such as their age, gender, education, department tenure (Tenure D) and organizational tenure (Tenure O), and assessed their perceived workplace ostracism and collectivism. The second wave (T2) was conducted four months later, and we then asked the participants to assess the degree of their felt obligation and job engagement over the last four months. In the end, we received 108 useable questionnaires, for an effective return rate of 78.83%.
More than half (63%) of the 108 participants were women. Most (81.5%) were under 29 years old and 14.8% were between 30 and 39 years old. Most had advanced degrees: 41.7% had a bachelor's degree and 46.3% had a master's degree or above. The average organizational tenure was 7.83 months (SD 5 1.07), and the average department tenure was 7.69 months (SD 5 1.29).

Measures
Although the original measures of the key scales were developed in the West, they have been applied and validated in China, so Chinese versions of the scales are available. In addition to demographic variables, we measured the four key variables based on the participants' ratings for the following scales. In each case, the participants' responses were given on a fivepoint Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).
Workplace ostracism. We used the ten-item scale originally developed by Ferris et al. (2008) and later applied in a Chinese setting by Wu et al. (2012) to measure perceived workplace ostracism, a typical item being "Others ignore you at work." Cronbach's alpha for this scale was 0.97.
Collectivism. We used the six-item scale originally developed by Dorfman and Howell (1988) and later applied in a Chinese setting by Wu et al. (2016) to measure collectivism as a cultural orientation at the individual level, a typical item being "Group welfare is more important than individual rewards." Cronbach's alpha for this scale was 0.91.
Felt obligation. We used six items from the seven-item scale originally developed by Eisenberger et al. (2001) and later applied in a Chinese setting by Wu et al. (2016) to measure employees' felt obligation. We excluded the reverse item due to its low factor loading. A typical item is "I feel a personal obligation to do whatever I can to help the organization achieve its goals." Cronbach's alpha for this scale was 0.91.
Job engagement. We used the nine-item Utrecht Work Engagement Scale (UWES-9) originally developed by Schaufeli et al. (2006) and later applied in a Chinese setting by Chen et al. (2013) to measure job engagement. A typical item is "At my work, I feel bursting with energy" (vigor). Cronbach's alpha for this scale was 0.94.
Control variables. Following Chen et al. (2013), we controlled for the five demographic variables of gender, age, education, department tenure and organizational tenure. Age was measured in years. with 'Age had four categories: 29 or under, 30 -39, 40 -49, and 50 or above'. Gender was measured as a dichotomous variable coded as 0 for men and 1 for women. Department tenure and organizational tenure were measured in months because the organization was newly established and its employees had only been working there for a few months, the longest tenure being 10 months.
To examine common method variance bias, we conducted a statistical remedy procedure by fitting our proposed model with an orthogonal method factor (Podsakoff et al., 2003). The result showed that the average of squared loadings on the common method factor was only 1.5%, lower than the 17.2% reported by Williams and McGonagle (2016). Thus, common method variance was not a serious issue. Table 1 shows the means, standard deviations and correlations of all of the variables.

Hypothesis testing
We conducted multiple regression analyses and PROCESS analyses (bootstrap 5 1,000) to examine the direct and indirect effects of workplace ostracism on job engagement in the hypothesized model. The multiple regression results are shown in Table 2. The relationship between workplace ostracism and job engagement was negative and significant (β 5 À0.58, p < 0.01), thus supporting Hypothesis 1. The relationship between workplace ostracism and felt obligation was significant and negative (β 5 À0.46, p < 0.01), and felt obligation had a significant positive relationship with job engagement (β 5 0.71, p < 0.01). The indirect effect of workplace ostracism on job engagement via felt obligation was À0.33. In addition, the  .20], excluding 0). Therefore, Hypothesis 2 was supported. Table 2 and Figure 2 show that collectivism significantly moderated the relationship between workplace ostracism and felt obligation (β 5 À0.34, p < 0.01) and that the negative relationship between workplace ostracism and felt obligation was stronger when the level of collectivism was high (β 5 À0.52, p < 0.01) rather than low (β 5 À0.13, n.s.). This supported Hypothesis 3. The PROCESS results show that the indirect effect of workplace ostracism on job engagement via felt obligation was significant and more negative when the level of collectivism was high (β 5 À0.39, SE 5 0.09, 95% CI [À0.61, À0.23], excluding 0), but it was not significant when the level of collectivism was low (β 5 À0.10, SE 5 0.10, n.s.; 95% CI [À0.29, 0.12], including 0). This supported Hypothesis 4.

Discussion
Workplace ostracism can be a disturbing experience (Mao et al., 2018). Based on multiple regression analyses and PROCESS analyses, this study provided evidence that ostracism has a negative effect on employees' working lives. Drawing on social exchange theory and taking a cultural perspective, we tested a model of workplace ostracism and job engagement. The findings support our predictions that workplace ostracism has a negative relationship with employees' job engagement via a reduced sense of felt obligation and that collectivism moderates the effects of ostracism and the mediating mechanism between workplace ostracism and job engagement.

Theoretical contributions
By linking workplace ostracism and job engagement, we contribute to the literature on job engagement, examining why some employees are less engaged in their work than others are. Some studies have shown that most workers lack job engagement (e.g. Byrne et al., 2016), and others have suggested that stable context-specific characteristics and individual differences influence this engagement (Green et al., 2017). However, daily interactions with others are a significant but underexplored factor in job engagement (Green et al., 2017). Our study adds to the literature by demonstrating that workplace ostracism, a universal negative experience (Nezlek et al., 2012(Nezlek et al., , 2015, inhibits employees' job engagement. Our study also extends the application of social exchange theory by explaining how exchange-based obligations mediate the relationship between workplace ostracism and employees' job engagement. Social exchange theory is commonly applied as a conceptual paradigm for understanding workplace phenomena (Cropanzano and Mitchell, 2005) and suggests that social exchange involves a series of interactions that generate obligations (Emerson, 1976). Research has examined the mediating role of felt obligation in various relationships. For example, Basit (2017) showed that trust in a supervisor influences job engagement by promoting psychological safety and felt obligation, supporting Saks's (2006) view that the supervisor-subordinate relationship involves social exchange. Our finding that workplace ostracism influences job engagement via felt obligation also extends the application of social exchange theory to job engagement. This mechanism provides a more detailed and specific explanation of how workplace ostracism influences employees' job engagement.
Finally, we make an important contribution to the literature by using a cultural factor (i.e. collectivism) as a boundary condition to explore the effects of workplace ostracism and thus respond to the call made by Mao et al. (2018) for further research in this area. The specific cultural context should be considered when examining the phenomenon of ostracism (Mao et al., 2018). However, the role played by collectivism as a boundary condition for the effects of ostracism has not been fully investigated. Therefore, our study provides a meaningful and significant analysis of the effect of this key cultural factor on the links between workplace ostracism, felt obligation and job engagement.

Limitations and recommendations for future research
Our study has several limitations. First, our sample size was relatively small due to limited resources. Fortunately, the reliability and discriminant validity met measurement standards. Nevertheless, future studies should use larger samples to make the research results more reliable and rigorous.
Second, our data came from a single source and were self-reported, which may have inflated the correlations and thus increased the risk of common method bias (Podsakoff et al., 2012). As noted, the average of squared loadings on the common method factor was low, indicating that common method variance was not a serious issue. We also applied a time-lagged research design to reduce the likelihood of common method bias (Spector, 2019). The concern of common method variance could be further mitigated in future research by refining the research design, for example, by collecting multi-source data and conducting longitudinal research (Podsakoff et al., 2003).
Third, mediation analysis provides inconsistent estimates of the indirect effect due to endogeneity issues. This can only be addressed by using experimental manipulations of both the independent variable and the mediator or through the use of two-stage least-squares regression. Future studies should take steps to address these endogeneity issues.
Fourth, the generalizability of our findings may be limited. Chinese society is characterized by collectivism (Oyserman and Lee, 2008). Thus, although Chinese people prefer to join groups and build interdependent relationships, this also means that they are more likely to be hurt when excluded. Our sample from China had a mean collectivism score of 4.26, which is high and close to the maximum score and may therefore hinder the application of our findings to Western contexts. In future studies, this problem could be avoided by conducting cross-cultural research. Chinese culture is also characterized by traditionalism and power distance (Farh et al., 2007), which may be important cues that influence how individuals make sense of workplace ostracism and how they respond to it. Therefore, we suggest that future studies explore the moderating roles played by traditionalism and power distance in the relationship between workplace ostracism and employee outcomes.

Practical implications
In practical terms, our findings show that workplace ostracism is a major factor contributing to a lack of engagement in the workplace. Ostracism is costly for employees and their organizations (Wu et al., 2016), as job engagement generally has positive effects on job performance (Rich et al., 2010). Thus, managers should pay attention to this phenomenon and could also play an important role as interceders. Taking action to reduce the incidence of ostracism in the workplace is also essential. For example, interactive activities (e.g. a freestyle salon) could be organized to enhance communication.
Drawing on the social exchange perspective, our study also indicates that felt obligation, which is generated through reciprocal action, mediates the relationship between workplace ostracism and job engagement. Thus, managers should put more emphasis on reciprocal interdependence to help their employees achieve their goals (Eisenberger et al., 2001). For instance, they could assign cooperative tasks and encourage friendly cooperation instead of ostracism, as the negative effects of ostracism are bidirectional.
Finally, our study suggests that the direct and indirect effects of workplace ostracism on job engagement are more negative when feelings of collectivism are high rather than low, so employees in collectivist cultures are likely to be more sensitive to exclusion. When interdependent and close relationships are broken, collectivist individuals are more likely to feel bad and to exhibit negative attitudes and behaviors. Collectivism is deeply rooted in Chinese culture, so this is unlikely to change in the foreseeable future. Managers should therefore be aware of situational cues, as these can help them identify why ostracism originally occurs, resolve employee relationship issues, and possibly avoid such problems in the future.

Conclusion
Drawing on social exchange theory and a cultural perspective, we examine how felt obligation mediates the relationship between workplace ostracism and employees' job engagement and the role played by collectivism in moderating this effect. The findings enrich our understanding of how cultural differences affect employees' reactions to workplace ostracism. We hope that our findings will provide a theoretical basis for future research.