As is known to all, voice behavior is a kind of organizational citizenship behavior which is beneficial to the development of enterprises. And with the rapid economic development and fierce competition, organizations put forward higher requirements for employees voicing their opinions. However, in organizations, there is a common phenomenon that employees are not willing to express themselves even though they know the truth and clearly know the problem or the correct direction of solution. Once a company loses employee voice behavior, it will reduce innovation opportunities and even leads to failure (Deniz, Noyan & Ertosun, 2013)  . This phenomenon has captured the scholars’ attention of the field of organizational behavior. Morrison and Milliken (2000)  defined this collective cultural phenomenon that “most employees choose to retain their views in face of potential problems in the organization” as employee silence. On this basis, Khalid & Ahmed (2016) further deepened the definition of employee silence behavior. He pointed out the phenomenon that members of the organization know what is the truth but they are unwilling to express even though they know the problem or the correct direction of solution named employee silence. We found that early researches about employee silence behavior have reflected the organization members know about clearly the problems existing in the management process and also know how to improve, but for various reasons they chose to withhold their ideas without vocal expression of behavior (Morrison & Milliken, 2000)  . Although silent behavior can sometimes contribute to quick and effective decision-making, its negative effects are obvious in general (Knoll & van Dick, 2013)  , Employee silence may decrease organizational change and innovation and reduce employee learning motivation and knowledge sharing engagement as well. At the same time, employees’ silence behavior is also considered as an important threat to organizational change and development (Morrison & Milliken, 2000)  . Therefore, how to effectively reduce the silence behavior of organization members has become a key issue of organizational change management.
In previous studies, the process mechanism of employees’ silence behavior is mainly shown in the following aspects: Firstly, leadership styles, such as abusive leadership, narcissistic leadership and authoritarian leadership are significantly positively correlated with employee silence behavior (Song, Qian, Wang, Yang, & Zhai, 2017; Xu, Loi & Lam, 2015)   ; Secondly, the organizational situations, such as the silence atmosphere, procedural justice climate and the work place ostracism have a significant impact on the employee’s silence behavior (Rai & Agarwal 2017; Tangirala & Ramanujam, 2008)   ; Thirdly, on the individual level, for example, the individual’s perception of power, organizational identity can negatively predict employee silence behavior (Morrison, See & Pan, 2015)  . However, the existing research shows us that there is a lack of research on the generation mechanism of employee silence behavior from the perspective of interpersonal relationship.
Colleague exclusion is a form of workplace interpersonal relationship. Its concept comes from workplace ostracism. Workplace ostracism is defined that individual perceives that they are being rejected, ignored, or ostracized by another individual in their work place (Ferris et al., 2008)  . It includes the rejection of colleagues, the exclusion of leaders, and the exclusion of language and rules. Previous studies have shown that workplace ostracism has significant positive affect on employee deviant behavior, dimission behavior and silence behavior (Zheng, Yang, Ngo et al., 2015; Balliet et al., 2012; Renn et al., 2013)    , but workplace ostracism can include both co-work rejection and leader rejection, we don’t know what is the difference impact of the different exclusions. However, previous studies only discussed the impact of workplace ostracism on silence behavior in general, and there was a lack of detailed studies on different types of rejection. We assume that employees are more likely to quit than silence when they are rejected by both the leader and the employee in the workplace. If employees only choose to be silent, it is proved that the individual resources are still in the retention stage, the pressure of employees’ is less than the situation that they are rejected by both colleagues and superiors. Therefore, on the basis of theoretical analysis and literature support, this study tested the effect of colleague exclusion on employee silence in the workplace, and explored the intermediary mechanism of psychological security and the boundary effect of supervisor support. On the one hand, it can deepen the theoretical and practical understanding on the process mechanism of employees’ silent behavior s; on the other hand, it can also help to make valuable countermeasures and suggestions of managing employees’ silent behavior.
2. Theory and Hypothesis
Conservation of the resource (COR) suggests that the individual has been working hard to maintain, save and create individual resources (objects, the conditions, personal characteristics, energies); pressure will destroy individual resources and individual resources conservation follows the principle of increment spiral and decrement spiral (Hobfoll, 1989, 2001)   . That means that when individuals have more resources, they are more willing to take risks to acquire new resources, such as voice behavior, innovation behavior, and so on (Wang, Zhao et al., 2017)  . When individual resources are scarce, individuals are more willing to adopt conservative strategies to maintain resource they have. Colleague exclusion is a kind of stress source that can weaken individual resources, so that this study tried to use COR to explain the main effect of the model. What’s more, Hobfoll (2001)  also pointed out that social support resources have a positive impact on individual resources which provides a strong theoretical support for the supervisor support as boundary mechanism of this study.
2.1. Colleague Exclusion and Employee Silence Behavior
Colleague exclusion refers to the perception that an individual is ignored, rejected, isolated and other negative behaviors by colleagues. Researches had shown that this kind of rejection often brings negative emotions, such as sadness, loneliness, jealousy, guilt, embarrassment and social anxiety (Yang, 2018; Wu, Liu, Kwan & Lee, 2016)   . Workplace ostracism can create interpersonal pressure, When the pressure of interpersonal rejection is too high, it even affects employees’ self-control ability and psychological cognitive function (Williams, Forgas, 2005)  . Dealing with these stress will consumes individual psychological resources (Hobfoll, 1989)  , and Conservation of the resource (COR) suggests that the individual has been working hard to maintain, save and create individual resources (objects, the conditions, personal characteristics, energies), when individual resources are reduced, individuals are more willing to take avoidance strategy to protect their resources. Colleague exclusion is obvious a destructive behavior to hurt others and make individuals feel isolated and excluded (Ferris et al., 2008)  , It causes individual resources to be constantly depleted. At the same time, lots of studies have shown that silence is an avoidance behavior that to protect employees from punishment or injury (Deniz, Noyan & Öznur Gülen Ertosun, 2013; Dedahanov & Rhee, 2016)   . Therefore, according to the COR, when individuals perceive colleague exclusion, in order to avoid the resources loss, they are more likely to adopt a state of inactivity to avoid risks. Therefore, based on the above inference, we made the following hypothesis:
H1: Colleague exclusion can positively predict employee silence behavior
2.2. The Mediating Effect of Psychological Security between Colleague Exclusion and Silence Behavior
From the definition, psychological security can be divided into individual level and team level, On the individual level, Kahn (1990)  defined psychological security as a kind of individual character, it reflects the employees’ internal psychological state and self-awareness, staff show themselves without worrying about the negative effects of self-image, position, or occupation, when individuals have stronger feeling of that, they more likely to show themselves in the work. Ednnondson (1999)  expounded psychological security from the perspective of team, defined psychological security as the common belief that employees generally accept that is safe to take interpersonal risk, and is free from punishment, rejection and resistance when they have public discussion. From the definition, no matter the sense of psychological security at the team level or individual level, they all are the experiences of the internal psychological states, it reflects that individual feels others whether will be difficult or trust and respect themselves (Silla & Gamero, 2018; Frazier & Tupper, 2016)   , and it is a reflection of the individual’s internal resource state. According to the theory of resource conservation, stress will reduce individual resources, change resource state. And from previous researches we found that colleague exclusion can produce stress through affect relationship with co-workers (Yang, 2018)  . Therefore, we hypothesized that high colleague exclusion will reduce Individual resources and change internal resource state, which let employees’ psychological security be low.
H2a: Colleague exclusion negatively affects employees’ psychological security
The theory of resource conservation also told us that, when individual resources are sufficient, individuals are more willing to take risks to acquire more resources, while when individual resources are less, individuals prefer to take evasive actions to prevent risks and protect existing resources (Hobfoll, 1989)  . Individuals with high psychological security means that the richer the internal resource state is, according to COR, the employees are more likely to take risk behaviors and express themselves in their work. On the contrary, individuals with low psychological security means low internal resource perception, so the possibility of avoidance behavior is higher, there before, we hypothesized that following.
H2b: Psychological security negatively affects employee silence
In summary, colleague exclusion is a kind of malicious behavior that hurts individual feelings (Wu, Liu, Kwan & Lee, 2016)  , weaken individual resources and which may decrease Psychological security. Psychological is the external manifestation of individual resources, according to COR, people who with low mental resource in the less possibility to take risk to voice or other proactive behaviors but more likely to be silence, there before we guess that Psychological security negatively predict silence behavior. What’s more, according to the foregoing deduction, the aggressiveness and exclusivity of colleagues’ exclusion may lead directly to employee silence, there before we hypothesized that following.
H2c: Psychological security mediates colleague exclusion and employee silence behavior
2.3. Supervisory Support as a Moderator in the Influence of Colleague Exclusion on Psychological Security
Organization is a complex network of relationships that includes relationships with colleagues, superiors, and subordinates. In previous studies, supervisor exclusion and colleague exclusion were combined as workplace ostracism to study the impact of on employees’ behavior. However, there are situations in which colleagues and leaders have different attitudes towards individuals in real life. Therefore, this study believes that it is necessary to explore the possibility of individual silence behaviors in organizational situations when employees perceive colleague exclusion and receive support from leaders.
Eisenberger (1986)  defined supervisor support as the individual perceived the supervisor’s evaluation of their contribution to organization and supervisor’s attention on their happiness, it develops employees’ comprehensive and integrated perception of organizational support. Under the combined effect of organizational support theory and social exchange theory, relevant theories of perceived supervisor support gradually formed. A large number of scholars have positive points of perceived supervisor support (Aydın, Esra, Basım & Nejat, 2017; Jose & Mampilly, 2015)   , and believed that perceived supervisor support is an important pre-dependent variable of perceived organizational support, which has a positive impact on employees’ proactive behaviors and extra-role behaviors (Akram, Kamran et al., 2018)  . In addition, from the COR theory, the supervisor support can be seen as a social support resources to promote individual resource (Hobfoll, 1989, 2001)   . When colleague exclusion reduces individual resources and decrease individual psychological security, supervisor support as a social support resources can make up for the loss of individual resources. What’s more, supervisor support can help employees to identify their values and develop their overall cognition of organizational support (Aydın, Esra, Basım & Nejat, 2017)  . When employees perceive supervisor support at high level, they will form an individual social support system and increase employees’ sense of belonging, responsibility and happiness (García-Cabrera, Lucia-Casademunt et al., 2018)  . At the same time, it will reduce the loneliness and sadness brought by the colleague exclusion and further weaken the negative impact of the colleague exclusion on the psychological security of employees. In contrast, perceived low levels of supervisor support, the effect of colleague exclusion on psychological security may be more obvious. Based on the above analysis, this study makes the following assumptions.
Hypothesis 3: The supervisor support negatively moderates the relationship between colleague exclusion and psychological security of employees. Specifically, the higher the leader’s support, the weaker the effect of colleague exclusion on the psychological security of employees, and on the contrary, the stronger the negative effect will be.
3.1. Participants and Procedures
In this study, we used a convenient sampling method to collect the questionnaires, and the sample was employees from enterprises in major cities in China. Before the formal research, researchers explain the purpose of the study subjects and processes. We promise that the anonymity and confidentiality and ensure that the results of data will be used only for scientific research and will not cause any harmful effects for the enterprise. As soon as participators fill out the questionnaires, they submit the questionnaires directly to the researchers. To avoid homologous method bias, the researchers used tracking method to collect data at three time points. Collect the colleague exclusion behavior data in time 1, provided by the employees; Time 2 (three months after the completion of the first survey) to collect data on employees’ psychological security and perceived supervisor support provided by employees; Time 3 (three months after the end of the second survey) to collect data on employee silent behavior. Researchers distributed 248 questionnaires in total. After eliminating the invalid questionnaires, 226 valid questionnaires were finally obtained, and the recycling rate was 91.12%. The descriptive statistics showed that in terms of gender, men accounted for 55.9%, 21 - 40 years old accounted for 80.5%, mostly with bachelor’s degree or above, accounting for 64.9%, and the working years were concentrated in 1 - 4 years, accounting for 50.2%.
In order to ensure the reliability of data in this study, researchers mainly adopted mature scales used in previous studies. Likert 5-point scoring method was adopted in the responses of the scales involved in this study, and the grades from “totally inconsistent” to “totally consistent” were 1 to 5 respectively. (1_strongly disagree to 5_strongly agree). The higher the score, the higher degree of compliance.
Colleague exclusion. We measured with 7-item scale developed by Hitlan & Noel (2009)  . Typical questions include “Co-workers giving you the silent treatment”. “Co-workers making you feel like you were not a part of the organization”. Its internal consistency coefficient in our study is 0.889.
Psychological security. We measured with 5-item scale developed by Edmondson et al. (1999)  . Representative items such as: “It is difficult to get help from other members of the team.” its internal consistency coefficient in our study is 0.768.
Perceived supervisor support. We measured with 3-item scale revised by Jokisaari and Nurmi (2009)  Representative items such as: “My leaders often advise me on how to do my work.” Its internal consistency coefficient in our study is 0.833.
Silence behavior. We measured with 7-item scale which was developed by van Dyne et al. (2003) and revised by Tangirala and Ramanujam (2008)  . Sample items include “Don’t talking to your boss even if you find potential problems in the company”. The internal consistency level of this scale is found in our study is 0.818.
4. Analyses and Results
In order to test the hypothesis of the study, Mplus 7.11 was used to establish Structure Model (Preacher, Zyphur, & Zhang, 2010)  . Theoretical model includes main effect, mediating effect and moderating effect. Firstly, our study examine the discriminant validity of colleague exclusion, psychological security, leadership support, and employee silence with confirmatory factor analysis model (CFA), Then use SPSS.22 to conduct descriptive statistics on latent variables, Thirdly, this study built a moderated mediating structural equation model with Mplus 7.11 and get the path coefficients of the structural equation by coding. Finally, we checked the mediating effect and moderating effect with bootstrap according to the distribution principle.
4.1. Confirmatory Factor Analysis Model
Harman’s one-factor test shows that, the first factor without rotation explained only 20.47% of the total variation, and did not account for most of the variation of the variable in this study. Further analysis of the competition model (Table 1) shows that the four-factor model has a good fitting degree (χ2 = 306.098, df = 198, χ2/df = 1.499, CFI = 0.957, TLI = 0.950, SRMR = 0.076, RMSEA = 0.049), All the fitting indexes are superior to other alternative models (three-factor model, two-factor model and single-factor model), which indicates that the core variables in this study have good discrimination validity, and the problem of homologous variance is effectively controlled to some extent.
Table 1. Result of confirmatory factor analysis for the measures of variables.
Notes: 4-factors model: colleague exclusion; perceived supervisor support; psychological safety; silence behavior; 3-factors model: colleague exclusion; perceived supervisor support + psychological safety; silence behavior; 2-factors model: colleague exclusion; perceived supervisor support + psychological safety + silence behavior; 1-factor model: colleague exclusion + perceived supervisor support + psychological safety + silence behavior.
4.2. Statistical Analysis
SPSS.22 was used for descriptive statistical analysis of the data. Table 2 shows the mean value, standard deviation and correlation coefficient of the study variables. As can be seen from Table 2, the colleague exclusion is significantly negatively correlated with the employee psychological safety (r = −0.430, p < 0.01) and silence behavior (r = 0.340, p < 0.01). Psychological safety is negatively correlated with silence behavior (r = −0.348, p < 0.01). In addition, perceived supervisor support was significantly negatively correlated with silence behavior (r = −0.429, p < 0.01). These results provide a good foundation for the following hypothesis testing.
4.3. Result of Structural Equation Model
To test the hypothesis by using Mplus 7.11 to construct a structural equation model. In order to improve the explanatory power of model and reduce the co-linearity of the model, the researchers treated the supervisor support with grand mean centered. And control the influence of employee age, working years and other control variables. In Figure 1, path a presents the impact of colleague exclusion on psychological safety, path b shows the impact of psychological safety on silence behavior, and path c shows the direct effect of colleague exclusion on silence behavior. Path i represents perceived supervisor support’s moderating effect on path b. In addition, a × b denotes the mediation effect, and a × b + c denotes the total effect of colleague exclusion on silence behavior.
As can be seen from Table 3, the colleague exclusion has a significant positive effect on the silence behavior (γ = 0.199, p < 0.01). Hypothesis 1 is verified. The colleague exclusion has a significant negative effect on psychological safety (γ = −0.356, p < 0.01) and psychological safety has a significant negative effect on silence behavior (γ = −0.234, p < 0.001). Moreover, the direct effect of the colleague exclusion on silence behavior (γ = 0.286, p < 0.001) is significant. Thus, psychological safety plays a partial intermediary role between colleague exclusion and silence behavior. Hypothesis 2a, hypothesis 2b, and hypothesis 2c are supported.
Since this study examines the mediating effects models by Mplus, BOOTSTRAP
Table 2. Means, standard deviations, and correlations.
Notes: N = 226; **: p < 0.01; *: p < 0.05.
Table 3. All the path coefficients of MSEM.
Notes: N = 226; ***: p < 0.001; **: p < 0.01; *: p < 0.05.
Figure 1. Model.
analysis is also needed. If the 95% confidence interval does not include 0, then the mediating effect is significant (Preacher et al., 2010)  . As can be seen from Table 4, the mediating effect of psychological safety in the relationship between colleague exclusion and silence behavior was significant, the effect value was 0.117 (p < 0.01), and the 95% confidence interval was (0.038, 0.143), excluding 0. Therefore, H2c is to be further supported.
At the same time, the perceived supervisor support played a negative regulatory role between the individual’s psychological safety and silence behavior (i = −0.050, p < 0.05), so 3a was assumed to be validated. To further intuitively reflect the moderating effect of supervisor support, researchers made moderating effect diagram of perceived supervisor support with the suggestion from Preacher, Curran and Bauer et al. (2006)  . As can be seen from Figure 2, no matter what’s the extent of colleague exclusion, when perceived supervisor support is high, psychological safety is always higher than that when perceived supervisor
Table 4. Bootstrap of mediation effect.
Figure 2. Moderating effect.
support is low. In addition, when perceived supervisor support is low, the change of psychological safety is more obvious in the different levels of colleague exclusion.
The investigators also used bootstrap analysis to check the moderating effects of this study. As can be seen from Table 5, in the case of high level of perceived supervisor support (mean + 1 standard deviation), the colleague exclusion affects the silence behavior by psychological safety at the 95% CI interval value excluding 0 (0.038, 0.142), It is indicated that under high-perceived supervisor support level, the effect of colleague exclusion on silence behavior by psychological safety is supported. In the case of low level of perceived supervisor support (mean − 1 standard deviation), colleague exclusion affects silence behavior by psychological safety at the 95% CI interval value does not include 0 (0.038, 0.144), indicating that the mediating effect of colleague exclusion behavior affecting silence behavior through perceived supervisor support is still supported under low perceived supervisor support level. The above results indicate that moderated mediation effects of colleague exclusion on individual silence behavior in this study were supported under the high-low level conditions of perceived supervisor support. It clearly showing that in the relation of colleague exclusion and silence behavior, the perceived Supervisor support plays an important role as boundary conditions.
5. Summary and Future Research Prospects
This study from the perspective of organizational environment discusses the possibility of employee silence behavior in the workplace when employees gain
Table 5. Bootstrap of moderating effect.
the supervisor’s support without colleagues’ kind. Combining with the viewpoint of resource conservation theory, this paper examines the mediating effect of psychological security in the process that colleague exclusion affects individual silent behavior, and, as well as the interaction of supervisor support and colleague exclusion. In summary, after the testing hypotheses, this study found that colleagues’ exclusion has a significant positive effect on employee silent behavior. This result is consistent with the relevant literature. What’s more, we proved that psychological security has a significant negative effect on employee silence behavior, and psychological security mediates the relationship between colleague exclusion and employee silence behavior. On the other hand, perceived supervisor support negatively moderates the relation between colleague exclusion and psychological security. That is to say, when the perceived supervisor support is high, the negative relationship between the colleague’s exclusion and the employee silence is weakened, and in contrast, the negative relationship is much stronger.
From previous studies, it has been found that there have been a large number of studies on employee negative behavior from the perspective of workplace ostracism, and there also has been a mature accumulation of research on employee silence. However, a very common situation has been ignored in the research on the relationship between workplace ostracism and employee silence. That is, the exclusion that individuals face in the organizational context comes from different objects, which may reject or support. For example, employees may be ostracized by their colleagues but supported by their leaders in organizational situations, or ostracized by their leaders and supported by their colleagues. When studying workplace ostracism, previous literatures did not distinguish the different situations, but comprehensively discussed all the possible exclusion. Therefore, in this study, the influence of different rejection attitudes from leaders and colleagues on employees’ silence behaviors was studied in a more detailed manner, which made up for the shortcomings of previous studies.
Despite the strengths, there are still some limitations in this study. The exclusion in organizational context is from two aspects, organizational support also comes from two aspects, and they can compose four different cases, but this study only explores part of them. The situation is that how to change individual silence behavior when they expose to colleague exclusion with different levels of perceived supervisor support. We still lack studies on the case of supervisor exclusion interacting with different levels of colleague support to affect employee silence behavior. Therefore, on the one hand, future researchers can study the opposite of this study to explore whether supervisor exclusion interacts with different levels of colleague support and also affects employee silence behavior. On the other hand, future researchers can also explore more in-depth, comparing the two situations of leadership exclusion interacting with different levels of colleague support and colleague exclusion interacting with different levels of supervisor support in the organizational situation to examine which situation employee’s silence behavior is higher.