Employee voice is becoming increasingly critical (Wilkinson & Fay, 2011) because it assists organizations in optimizing processes and promoting effectiveness (Gollan & Wilkinson, 2007; Janssen & Gao, 2015). Interest in understanding the consequences of voice at the employee level has been expanding recently (Bashshur & Oc, 2015), as scholars have noted the importance of voice for employees who exhibit such behavior (Hung, Yeh, & Shih, 2012; Timming & Johnstone, 2015). However, the knowledge of voice is limited by the inconclusive findings on the relationship between voice and employee performance (Bashshur & Oc, 2015; Kim, MacDuffie, & Pil, 2010; Wood & Wall, 2007). Although several studies have shown positive associations between voice and employee performance (e.g., Bryson, Charlwood, & Forth, 2006; Chen & Hou, 2016; Ng & Feldman, 2012; Whiting, Podsakoff, & Pierce, 2008), other studies report negative associations (e.g., Hung et al., 2012; Seibert, Kraimer, & Crant, 2001). One possible reason for these mixed findings is that extant research has largely focused on the amount of voice, thereby ignoring the types of messages and information delivered by voice. The effect of voice may not only depend on the amount of voice but also on its types (Morrison, 2011); different types of voice may have varying effects on employee performance (Hung et al., 2012; Maynes & Podsakoff, 2014). Hence, the current study aims to analyze the voice–employee performance relationship by considering the types of voice.

Liang, Farh, and Farh (2012) used the content of voice as the basis of the promotive–prohibitive voice framework, which has been popularly applied in the voice literature (e.g., Kakkar, Tangirala, Srivastava, & Kamdar, 2016; Qin, DiRenzo, Xu, & Duan, 2014). Promotive voice refers to “employees’ expression of new ideas or suggestions for improving the overall functioning of the work unit or organization” (Liang et al., 2012, p. 74). By contrast, prohibitive voice refers to “employees’ expressions of concern about work practices, incidents, or employee behavior that are harmful to their organization” (ibid, p. 75). Although both promotive and prohibitive voice aim to benefit the teams where the employee works (Kong, Huang, Liu, & Zhao, 2017), they are different in content, aim, and function (Liang et al., 2012). Promotive and prohibitive voice are expected to differently influence employee performance, which is defined as the performance in tasks that are specifically related to the job description (Welbourne, Johnson, & Erez, 1998).

On the one hand, given that the good intention behind promotive voice can be easily recognized (Kong et al., 2017), leaders will represent the team (Blanc & Gonzalez-Roma, 2012) to reciprocate employees by providing task-related support (Cheng, Lu, Chang, & Johnstone, 2013). On the other hand, the attention of employees who are overly involved in promotive voice may be diverted from their immediate tasks to the team’s future ideal state (Dane, 2011; Smallwood & Schooler, 2006), which is the aim of promotive voice (Ward, Ravlin, Klaas, Ployhart, & Buchan, 2016). Moreover, excessive promotive voice conveys numerous new ideas and clear solutions that are generated by consuming resources (i.e., time) available for tasks (Qin et al., 2014). Excessive information also makes leaders suffer from information overload, become numb (Bettis-Outland, 2012; Fukukura, Ferguson, & Fujita, 2013), and end up decreasing support for tasks. By integrating social interaction and the resource process, we assume that the relationship between promotive voice and employee performance may follow an inverted U-shaped trajectory. The reason is that employees who exhibit moderate promotive voice gain support from their leaders and can afford time and attention resources to achieve task goals.

Prohibitive voice is problem- and past-oriented by identifying the harmful factors that exist within teams without necessarily offering clear and novel solutions (Liang et al., 2012). Employees with prohibitive voice will engage in behavior and use vigilant strategies at work to avoid loss (Lin & Johnson, 2015). Consequently, the levels of vigilance and accuracy at work are promoted, while the error rate is reduced (Förster, Higgins, & Bianco, 2003). Given that performing well is regarded as no-loss (Wallace & Chen, 2006), employees with prohibitive voice will also have a high level of work motivation to retain this no-loss state. Moreover, prohibitive voice precisely serves as a channel to express dissatisfaction (Avery & Quiñones, 2002) by identifying harmful factors, thereby contributing to employees’ motivation regarding task goals (Tsai, Chen, & Liu, 2007). Collectively, prohibitive voice is expected to be positively and linearly related to employee performance because this voice assists employees in enhancing work motivation and accuracy.

Furthermore, we contend that power distance orientation, defined as the extent to which a person accepts the inequality of power distribution within teams (Lian, Ferris, & Brown, 2012), is a moderator in the aforementioned relationships. Power distance orientation has a special relevance to voice research (Botero & Van Dyne, 2009), owing to its implications for understanding the recipient’s reactions when voice is expressed (Bashshur & Oc, 2015). The extant literature (e.g., Kirkman, Chen, Farh, Chen, & Lowe, 2009; Lian et al., 2012) has highlighted the important role of power distance orientation in the interactive process with leaders. Leaders are the prime recipients of employee voice, thereby resulting in the variation in the influencing process of voice with the level of power distance orientation. Employees with high power distance orientation treat leaders with considerable worship and respect because the former believes that the latter possesses superiority and status (Kirkman et al., 2009). Thus, we argue that employees with high power distance orientation substantially depend on (Wang, Mao, Wu, & Liu, 2012) and create added value from the leaders’ support, which can be gained via promotive voice. These employees are likely to generate negative emotions about themselves when prohibitive voice is expressed because such a voice accompanied by criticism of supervisory behaviors (Gross & John, 2003; Liang et al., 2012) is inconsistent with their power value belief. We theorize that the effects of promotive and prohibitive voice on employee performance depend on power distance orientation.

This study makes four important contributions. First, prior findings about the effect of voice on employee performance are inconclusive, showing positive (e.g., Ng & Feldman, 2012; Whiting et al., 2008) and negative (e.g., Hung et al., 2012; Seibert et al., 2001) relationships. Separating from previous empirical works, we apply the promotive–prohibitive voice framework (Liang et al., 2012) and propose a curvilinear and a linear linkage between the different types of voice and employee performance. A possible reason for the apparently paradoxical findings on the relationships that link voice and individual performance is provided. Moreover, doing so may guide practitioners and promote an in-depth understanding of when and which type of voice will lead to improved employee performance.

Second, although the antecedents of promotive and prohibitive voice have been the focus of empirical research (e.g., Kakkar et al., 2016; Loi, Ao, & Xu, 2014; Ward et al., 2016), information is limited regarding their unique consequences (Lin & Johnson, 2015). The introduction of employee performance as an outcome has enriched the existing research on the consequences of promotive and prohibitive voice.

Third, studies have been conducted on the establishment of boundary conditions from a cultural perspective to explain the consequences of voice (Hung et al., 2012). Bashshur and Oc (2015, p. 1549) stated that “perhaps most notable is the shortage of work exploring the role of culture in voice.” The analysis of the cultural moderators of the voice–employee performance relationship responds to these calls. Specifically, power distance orientation has been argued to be an important factor in the interactive process between leaders and employees (Kirkman et al., 2009; Lian et al., 2012). In the current study, leaders are the prime recipients of employee voice. Power distance orientation may affect employees’ perceptions of leaders’ reactions to voice. Thus, the effects of voice may be bound by employee power distance orientation, thereby adding to the expanding research on voice and culture.

Lastly, culture value is key to understanding employees’ reactions to their work (Gelfand, Erez, & Aycan, 2007; Kirkman, Lowe, & Gibson, 2006). Thus, managing employees with different cultural backgrounds has been an immense challenge for organizations. Indeed, due to the growing foreign investment in China, China has been the focus of recent management research (Wang et al., 2012). The Chinese society is characterized by high levels of power distance (Hofstede, 1980). The introduction of power distance orientation as a moderator has deepened our understanding of the voice–performance relationship among Chinese employees and provided important guidance for managing employees from other high power distance countries (e.g., Japan, Korea, and the Philippines).

Theoretical development and hypotheses

Voice can occur directly or indirectly. Indirect voice exerts an influence via employee representatives, such as unions and nonunion representative voice (Bryson, 2004), whereas direct voice describes the degree to which an employee directly influences management (Kim et al., 2010). An increasing number of studies have analyzed the relationship between direct and indirect voice and organizational outcomes in the fields of human resource management (HRM), industrial relations (IR), and organizational behavior (OB) (Wilkinson & Fay, 2011). The HRM literature has highlighted the role of voice on performance (Kim et al., 2010). In a study of 18 cases from the UK and Ireland, Wilkinson and his colleagues (Wilkinson, Dundon, Marchington, & Ackers, 2004) suggested that voice could improve general productivity and individual performance through great teamwork and little absenteeism. Voice also aids in identifying and managing problems and positively impacts quality and productivity (Gollan & Wilkinson, 2007). In the IR literature, voice has been deemed to benefit the company and workforce. Bryson (2004) found that nonunion voice is more effective than union voice in eliciting managerial responsiveness in British workplaces and that direct voice is more effective than indirect voice. Kim et al. (2010) empirically found that the positive impact of team voice or worker representative voice on labor productivity is stronger at low levels of other types of voice.

In OB research, voice has been linked to several organizational outcomes, including learning, adaptability, decision making, and performance (Morrison & Milliken, 2000). At the individual level, research has revealed the relationships between voice and several outcomes (e.g., Cheng et al., 2013). The most frequently studied outcome is employee performance (Bashshur & Oc, 2015). However, scholars have not yet reached a consensus. Studies have found positive (e.g., Bryson et al., 2006; Whiting et al., 2008) or negative (e.g., Hung et al., 2012; Seibert et al., 2001) relationships between voice and employee performance. The types of voice may account for this inconsistency because previous research has largely focused on the amount of voice while ignoring its different types (Morrison, 2011).

Liang et al. (2012) used a broad conceptualization of voice (Van Van Dyne, Ang, & Botero, 2003) as the basis of their promotive–prohibitive voice framework, which has been popularly adopted by other studies (e.g., Ward et al., 2016). Promotive voice expresses innovative suggestions for improving team effectiveness, whereas prohibitive voice identifies factors that harm teams. Despite the common goal of benefiting the teams where the employee works (Kong et al., 2017), remarkable differences exist between the two forms of voice. For example, promotive voice is suggestion-oriented, which is denoted by the expression of new suggestions and clear solutions for improving team function (Liang et al., 2012). By contrast, prohibitive voice is problem-oriented, which is denoted by the expression of concerns regarding factors that serve as a check on team development without necessarily providing clear solutions or new suggestions (ibid). Moreover, promotive voice is future-oriented, which is aimed at ideal states, whereas prohibitive voice is past-oriented, which is aimed at addressing existing problems (Svendsen, Jønsson, & Unterrainer, 2016).

On account of their distinctions, the results of promotive and prohibitive voice may be unique, yet the issue regarding their unique consequences (e.g., employee performance) remains an important area of inquiry (Lin & Johnson, 2015). To address this research gap and enhance our understanding of the voice–employee performance relationship, we aim to test the effects of promotive and prohibitive voice on employee performance.

Promotive voice and employee performance

Employee performance is expected to improve as promotive voice increases from low to moderate levels because this voice promotes the social process through which the actor can gain the leader’s task-related support. Specifically, social exchange theory suggests that when one party behaves in a manner that will benefit the other party, the latter will reciprocate the former through actions that favor of the initiating party (Cheng et al., 2013). Promotive voice is a type of pro-social action that aims to help teams adapt to a changing environment and obtain innovative and successful work outcomes (Mo & Shi, 2016). When employees engage in promotive voice, the good intention behind this voice can be easily recognized (Liang et al., 2012). As the primary recipient, leaders will represent the teams (Blanc & Gonzalez-Roma, 2012) to provide task-related information and emotional support to employees who demonstrate promotive voice (Cheng et al., 2013), thereby contributing to employee performance. By contrast, employees with minimal promotive voice do not develop solid reciprocal relationships. Thus, they will experience difficulty in securing the leader’s assistance, feedback, and advice regarding tasks or receiving the leader’s understanding during peak work times. Accordingly, the performance of employees who exhibit minimal promotive voice is expected to be poor.

On the other hand, the performance of employees who overly engage in promotive voice is also expected to be poor because in the case of excessive promotive voice, they lose focus on their core task activities and lack sufficient time to fulfill their tasks successfully, and leaders’ support for their tasks reduces. First, a high level of promotive voice distracts employees’ attention from their core tasks. Employees have limited attention resources, which must be distributed between on- and off-task activities (Shaw, Zhu, Duffy, Scott, Shih, & Susanto, 2011). Promotive voice aims to achieve the future ideal states of teams (Mo & Shi, 2016). When employees are overly involved in promotive voice, they are imagining what and how to achieve the future ideal states of teams (Qin et al., 2014). In this case, their mind veers away from their immediate tasks and explores ideas for achieving the team’s future ideal state in an undisciplined manner (Dane, 2011; Smallwood & Schooler, 2006). That is, the actor’s thoughts are off-task at the moment when promotive voice reaches a high level. Thus, the actor’s considerable attention has been allocated to the team’s future development rather than the actor’s primary tasks during the generation process of promotive voice. Moreover, these actors have difficulty reconcentrating on their task activities immediately after engaging in a high level of promotive voice because their suggestions are focused on the team’s future (Liang et al., 2012) rather than the task at hand. Consequently, employees with a high level of promotive voice may be unable to concentrate on and complete their core task activities, thereby achieving poor performance.

Second, a high level of promotive voice leads to a shortage in time for completing tasks. The voice literature (Qin et al., 2014) has suggested that promotive voice consumes substantial time to generate innovative ideas and solutions that guide the entire team to achieve the ideal states. The time available for tasks decreases with an increase in promotive voice. Resource allocation theory argues that activities to which less time is allocated should lead to decreased performance (Rapp, Bachrach, & Rapp, 2013; Schmidt & Dolis, 2009). Thus, employees who are overly engaged in promotive voice may exhibit poor performance because they do not have sufficient time to complete their tasks successfully.

In addition, a high level of promotive voice results in the reduction of leaders’ support for the achievement of job goals. Promotive voice is accompanied by new and constructive information intended to improve the teams’ effectiveness (Liang et al., 2012). When employees engage in a high level of promotive voice, excessive new and constructive information is transmitted to their leaders. In this condition, leaders will experience information overload and suffer from paralysis (Bettis-Outland, 2012; Fukukura et al., 2013) and thus ignore the benefits of employees’ promotive voice. Consequently, leaders will provide diminishing task-related information and emotional support to actors who express promotive voice. Thus, employees who are overly involved in promotive voice may perform poorly.

Low or high levels of promotive voice may lead to social (i.e., leaders’ support) or resource (i.e., time and attention) constraints that harm performance. We argue that employees who engage in a moderate level of promotive voice are likely to elicit high levels of performance. They are likely to obtain the leader’s assistance, feedback, and advice regarding tasks or receive understanding during peak work times, thereby leading to improved performance. Employees are able to concentrate on immediate task activities and have time to complete immediate tasks when they exhibit moderate promotive voice. Recent evidence (Rapp et al., 2013) on organizational citizenship behavior (OCB) and task performance supports this expectation. Specifically, evidence demonstrates that moderate levels of OCB are related to high levels of task performance because employees who engage in moderate OCB obtain support from coworkers and have time to perform their tasks (ibid). Thus, employees who show moderate promotive voice not only receive leaders’ support but also have sufficient time and attention to complete their work successfully, thereby gaining a high level of performance. The relationship between promotive voice and employee performance is expected to follow an inverted U-shaped pattern. Hence, we predict the following:

Hypothesis 1

Promotive voice has a curvilinear relationship (inverted U-shaped) with employee performance.

Prohibitive voice and employee performance

Prohibitive voice is expected to be positively related to employee performance because the former enables employees to enhance their work motivation and increase their accuracy at work. First, prohibitive voice aims to stop harmful factors or protect the team from them (Liang et al., 2012). Thus, employees with prohibitive voice primarily aim for the timely prevention of losses. These employees are sensitive to losses and behave in such a manner to avoid loss (Lin & Johnson, 2015). In general, performing well is regarded as no-loss because such action can reduce negative consequences (Wallace & Chen, 2006). To achieve this no-loss state, employees who express prohibitive voice will have strong work motivation and thus achieve improved performance. Moreover, research has suggested that actual problems may trigger employees’ dissatisfaction with their work (Zhou & George, 2001). Prohibitive voice serves as a channel by which employees can vent their discontent at work (Avery & Quiñones, 2002) and through which they can identify existing problems (Lin & Johnson, 2015). Their emotional dissatisfaction decreases with the increase in satisfaction. In this case, employees may set high goals and have strong motivation to meet task goals (Tsai et al., 2007). Hence, employees with prohibitive voice are motivated to meet performance expectations and ultimately achieve a high level of employee performance.

Second, to detect and avoid problems that may result in loss, failure, or punishment, employees who are involved in prohibitive voice will resort to vigilant strategies at work (Lin & Johnson, 2015). Consequently, the levels of vigilance and accuracy at work are promoted, thereby decreasing the possibilities of committing errors (Förster et al., 2003) and eventually improving employee performance. Moreover, prohibitive voice aims to solve the real underlying problems within teams (Liang et al., 2012). Therefore, employees who utilize prohibitive voice should master the ability to provide valid information on harmful behaviors or practices through observation of colleagues’ work and reflection of their present work and experiences from previous projects. During this process, employees who engage in prohibitive voice do not disregard their work but rather pay considerable attention to it. Furthermore, they can recognize the existence of issues that are obstacles to work efficiency in advance. Given this advantage, mistakes embedded in the task process can be prevented or corrected. Thus, employees who demonstrate prohibitive voice can act in a manner that results in a low error rate while performing tasks, thereby eventually leading to an increase in employee performance. Hence, we predict the following:

Hypothesis 2

Prohibitive voice is positively related to employee performance.

Power distance orientation

Power distance stems from Hofstede (1980) and is an important culture value in organizational research (Lian et al., 2012). Hofstede (1980) claimed that culture value research is meaningful only at the societal level. However, researchers have noted that culture values substantially vary among individuals (Lin, Wang, & Chen, 2013). Daniels and Greguras (2014) reviewed the empirical research on power distance and found that the majority of studies manipulate power distance at the individual level rather than analyzing it at the societal level. To distinguish power distance at the individual level from that at the societal level, power distance orientation is used to represent an individual-level concept (see Kirkman et al., 2009) and is defined as “the extent to which one accepts the legitimacy of unequally distributed power in institutions and organizations” (Lian et al., 2012, p. 108). Previous studies have shown that power distance orientation influences individual reactions to and perceptions of authority (e.g., Farh, Hackett, & Liang, 2007; Lin et al., 2013). For example, employees with high power distance orientation legitimize the unequal distribution of power and regard authorities as superior and elite (Kirkman et al., 2009). Accordingly, employees with high power distance orientation are willing to trust, respect, and defer to their leaders (De Luque & Sommer, 2000). Moreover, these employees hold the belief that an individual should accept and obey their leaders’ decisions instead of contradicting them (Bochner & Hesketh, 1994).

Previous research has indicated that promotive and prohibitive voice have important and varying implications during interactions with leaders (e.g., Kong et al., 2017; Liang et al., 2012; Loi et al., 2014). Power distance orientation influences the interpretation and evaluation of social processes in relation to authorities (Lin et al., 2013). The employee silence literature (Timming & Johnstone, 2015; Van Van Dyne et al., 2003) has also acknowledged the potential moderating role of individual differences. Building upon work on voice, silence, and power distance orientation, in this study, power distance orientation is used as a moderator to completely understand the relationships between promotive and prohibitive voice and employee performance.

Power distance orientation as a moderator of the promotive voice–employee performance relationship

We previously argued that promotive voice, which leads to increased employee performance, triggers social processes through which employees obtain support from leaders to achieve task goals. Employees with high power distance orientation believe that their leaders possess superiority and higher status (Kirkman et al., 2009). As the inferior party, these employees are more inclined to believe that leaders control crucial work-related resources, thereby placing more value on resources that are obtained from the leaders (Wang et al., 2012). Thus, when promotive voice brings about leaders’ support for tasks, employees with high power distance orientation are more effective in utilizing these resources. Consequently, employees characterized by high power distance orientation can extract more value from promotive voice, thereby amplifying the positive effects of promotive voice on performance. Therefore, employees with high power distance orientation are expected to reach a higher level of performance after which promotive voice negatively affects employee performance.

By contrast, employees with low power distance orientation place more emphasis on egalitarianism (Loi, Lam, & Chan, 2012) and hold a stronger belief that they are equal to the leaders in status (Lin et al., 2013). These employees, compared with those with high power distance orientation, may attach less importance to task-related support from leaders accompanied by promotive voice and be less likely to utilize these resources effectively. As such, employees with low power distance orientation may gain less value for their performance via promotive voice. Thus, we expect that the threshold for the point at which promotive voice begins to negatively affect employee performance is higher when power distance orientation is high than when it is low. Therefore, we predict the following:

Hypothesis 3:

Power distance orientation moderates the relationship between promotive voice and employee performance, such that the inverted U-shaped relationship is more pronounced when power distance orientation is high.

Power distance orientation as a moderator of the prohibitive voice–employee performance relationship

We also argue that prohibitive voice is conducive to employee performance. People with high power distance orientation firmly believe that one should be submissive to leaders’ decisions (Lian et al., 2012). Moreover, disagreement with leaders is tantamount to insubordination (Vitell, King, Howie, Toti, Albert, Hidalgo, & Yacout, 2016). Prohibitive voice includes a direct or implied criticism of supervisory behaviors and policies (Liang et al., 2012). Thus, such a voice is inconsistent with employees’ power distance beliefs. This inconsistency between inner values and outer behavior may lead to increased negative emotions about oneself (Gross & John, 2003). Moreover, people negatively interpret their external environment when they have increased negative emotions (Rothbard & Wilk, 2011). Accordingly, compared with those with low power distance orientation, maintaining positive emotions is more difficult for employees with high power distance orientation, thereby weakening the effect of prohibitive voice on employee performance.

Conversely, employees with low power distance orientation place more emphasis on egalitarian and are less likely to completely obey authority (Loi et al., 2012). These employees prefer frequent and open communication with leaders (Kirkman et al., 2009). For low power distance individuals, one’s criticism of and disagreement with authorities is appropriate (Farh et al., 2007; Lin et al., 2013). Although prohibitive voice contains criticism of leaders’ behaviors and policies (Liang et al., 2012), employees with low power distance orientation are less likely to feel discomfort because such a voice fits more with their beliefs of power. Accordingly, these employees can better concentrate on tasks and maintain positive emotions. Therefore, we expect that prohibitive voice has a stronger effect on employee performance among employees with low power distance orientation. Thus, we predict the following:

Hypothesis 4:

The relationship between prohibitive voice and employee performance is moderated by power distance orientation, such that the relationship is weaker when power distance orientation is high.

Methods

Procedures and sample

Data were collected from Anhui Province in China. We selected 39 organizations (through convenience sampling) from the “Yellow Pages” directories, which mainly operated in traditional manufacturing, business services (e.g., tourism, banking, and insurance), construction and real estate, and high-tech industries (e.g., the bio-pharmacy, communication, aerospace, and electronic industries). To make this research more convincing, we first sought the assistance of a responsible person in a local government department who had personal contact with the heads of these organizations. Our researchers formally invited the heads who showed interest in this research. We explained the purpose of this study and requested assistance to recruit leader–employee dyads in their organizations through voluntary participation. Overall, 36 organizations participated in our study. Each portfolio contained several employee questionnaires and a leader questionnaire. Each questionnaire had a researcher-assigned identification number that was used to match the employees’ responses with their direct leaders’ evaluations. Certain explanatory materials in each questionnaire emphasized voluntary participation, ensured the confidentiality of information, and verified the identity of the researchers who are from a university in southeastern China. The employees were asked to report their promotive and prohibitive voice, power distance orientation, and personal demographic information. The leaders were asked to evaluate the performance of the employees under their supervision and provide personal demographic information. The employees and their leaders completed the surveys at separate locations during work. To ensure confidentiality, each participant was instructed to place his/her questionnaire in a sealed envelope and return the envelope to the researchers directly.

Paper-and-pencil questionnaires were distributed to 721 employees and 132 leaders. After eliminating missing data and excluding unmatched questionnaires, our final data included information on 431 employees rated by 80 leaders. The final response rates were 75.73% and 60.61% for the employees and leaders, respectively. Among the employees, 229 (55.9%) were more than 33 years old, 140 (32.5%) were female, 221 (51.3%) had obtained a bachelor’s degree, and 22 (7.4%) had obtained a master’s degree or higher. Among the leaders, 39 (48.8%) were less than 40 years old, 25 (31.3%) were female, and 45 (56.2%) had obtained a bachelor’s degree.

Measures

All scales were derived from existing empirical research. In line with translation and back-translation procedures (Brislin, 1986), all English measurements were translated into Chinese and then back into English by different professors in the management field. To evaluate the convergent validity of our measurement, we employed a four-factor, principal component analysis followed by varimax rotation (Aupperle, Carroll, & Hatfield, 1985; Van Dyne & LePine, 1998). Our analysis produced four factors, namely, promotive voice, prohibitive voice, power distance orientation, and employee performance. Table 1 provides all the measurement items and their loading values.

Table 1 Measurement items and loading
Full size table

Promotive and prohibitive voice

Employees used a 10-item scale to assess their promotive and prohibitive voice. The adopted scale, which was developed by Liang et al. (2012; 1= strongly disagree; 5= strongly agree), included 5 promotive items and 5 prohibitive items. Sample promotive items are “Proactively develop and make suggestions for issues that may influence the unit” and “Make constructive suggestions to improve the unit’s operation” (α = .90). Sample prohibitive items are “Advise other colleagues against undesirable behaviors that would hamper job performance” and “Proactively report coordination problems in the workplace to the management” (α = .80).

Employee performance

Employee performance was rated by the direct leader using the subscale of the Role-based Performance Scale (RBPS) developed by Welbourne et al. (1998). This scale consists of four items. Example items are “Quantity of work output” and “Accuracy of work” (1 = needs much improvement; 5 = excellent; α = .74).

Power distance orientation

Following prior individual-level research (e.g., Kim & Leung, 2007; Kirkman et al., 2009), the employees assessed power distance orientation using an eight-item measure from Earley and Erez (1997). Due to convergent validity, we excluded three items with loading values below .5 based on previous research (e.g., Umphress, Bingham, & Mitchell, 2010; Wu & Wang, 2006). Sample items were “Employees should not express disagreement with their managers” and “Managers should be able to make the right decisions without consulting with others” (1 = strongly disagree; 5 = strongly agree; α = .77).

Controls

The demographic variables of the employees (i.e., gender, age, and education level) were assigned as the control variables. Specifically, age is considered to be related to performance because it reflects accumulated work experience (Shirom, Shechter Gilboa, Fried, & Cooper, 2008). In the workplace, men typically receive more favorable performance evaluations than women (Watkins, Kaplan, Brief, Shull, Dietz, Mansfield, & Cohen, 2006). Thus, gender is important for performance outcomes. Previous research has suggested that educational level is crucial for job performance (Janssen, 2001). Accordingly, the employees’ education was also controlled for in this study.

Data analysis

The current study adopted moderated hierarchical regression analyses, which was in line with previous studies (e.g., Lam, DeRue, Karam, & Hollenbeck, 2011; Qin et al., 2014) that tested the moderating effects of curvilinear relations. Table 4 displays the results of our regression analyses. All independent variables were standardized, thereby reducing multicollinearity (Aiken & West, 1991).

Results

Table 2 presents the means, standard deviations, and correlations of the study constructs and the square roots of the average variance extracted (AVE). Prior to testing the hypotheses, confirmatory factor analyses (CFA) and AVE were conducted to rate the discriminant validity of our survey measures. The CFA results provided in Table 3 indicate that the four-factor model (i.e., promotive voice, prohibitive voice, employee performance, and power distance orientation) had a satisfactory fit (χ2(146) = 409.11, p < .001, RMSEA = .066, CFI = .95, IFI = .95, NFI = .91) that was better than that of the models with fewer factors. Furthermore, Table 2 shows that the AVEs for the four constructs were higher than the recommended .50 level (Hair, Anderson, Tatham, & Black, 1992; Lam, Huang, & Chan, 2015). These findings documented the discriminant validity of the four constructs.

Table 2 Means, standard deviations, correlations, and square roots of AVE in diagonals
Full size table
Table 3 The CFA results for various models
Full size table

Hypothesis testing

Table 4 summarizes the regression analysis results for testing Hypotheses 1–4. The variance inflation factors (VIFs) ranged from 1.04 to 2.10, indicating that multicollinearity was not a major concern. To test Hypothesis 1, we incorporated promotive voice and its quadratic term into the model (see Model 3). The coefficient of the quadratic term for promotive voice was significant and negative (β = −.13, p < .05), indicating that an inverted U-shaped relationship existed between promotive voice and employee performance. We graphed this result following the guidelines of Aiken and West (1991). Figure 1 shows that employee performance increased with an increase in promotive voice. Upon reaching a certain level of promotive voice, performance peaked and declined thereafter as promotive voice further increased.

Table 4 Results of hierarchical regression analysis
Full size table
Fig. 1
figure 1

The Relationship Between Promotive Voice and Employee Performance

Full size image

To test the inverted U-shaped relationship between promotive voice and employee performance, we computed the simple slopes of the curve in conditions of low (2 SD below the mean), medium (mean), and high (2 SD above the mean) levels of promotive voice. When the levels of promotive voice were low, medium, and high, the simple slope coefficient values of the curve were .55 (p < .001), .23 (p < .001) and − .08 (ns), respectively. This slope analysis confirmed that the positive slope apparent at low and medium levels of promotive voice turned at high levels of promotive voice, further supporting the inverted U form of the overall relationship. Thus, Hypothesis 1 is supported.

Hypothesis 2 predicted that a positive relationship exists between prohibitive voice and employee performance. To test this hypothesis, we entered prohibitive voice and its quadratic term into the model (see Model 3). The coefficient of prohibitive voice was significant and positive (β = .13, p < .01), and its quadratic term was nonsignificant (β = .02, ns), indicating that prohibitive voice was positively and linearly related to employee performance. Hence, Hypothesis 2 is supported.

We hypothesized that the inverted U-shaped relationship between promotive voice and employee performance (Hypothesis 3) and the linear relationship between prohibitive voice and employee performance (Hypothesis 4) were moderated by individual power distance orientation. To test these hypotheses, we entered individual power distance orientation (see Table 4, Model 4) and its interactive term with promotive and prohibitive voice (see Table 4, Model 5, Model 6) into the regression equation.

Table 4 (Model 6) shows that the interaction term for power distance orientation and the linear term of promotive voice were statistically nonsignificant (β = .04, ns). However, the interaction term for power distance orientation and the squared term of promotive voice were significant (β = −.14, p < .05). Thus, Hypothesis 3 is supported.

We analyzed the simple slopes of the regression curves that corresponded to all possible combinations of very high (2SD above the mean), high (1SD above the mean), moderate (mean), low (1SD below the mean), and very low (2SD below the mean) levels of promotive voice with high (1SD above the mean) and low (1SD below the mean) levels of power distance orientation (Aiken & West, 1991). Table 5 shows that among individuals with high power distance orientation, the slope for promotive voice was positive at very low (β = .29, p < .001), low (β = .19, p < .001) and medium (β = .09, p < .01) levels of promotive voice, whereas the slope coefficient value was negative at high (β = −.02, ns) and very high (β = −.12, ns) levels of promotive voice. The slope coefficient for promotive voice for employees with low levels of power distance orientation was positive at all levels of promotive voice (see Table 5). Figure 2 shows that the relationship between promotive voice and employee performance followed an inverted U-shaped pattern in cases of high (1 SD above the mean) and low (1 SD below the mean) levels of power distance orientation.

Table 5 Tests of simple slope
Full size table
Fig. 2
figure 2

Interaction between promotive voice and power distance orientation on employee performance

Full size image

The interaction term for power distance orientation and the linear term of prohibitive voice was statistically significant (β = −.11, p < .05). However, the interaction term for power distance orientation and the squared term of prohibitive voice was not significant (β = −.03, ns). This finding indicated that the positive relationship between prohibitive voice and employee performance was stronger when the level of power distance orientation was low, which was consistent with Hypothesis 4. To illustrate this finding in depth, we represented the interactions of prohibitive voice and power distance orientation following the graphing suggestions of Aiken and West (1991). Figure 3 shows that prohibitive voice was more positively related to employee performance when power distance orientation was low (β = .28, p < .01) than when it was high (β = .16, p < .05). Accordingly, Hypothesis 4 is supported.

Fig. 3
figure 3

Interaction of prohibitive voice and power distance orientation on employee performance

Full size image

Discussion and conclusions

We analyzed the relationships among promotive voice, prohibitive voice and employee performance and whether power distance moderated these relationships. As expected, the results showed that promotive and prohibitive voice have inverted U-shaped and positive linear relationships with employee performance, respectively. Power distance orientation moderates these relationships.

Theoretical implications

This study makes several contributions to the voice literature. First, our results indicate that the relationship between promotive voice and performance is curvilinear, whereas that between prohibitive voice and employee performance is positive and linear. These findings demonstrate that the voice–employee performance relationship can be better understood by considering the types of voice. The content conveyed by voice is vital for predicting employee performance, supporting the assumption of Morrison (2014) that the content of voice may influence outcomes. Note that the existing findings on voice and individual performance are mixed (Bashshur & Oc, 2015; Kim et al., 2010). Given its constructive nature, voice is believed to be positively linked to individual performance (e.g., Bryson et al., 2006; Chen & Hou, 2016). However, Hung et al. (2012) and Seibert et al. (2001) documented a negative relationship. The present study may provide elementary evidence that the effect of voice on performance outcomes can be complex and that the types of voice may be a factor. Our findings may provide a possible explanation for the inconsistent and apparently paradoxical findings on the relationships that link voice and individual performance. Our study also broadens the research on the voice–employee performance relationship by focusing on the types of voice. The current research paves the way for future research to explore the effects of the types of voice on employee performance. Moreover, our adoption of the promotive–prohibitive voice framework has enabled us to respond to recent calls (e.g., Morrison, 2011; Song, Wu, & Gu, 2017) to substantially consider different types of voice.

Second, our findings concerning employee performance demonstrate the distinction between promotive and prohibitive voice. In particular, our results indicate that promotive and prohibitive voice have a curvilinear and linear relationship, respectively, with employee performance. The antecedents of promotive and prohibitive voice have garnered attention in empirical research (e.g., Kakkar et al., 2016; Loi et al., 2014; Ward et al., 2016). However, information on their consequences is limited (Mo & Shi, 2016). The introduction of employee performance as an outcome has enriched the existing research on the unique consequences of promotive and prohibitive voice.

This study provides new insights into the relationship between promotive voice and individual performance. Mo and Shi (2016) focused on social interactions with leaders and confirmed a positively linear relationship between promotive voice and individual performance. To further develop previous work (ibid), we integrated the processes of social interaction and resource allocation, thereby revealing that a curvilinear relationship exists between promotive voice and employee performance. This finding empirically supports the argument of Rapp et al. (2013) that employees can achieve a high level of performance when they gain support from outside and have time to complete tasks. Moreover, prior studies have suggested that nonlinear models have a better data fit in organizational contexts than linear models (Giorgi, Leon-Perez, & Arenas, 2015; KaranikaMurray, Antoniou, Michaelides, & Cox, 2009). Therefore, the current study offers insights into the promotive voice–performance relationship by identifying the inverted U-shaped association between promotive voice and employee performance.

This study also has important implications for understanding the prohibitive aspect of voice. Prohibitive voice has attracted less empirical attention than promotive voice (Liang et al., 2012). The analysis of the role of prohibitive voice in employee performance has extended previous research (Mo & Shi, 2016) and advanced our understanding of the outcomes of prohibitive voice. The extant literature generally emphasizes the negative other-directed implications of prohibitive voice for recipients but disregards the effects on employees who engage in prohibitive voice (Lin & Johnson, 2015). We adopt a self-directed perspective to analyze the relationship between prohibitive voice and employee performance. By adopting the self-directed perspective, the proximal effects of the subjective experience of work on employees who exhibit prohibitive voice can be better revealed (Weiss & Rupp, 2011). The findings support the positive aspect of prohibitive voice. That is, prohibitive voice helps employees enhance their work motivation and increase their accuracy at work, thereby improving their performance. Therefore, this evidence enriches the knowledge of prohibitive voice and may facilitate an improved understanding of the psychological process that links prohibitive voice to its consequences.

Third, the present results confirm the moderating role of power distance orientation in the relationships among promotive, prohibitive voice and employee performance. The inverted U-shaped relationship between promotive voice and performance is more pronounced, whereas the positive linear relationship between prohibitive voice and performance is weaker among employees with a high (vs. low) power distance orientation. This finding demonstrates the importance of individuals’ culture in understanding the effects of voice on performance. Scholars on voice have recently shown increasing interest in the boundary conditions of the effect of voice on performance outcomes (e.g., Hung et al., 2012; Song et al., 2017). However, our understanding remains restricted due to our limited knowledge of the importance of culture. By establishing boundary conditions from a cultural perspective, the current research extends the view of the existing literature on the boundary conditions of the voice–employee performance relationship. We also respond to voice scholars’ call for increased attention to the role of culture in the voice literature (Bashshur & Oc, 2015). Furthermore, our study is consistent with previous studies (e.g., Astakhova, 2015) that observed the moderating effects of culture value on the link between one’s behavior or attitude and work outcomes.

Managerial implications

The results of this study have important managerial implications. First, our results indicate that prohibitive voice and moderate promotive voice can improve employee performance. Thus, managers should be completely aware of the role of the types of voice in employee performance. Encouraging voice, particularly one that considers the inverted U-shaped pattern that links promotive voice and performance, may be a worthwhile endeavor. Managerial attitudes, specifically trust, are deemed key to employees’ high level of engagement in voice (Gollan & Wilkinson, 2007). Managers should focus more on the quality of voice (rather than quantity) by setting standards regarding the content of employee expression. Moreover, managers should provide timely feedback to avoid repeated voice. Likewise, strengthening employees’ resource management capability is important. This notion indicates that organization should provide training programs designed to improve employees’ ability to reasonably allocate resources (e.g., time management skills). In doing so, employees may better balance resource allocation as they promote voice and accomplishing tasks. Note that the channels of voice are numerous, including direct and indirect mechanisms and the combination of both (Bryson, Willman, Gomez, & Kretschmer, 2013). Direct and indirect voice complement each other (Wilkinson et al., 2004); when combined, they have the best effect (Bashshur & Oc, 2015). Thus, we suggest that managers provide various channels for employees to voice.

Second, this study elucidates the moderating role of power distance orientation in the relationship between voice and employee performance. This research implies that managers should focus on individuals’ culture values when implementing management practices. Organizations can collate employee information related to power distance orientation through surveys and adopt a flexible and strategic approach for employees with varying levels of power distance orientation. On the one hand, important resources overseen by the leaders should be highlighted among employees with low levels of power distance orientation. On the other hand, organizations should invite employees, particularly those with high power distance orientation, to participate in open communication and send a signal of acceptance in this invitation.

Limitation and directions for future research

This study has several limitations that suggest possibilities for future research. First, the cross-sectional design adopted is an obstacle to causal inference. Future studies could identify causal associations by using instrumental variables (Angrist, Imbens, & Rubin, 1996). Moreover, the relationship between voice and employee performance is complicated and possibly changes with time. We cannot capture these changes over time and make causal inferences to disclose this complex relationship without using longitudinal data (Lam, Chen, & Schaubroeck, 2002; Rapp et al., 2013). However, for now, the best we can conclude is that the relationships we demonstrated support our arguments. Nevertheless, a longitudinal design is preferable in future research.

Second, this study only identified power distance orientation as a moderator. However, the role of power distance orientation varies with voice stages. Apart from the moderating role studied in this research, power distance orientation may also have a predictive function, particularly for prohibitive voice, because employees with high power distance orientation are characterized by obedience to authority (Lian et al., 2012) and may not display behaviors that contradict this characteristic. On the other hand, Gross and John (2003) suggested that a suppression regulatory strategy presents oneself in such a manner that deviates from one’s inner belief to protect what they care about and avoid what they fear losing. Although prohibitive voice differs from one’s beliefs about power, individuals with high power distance orientation may still take it via a suppression regulatory strategy. The reason is that these individuals care about organizational development and fear the factors that harm their organizations. Future research is suggested to obtain the whole picture of the role of power distance orientation in the voice process. We argue that the influence of power distance orientation on prohibitive voice in the present study is not a major issue because the relationship between power distance orientation and prohibitive voice is nonsignificant (β = −.09, ns; see Table 2). Moreover, the question of whether power distance has a similar moderating effect at the country level remains unclear. Thus, we recommend that scholars further analyze this question in future cross-cultural research.

Third, in line with previous studies (e.g., Qin et al., 2014), promotive and prohibitive voice were rated using scales. We may acquire an improved understanding of leaders’ reactions by analyzing the proportion of each type of voice. Hence, we recommend that future research involve a 2 × 2 study of a high/low proportion of promotive voice versus a high/low proportion of prohibitive voice. Apart from the promotive–prohibitive framework, other types of voice (e.g., Maynes & Podsakoff, 2014) deserve attention in future studies. In addition, we focused only on direct voice and did not pay attention to indirect voice (e.g., representative voice; Bryson et al., 2013). Thus, it is suggested that future research focus on diverse channels or combinations of the channels of voice.

Fourth, in line with prior studies (e.g., Lam et al., 2002; LePine, Zhang, Crawford, & Rich, 2016; Ozer, 2011), we adopted subjective measures and asked the direct leaders to rate employee performance. However, subjective measures are vulnerable to systematic bias (Bommer, Johnson, Rich, Podsakoff, & MacKenzie, 1995) and random error (Bollen & Paxton, 1998). For example, leaders’ evaluations may be bounded by their observational opportunities and cognitive abilities (Feldman, 1981). Although objective measures are free from systematic bias and random error, the applications of objective measures are excessively narrow. That is, such measures can be applied only for limited types of jobs (Bommer et al., 1995; Murphy, 2008). Thus, future research could improve on the present study by adopting 360-degree evaluations from leaders, peers, and, occasionally, clients (Murphy, 2008) or a combination of objective and subjective data.

Fifth, we controlled for only employee demographic variables and ignored the potential influence of human resource practices (e.g., reward systems and training) on employee performance (Chang & Chen, 2011; Cherrington, Reitz, & Scott, 1971), which should be controlled for in future research. Sixth, we examined the direct relationships between only two types of voice and employee performance, without taking internal mechanisms into account. Future research is expected to explore the potential mediating mechanisms, such as resource consumption and leaders’ attitudes. Finally, the limitation regarding the sample (i.e., using a convenience sample from China) limits the generalizability of our findings. We expect future research to adopt the random sampling method and the result will duplicate findings of the current study.