Introduction

Employee voice is defined as the discretionary communication of ideas, suggestions, and concerns about work-related issues and problems to individuals who might be able to take appropriate action (Morrison, 2014). Over the past few decades, voice has attracted much attention and is an important factor of organizational development (Bashshur & Oc, 2015; Chamberlin et al., 2017; Weiss & Morrison, 2019). Given its importance, scholars have exerted considerable efforts in identifying its motivational antecedents, such as felt obligation for constructive change (Liang et al., 2012), organizational identification (Tangirala & Ramanujam, 2008a), and psychological attachment (Burris et al., 2008). Although these antecedents can explain why employees voice, they cannot reveal and differentiate employees' underlying motives. Essentially, employees can speak up and make suggestions because they genuinely want the organization to improve or they want to be rewarded for an extra-role behavior (Klaas et al., 2012).

Predominant literature has usually considered voice a type of other-serving behavior, motivated by an intent to bring about change that is beneficial for the organization (Morrison, 2014). However, few studies have empirically examined voice as the result of a self-serving motive, such that it helps employees to promote positive self-image (Klaas et al., 2012; Morrison, 2014). Although voice challenges the status quo, it simultaneously allows employees to display their competences and skills (Burris, 2012; Grant, 2013; Yun et al., 2007), thereby enhancing their status (Weiss & Morrison, 2019). A lack of consideration of other- and self-serving motives in voice enactment and the contextual factors that shape whether these different motives give rise to voice behaviors are important theoretical issues worthy of empirical investigations. Practically, by understanding that employee voice behavior can be driven by diverse motives, organizations can seek to tap into those different motives when cultivating greater voice. For instance, organizations not only can emphasize the importance of being a good citizen, but also suggest that there are real personal benefits (e.g., status) to enacting voice.

Voice reflects a deliberate decision process wherein the enactor weighs the potential benefits and costs of the action (Burris, 2012; Detert & Bruno, 2017; Huang & Paterson, 2017; Morrison, 2011). This is because suggesting changes to existing practices and procedures can alter the status quo and move relevant others out of their comfort zones, which may elicit negative responses from them (Burris, 2012; Morrison, 2011).Footnote 1 Employees who are driven by other-serving motives are less likely to be calculative of the potential interpersonal risks involved with voice enactment, which likely differs from those who are driven by self-serving motives (Donia et al., 2016). As such, observing employee inclinations to “read the wind” can shed light on their motives underlying voice, as “reading the wind” can better inform employees about the potential risks associated with voice. Tangirala and Ramanujam (2008b) echoed this proposition by claiming that although voice is influenced by individual-level factors, the social context is likely to have a significant impact on “whether or not this motivation finds expression as behavior” (p. 44). “Good soldiers” are more likely to stand out and voice when others remain silent because they are more concerned with organizational development and less concerned with their personal utilities, whereas “good actors” tend to avoid the potential risks and speak up when others are also speaking up. Given that employee perceptions of voice levels in the work context can provide direct contextual cues about the favorability of voice, it is a critical boundary condition that can separate “good soldiers” from “good actors.”

Integrating Rioux and Penner’s (2001) citizenship motives framework with the contingency perspective (e.g., Mischel, 1977), we investigate the influences of motives on voice and an important boundary condition. Specifically, we examine the impacts of two relevant motives, organizational concern (OC) and impression management (IM) motives, on promotive (i.e., “expressions of ways to improve existing work practices and procedures to benefit organizations”; Liang et al., 2012, p. 71) and prohibitive voice (i.e., “expressions of individuals’ concern about existing or impending practices, incidents, or behaviors that may harm their organization”; Liang et al., 2012, p. 72). OC motives center on the desire to help the organization; IM motives focus on the desire to be viewed positively by others (Rioux & Penner, 2001). Therefore, OC motives are considered other-serving motives, whereas IM motives illustrate self-serving motives. Furthermore, as employee behaviors are often influenced by the extent to which similar behaviors are demonstrated by others in the work context (Buunk & Gibbons, 2007; Festinger, 1954), we investigate perceived voice level in the work context as a critical contingent influence.Footnote 2

This study makes three primary contributions. First, we advance the voice literature by identifying the effects of motives on voice, of which the knowledge about is rather limited (for reviews, see Morrison, 2011, 2014). We not only expand the antecedent side of the voice literature (Chamberlin et al., 2017), but also respond to calls for additional work on discovering the antecedents of promotive and prohibitive voice (e.g., Liang et al., 2012; Morrison, 2014). Second, we extend the voice literature by specifically linking IM motives to voice. Most studies have associated voice with other-serving (e.g., OC) motives (e.g., Grant & Mayer, 2009; Kim et al., 2013; Morrison, 2011), but have comparatively neglected how voice can also be the result of self-serving motives (Van Dyne et al., 2003). We directly address this issue by examining both OC and IM motives as key determinants of voice, enriching understandings of why employees speak up in the workplace. Third, we enrich the voice literature by exploring the boundary condition of the links between OC/IM motives and voice. So far, the picture of when individual motives are more or less related to voice (Grant & Mayer, 2009) is underexplored. We demonstrate how employees with different motives respond to perceptions of voice levels in the work context. An illustration of the present research model is presented in Fig. 1.

Fig. 1
figure 1

The hypothesized model

Full size image

Theory and Hypotheses

Overview of Motives and Voice

Unlike an in-role behavior, an extra-role behavior is a work-related act that is not formally required by job descriptions (Van Dyne & LePine, 1998). Due to its voluntariness, there exist various reasons to which individuals engage in extra-role behavior. Rioux and Penner (2001) suggested that there are three distinct motives underlying extra-role behaviors: people want to maintain their image and obtain rewards (i.e., IM motives); people want to be helpful and build positive relationships with others (i.e., prosocial values or PV motives); and people want the organization to do well (i.e., OC motives). It is worth noting that PV and OC motives are related to intrinsic motivation because they are unitary motivations in which the intrinsic values are embedded (Grant, 2008). This means that people with PV and OC motives engage in extra-role behavior due to feelings of enjoyment, differing from prosocial motivation in that it does not assume an intrinsic value in its outcome effort (Grant, 2008).

Although all three motives have been suggested to facilitate citizenship behavior, they exert different influences on the specific dimension of citizenship behavior (Rioux & Penner, 2001). Since PV motives concern about interpersonal relationships, they are more related to the interpersonal dimension of citizenship behavior, such as helping and altruism (Kim et al., 2013). OC motives directly focus on organizational welfare, and therefore have been found to stimulate citizenship behaviors toward the organization (Bourdage et al., 2012). Whereas PV and OC motives emphasize certain dimension of citizenship behavior, IM motives can lead to both dimensions of citizenship behavior, as IM can be used to gain favorable evaluations from both a particular person and the larger collective (Bolino et al., 2008). As such, IM motives can lead to citizenship behavior toward particular individuals or the organization.

Voice is a type of citizenship behavior that involves expression of ideas or concerns about work-related issues (Morrison, 2011; Van Dyne et al., 2003). Since it is often made toward the organization about potential improvements in its existing work practices and procedures (Van Dyne & LePine, 1998), it is generally considered a citizenship behavior toward the organization (Chiaburu et al., 2013). Therefore, we focus on OC and IM motives and investigate their relationships with voice (i.e., promotive voice and prohibitive voice). The reasons are twofold. First, although voice is typically assumed as a form of other-serving behavior to improve the organization, it can also elicit positive influences on enactors’ image (Burris, 2012; Grant, 2013; Yun et al., 2007). Thus, voice is likely to be influenced by both other-serving (e.g., OC) and self-serving (e.g., IM) motives. Second, compared to an affiliative citizenship behavior (e.g., helping a coworker) which mainly involves maintaining the status quo and building positive relationships with others, voice is a more challenging form of citizenship behavior because it challenges the status quo and may harm existing relationships with others (Grant, 2013; Van Dyne et al., 1995). Therefore, voice is theoretically less relevant to PV motives. Consequently, we focus on the influences of OC and IM motives on voice.

OC Motives and Promotive and Prohibitive Voice

Voice is essential to the improvement, development, and even survival of organizations (Liang et al., 2012; Morrison, 2011, 2014; Van Dyne et al., 2003). In accordance with Ajzen and Fishbein’s (1977) principle of compatibility, OC motives and voice share a common goal, which is to seek ways to help organizations progress. Hence, they are expected to be positively related to each other. Studies have associated voice with a motivation to voluntarily help the organization and others. For example, Rioux and Penner (2001) suggested that showing concerns for the organization promotes employees to engage in citizenship behaviors toward the organization. Morrison (2011, 2014) argued that voice results from employees having a desire to promote constructive changes for the organization. Mowbray et al. (2015) emphasized that voice is a form of prosocial behavior, which is less self-focused and more other-focused. As such, employees with strong OC motives place great values on organizational functioning (Bolino et al., 2012; Grant, 2008; Ilies et al., 2006) and have high expectations for organizational development (Kim et al., 2013; Rioux & Penner, 2001). They are likely to generate ideas and suggestions as well as detect errors and pitfalls (Meglino & Korsgaard, 2004). In addition, due to their other-serving nature, they are less likely to regard the potential risks associated with voice enactment. In short, employees with OC motives are likely to propose constructive suggestions (i.e., promotive voice) and express concerns (i.e., prohibitive voice). These arguments lead to the first hypothesis:

  • Hypothesis 1: OC motives are positively associated with (a) promotive and (b) prohibitive voice.

IM Motives and Promotive and Prohibitive Voice

Existing literature has suggested that voice can be potentially motivated by self-serving motives. For instance, Bolino et al. (2006) found that organizational citizenship behavior is influenced by an employee’s inherent need to make a good impression. Klaas et al. (2012) argued that “a self-promotional focus may also be relevant” (p. 329) to voice. Morrison (2014) furthered echoed their viewpoints by arguing that “voice may be shaped, at least in part, by the desire to achieve positive self-relevant outcomes” (p. 184). Therefore, in addition to its potential benefits for the organization, voice can bring about personal benefits to the enactor, such as reward and recognition (Burris, 2012; Chamberlin et al., 2017).

IM motives are rooted in the pursuit of a positive self-image (Kim et al., 2013; Rioux & Penner, 2001). Employees with strong IM motives are eager to be identified and acknowledged by others in an organizational setting (Bolino et al., 2008). By engaging in voice behavior, employees can showcase their knowledge, skills, and abilities to others (Detert & Bruno, 2017; McClean et al., 2018) as well as advertise their altruistic and communal orientations (McClean et al., 2018; Weiss & Morrison, 2019). Voice is a viable strategic behavior to inform others about their prosocial mentality, therefore building a favorable image (Fuller et al., 2007). These arguments lead to the next hypothesis:

  • Hypothesis 2: IM motives are positively associated with (a) promotive and (b) prohibitive voice.

The Differential Moderating Effects of Perceived Voice Level in the Work Context

Voice is improvement-oriented, meaning that it is spoken with an intent to resolve particular issues by offering potential applicable suggestions. Hence, enactors are not only required to have sufficient and specific knowledge about the issue but also courage (Detert & Bruno, 2017). For example, making suggestions about product designs can be rejected by certain individuals due to the extra workload elicited on them. As voice challenges the status quo and may engender negative responses from others (Burris, 2012; Morrison, 2011), employees tend to consider the potential benefits and costs associated with enactment of voice (Burris, 2012; Huang & Paterson, 2017).

Studies have suggested that employees tend to be sensitive in detecting situational cues from the work environment (Cai et al., 2019; Lam et al., 2018) and are inclined to “read the wind” and adjust their voice behavior (Morrison, 2014; Morrison et al., 2011). This is because the work environment can inform employees about the status quo as well as the potential consequences associated with voice (Hussain et al., 2019). We suggest that an employee’s perceived voice level in the work context,Footnote 3 refering to the extent to which one believes that others within one’s work context engage in voice behavior, can offer an important contextual cue for an employee’s voice decisions. However, due to the distinct underlying motives, perceived voice level in the work context is likely to have differential moderating influences on the relationships between OC/IM motives and voice, respectively.

When others remain silent (i.e., perceived low levels of voice), although speaking up requires greater courage and involves more potential risks, it also brings more value to the organization due to its rarity and uniqueness (Hussain et al., 2019). Since employees with OC motives have a strong desire to see the organization develop, they are likely to disregard their potential risks in favor of organizational development (Takeuchi et al., 2015). Moreover, as employees’ constructive suggestions are vital to organizational success and progression, a lack of voice in the work context suggests that the organization is usually stagnant and may harm the survival of the organization (Liang et al., 2019; Morrison, 2014). This situation is especially likely to facilitate employees who have strong OC motives to be concerned with organizational development and eager to bring about constructive change (Lam et al., 2018). As such, they are more likely to offer both constructive suggestions (i.e., promotive voice) and voice concerns (i.e., prohibitive voice) when perceiving lower levels of voice in their work contexts.

On the other hand, employees with IM motives care much about their personal benefits and others' views of them (Halbesleben et al., 2010). They are likely to account for the potential risks associated with voice enactment and be more selective in the context in which they voice. For instance, they may avoid making bold suggestions and carefully choose the recipients of their voice (Giacalone & Rosenfeld, 2013; Morrison, 2011). In addition, they may be more mindful about their tone of voice and body language when making suggestions (Giacalone & Rosenfeld, 2013; Hosman & Siltanen, 2011). As such, we argue that the extent to which employees with IM motives voice is contingent on them “reading the wind.”

When perceiving others as active in voice behavior (i.e., perceived high levels of voice), employees with IM motives are likely to conform to others and voice as well. The reason is twofold. First, conforming to others sends a message that one cares about group values (Hewlin, 2003), helping gaining social approval from others (Cialdini & Goldstein, 2004). Indeed, the IM literature suggests that speaking in ways consistent with others engenders a favorable image (Bolino et al., 2008). Second, speaking up when others are also speaking up minimizes the negative consequences that one might bear. Essentially, one can enhance one’s personal image without having to bear much of a risk. Consequently, we posit the following hypotheses:

  • Hypothesis 3: Perceived voice level in the work context moderates the positive relationships between OC motives and voice (i.e., promotive and prohibitive voice), such that the associations between OC motives and voice are stronger when employees perceive lower levels of voice in their work contexts.

  • Hypothesis 4: Perceived voice level in the work context moderates the positive relationships between IM motives and voice (i.e., promotive and prohibitive voice), such that the associations between IM motives and voice are stronger when employees perceive higher levels of voice in their work contexts.

Method

Samples and Procedures

We designed and conducted a multi-source survey by using a snowball sampling technique to collect data (Heckathorn, 1997). In order to reduce common method variance (Podsakoff et al., 2003; Podsakoff et al., 2012), we invited supervisor-subordinate dyads to participate in the study. Following previous studies that have adopted a similar technique (e.g., Bayl-Smith & Griffin, 2015; Ezeofor & Lent, 2014; Meyer et al., 2013), we first sent the respective online survey links (i.e., supervisor or subordinate survey) to our contacts (i.e., supervisor or subordinate) working in China via an online survey platform (http://www.wjx.cn/). After completing the surveys, we asked them to invite their counterparts to participate in the study. Specifically, if the contacts were subordinates, they were asked to invite their direct supervisors to complete the supervisor survey. If the contacts were supervisors, they were asked to randomly evaluate one of their subordinates and invite that subordinate to participate in the subordinate survey. In addition, we asked the contacts to invite other appropriate full-time working personnel (supervisors or subordinates) to participate in the study following the abovementioned procedures. We stated that participation in the study is voluntary and explained the research purpose and procedures in the beginning of the survey. Each participant was only allowed to submit one completed questionnaire.

Subordinates were asked to self-report their OC and IM motives and report on their perceptions of voice levels in their work contexts. Supervisors evaluated their sampled subordinate’s promotive and prohibitive voice. The participants provided their demographic information. In three weeks, we received 258 and 210 responses from subordinates and supervisors, respectively. After matching the surveys with a unique identification code, the final valid sample consisted of 140 pairings of supervisor-subordinate dyads.

For the subordinate sample, 44.29% were male; 72.14% had received a bachelor’s degree; and 68.57% were frontline employees. Their average age was 2.67 (1 = 18–25 years old, 2 = 26–30 years old, 3 = 31–40 years old, 4 = 41–50 years old, 5 = 51–60 years old, 6 = 61 years old or above; standard deviation [SD] = 1.06). For the supervisor sample, 65.71% were male; 84.29% had received a bachelor’s degree; and 59.20% were middle-level managers. Their average age was 3.46 (1 = 18–25 years old, 2 = 26–30 years old, 3 = 31–40 years old, 4 = 41–50 years old, 5 = 51–60 years old, 6 = 61 years old or above; SD = 0.86). Among the supervisor-subordinate dyads, 30.00% were from government sectors; 24.29% were from public institutions; 27.14% were from state-owned enterprises; 10.00% were from private enterprises; 6.43% were from foreign enterprises; and 2.14% were from none of the above.

Measures

The survey items were presented in Chinese, following standard translation and back-translation procedures (Brislin, 1986). Unless otherwise stated, all measures were rated on a seven-point Likert scale (1 = strongly disagree to 7 = strongly agree).

OC and IM motives

We measured OC and IM motives using the Citizenship Motives Scale developed by Rioux and Penner (2001). Adapting to our research, we asked participants to indicate why they actively reported related concerns, ideas, problems, or suggestions to their supervisors. OC motives were assessed with ten items, such as “Because I care what happens to the company” and “Because I am committed to the company” (Cronbach’s alpha [α] = 0.90). IM motives were evaluated with ten items, such as “To avoid looking lazy” and “To look better than my coworkers” (α = 0.93).

Employee promotive and prohibitive voice

We used the ten-item scale developed by Liang et al. (2012) to measure promotive and prohibitive voice (five items each). A sample item for promotive voice was, “This employee makes constructive suggestions to improve the unit’s operation” (α = 0.95). A sample item for prohibitive voice was, “This employee speaks up honestly with problems that might cause serious loss to the work unit, even when/though dissenting opinions exist” (α = 0.91).

Although promotive and prohibitive voice have been validated as two independent constructs (e.g., Kakkar et al., 2016; Liang et al., 2012; Wei et al., 2015), we examined their distinctiveness. Specifically, we used a total of 210 responses from supervisors to perform confirmatory factor analysis (CFA). The results showed that the two-factor measurement model (i.e., promotive and prohibitive voice were treated as separate constructs) had better fit indices (χ2(34) = 109.04, p < 0.001; SRMR = 0.04, TLI = 0.95, CFI = 0.96, RMSEA = 0.10; Δχ2(1) = 145.92, p < 0.001) than those of the one-factor measurement model (i.e., promotive and prohibitive voice were treated as one construct; χ2(35) = 254.96, p < 0.001; SRMR = 0.06, TLI = 0.85, CFI = 0.88, RMSEA = 0.17). Thus, promotive and prohibitive voice were empirically distinct.

Perceived voice level in the work context

Following previous studies (e.g., Morrison et al., 2011), we used a referent shift technique (Chan, 1998) and asked participants to report on their perceptions of voice levels in their work contexts. Items were adapted from Van Dyne and LePine’s (1998) six-item scale. A sample item was, “In my work context, employees speak up with ideas for new projects or changes in procedures” (α = 0.91).

Control variables

We included employee position (1 = frontline employee, 2 = frontline manager, 3 = middle-level manager, 4 = high-level manager) as a control variable due to both theoretical and empirical concerns. Theoretically, employee position reflects one’s power, resources controlled, status, and span of influence (Milliken et al., 2003; Morrison, 2011), which are important determinants of one’s confidence in speaking up and others’ attitude toward and acceptance of one’s voice behavior (McClean et al., 2018; Morrison, 2014). Empirically, studies have shown that employees in higher positions feel a greater responsibility to speak up and make suggestions (Fuller et al., 2006) and are more likely to voice opinions (Chamberlin et al., 2017).Footnote 4

Results

Preliminary Analyses

We used Lisrel 8.8 to conduct CFA to ensure the discriminant validity of the main variables. As shown in Table 1, the hypothesized five-factor model demonstrates appropriate fit (χ2(570) = 887.27, p < 0.001; SRMR = 0.08, TLI = 0.90, CFI = 0.91, RMSEA = 0.06), resembling a significant improvement in chi-square value (χ2) over all the alternative models. Thus, construct distinctiveness of the main variables is established.

Table 1 Results of Confirmatory Factor Analysis
Full size table

To test common method variance, we performed the Harman’s single-factor test (Podsakoff et al., 2003). The results of exploratory factor analysis (with the unrotated factor solution) show four factors. The first factor accounts for 28.33% of the variance, which is lower than the threshold of 50% (Hair et al., 1998). In addition, as shown in Table 1, the hypothesized five-factor model has superior model fit indices than the alternative one-factor model (χ2(580) = 3927.54, p < 0.001; SRMR = 0.22, TLI = 0.51, CFI = 0.55, RMSEA = 0.20). Thus, common method variance was not a substantial issue in our findings (Podsakoff et al., 2003).

Hypothesis Tests

Table 2 presents the means, SDs, correlations, and reliabilities of study variables. To test the hypotheses, we used SPSS 25.0 to conduct regression analysis. The results are summarized in Table 3.

Table 2 Means, SDs, Correlations, and Reliabilities of Study Variables
Full size table
Table 3 Regression Results
Full size table

Tests of main effects

Hypothesis 1 proposes that OC motives are positively related to promotive (H1a) and prohibitive voice (H1b). The results displayed in Table 3 indicate that OC motives are positively associated with promotive voice (B = 0.36, SE = 0.12, p < 0.01; see Model 2) and prohibitive voice (B = 0.33, SE = 0.11, p < 0.01; see Model 6). Therefore, Hypotheses 1a and 1b receive support.

Similarly, Hypothesis 2 argues for the positive links between IM motives and promotive (H2a) and prohibitive voice (H2b). The results presented in Table 3 show that IM motives are indeed positively related to promotive voice (B = 0.16, SE = 0.08, p < 0.05; see Model 2) and prohibitive voice (B = 0.20, SE = 0.07, p < 0.01; see Model 6). Thus, Hypotheses 2a and 2b are both supported.

Tests of moderation effects

Hypothesis 3 proposes that perceived voice level in the work context moderates the positive relationships between OC motives and voice (i.e., promotive and prohibitive voice), such that the relationships are more salient when employees perceive lower levels of voice in their work contexts. The results displayed in Table 3 reveal that the interaction term of OC motives and perceived voice level in the work context is negatively related to employee promotive voice (B = -0.25, SE = 0.08, p < 0.01; see Model 4). Figure 2 depicts the interaction plot, showing that when perceived voice level in the work context is “low” (i.e., one SD below the mean), OC motives are positively related to promotive voice (simple slope = 0.54, SE = 0.15, t = 3.49, p < 0.001). When perceived voice level in the work context is “high” (i.e., one SD above the mean), OC motives are not significantly related to promotive voice (simple slope = 0.07, SE = 0.18, t = 0.37, p > 0.05). In addition, the results presented in Table 3 show that the interaction term of OC motives and perceived voice level in the work context is negatively related to employee prohibitive voice (B =  − 0.21, SE = 0.07, p < 0.01; see Model 8). The interaction plot in Fig. 3 presents a similar pattern in that when perceived voice level in the work context is “low,” OC motives are positively related to prohibitive voice (simple slope = 0.51, SE = 0.14, t = 3.54, p < 0.001). When perceived voice level in the work context is “high,” OC motives are not significantly related to prohibitive voice (simple slope = 0.11, SE = 0.17, t = 0.67, p > 0.05). These results provide support for Hypothesis 3.

Fig. 2
figure 2

The moderating effect of perceived voice level in the work context on the relationship between OC motives and promotive voice

Full size image
Fig. 3
figure 3

The moderating effect of perceived voice level in the work context on the relationship between OC motives and prohibitive voice

Full size image

Hypothesis 4 argues that perceived voice level in the work context moderates the positive relationships between IM motives and voice (i.e., promotive and prohibitive voice), such that the relationships are stronger when employees perceive higher levels of voice in their work contexts. The results presented in Table 3 show that the interaction term of IM motives and perceived voice level in the work context is positively linked to promotive voice (B = 0.18, SE = 0.09, p < 0.05; see Model 4). Figure 4 displays the interaction plot, indicating that when perceived voice level in the work context is “high,” IM motives are positively related to promotive voice (simple slope = 0.23, SE = 0.11, t = 2.07, p < 0.05). When perceived voice level in the work context is “low,” IM motives are not significantly related to promotive voice (simple slope =  − 0.11, SE = 0.13, t =  − 0.89, p > 0.05). In addition, the results shown in Table 3 indicate that the interaction term of IM motives and perceived voice level in the work context is positively related to prohibitive voice (B = 0.18, SE = 0.08, p < 0.05; see Model 8). The interaction plot presented in Fig. 5 suggests that when perceived voice level in the work context is “high,” IM motives are positively related to prohibitive voice (simple slope = 0.29, SE = 0.10, t = 2.94, p < 0.01). When perceived voice level in the work context is “low,” IM motives are not significantly related to prohibitive voice (simple slope =  − 0.05, SE = 0.12, t =  − 0.41, p > 0.05). These results provide support for Hypothesis 4.

Fig. 4
figure 4

The moderating effect of perceived voice level in the work context on the relationship between IM motives and promotive voice

Full size image
Fig. 5
figure 5

The moderating effect of perceived voice level in the work context on the relationship between IM motives and prohibitive voice

Full size image

Supplemental Analysis

Since employees in higher positions might feel more in control than those in lower positions (Milliken et al., 2003; Morrison, 2011), employee position may moderate the relationships between motives and voice. As such, we conducted supplemental analysis to explore the moderating role of employee position. The results indicate that the interaction term of OC motives and position is neither significantly related to promotive voice (B =  − 0.16, SE = 0.16, p > 0.05) nor prohibitive voice (B =  − 0.01, SE = 0.15, p > 0.05), suggesting that the effect of OC motives on voice is not contingent on employee position. On the other hand, the interaction term of IM motives and position significantly and negatively predicts promotive voice (B =  − 0.27, SE = 0.10, p < 0.05) and prohibitive voice (B =  − 0.20, SE = 0.09, p < 0.05), indicating that position weakens the relationship between IM motives and voice. An explanation to this interesting finding is that employees in higher positions have a lesser need to manage their impressions through means of speaking up than those in lower positions.

Discussion

Integrating the framework of citizenship motives (Rioux & Penner, 2001) with the voice literature (Liang et al., 2012; Morrison, 2011, 2014), we developed a contingency model to investigate the effects of OC and IM motives on promotive and prohibitive voice as well as the differential moderating effects of perceived voice level in the work context on these relationships. As hypothesized, we found that both OC and IM motives were positively related to promotive and prohibitive voice. Moreover, we showed that the relationships between OC motives and promotive and prohibitive voice were stronger when employees perceived lower levels of voice in their work contexts and that the associations between IM motives and promotive and prohibitive voice were stronger when employees perceived higher levels of voice in their work contexts. Overall, this study provides implications for both theory and practice, which are to be discussed in the following sections.

Theoretical Implications

This study makes three theoretical contributions. First, we contribute to the voice literature by identifying the impacts of motives on voice. Whereas existing studies have mostly investigated individual personalities (e.g., LePine & Van Dyne, 2001), attitudes (e.g., Liang et al., 2012; Tangirala & Ramanujam, 2008b), and contextual influences (Detert & Burris, 2007; Morrison et al., 2011; Walumbwa et al., 2012) as predictors of voice, only a few have examined individual motives as antecedents of voice (for reviews, see Morrison, 2011, 2014). As suggested by Rioux and Penner (2001) that there is a need to further scrutinize motives underlying voice, we explore the influences of OC and IM motives on voice (Klaas et al., 2012; Morrison, 2011, 2014), broadening understandings of antecedents of voice.

Second, we extend the voice literature by paying particular attention to the antecedent role of IM motives. Research to date has generally assumed that voice, including both promotive and prohibitive voice, is motivated by “the desire to help the organization or work unit” (Morrison, 2011, p. 381). However, scholars have paid scant attention to other potential underlying motives (Morrison, 2014), resulting in an incomplete understanding of why employees speak up in the workplace. We address this issue and suggest that employees engage in promotive and prohibitive voice for reasons of “doing good” (i.e., OC motives) and “looking good” (i.e., IM motives). Our findings suggest that voice can be driven by self-serving motives, complementing existing studies that have mostly associated voice with other-serving motives (e.g., Grant & Mayer, 2009; Kim et al., 2013; Morrison, 2011).

Third, we enrich the voice literature by exploring the differential moderating effects of perceived voice level in the work context on the links between OC/IM motives and voice. Based on the nature of voice, the environment in which employees are embedded plays an extremely crucial role in employee decisions to speak up or remain silent (Kakkar et al., 2016; LePine & Van Dyne, 1998; Tangirala & Ramanujam, 2008b). Surprisingly, extant literature has paid little attention to examining this boundary condition (Morrison et al., 2011). As such, we have insufficient knowledge about when inherent motives impact employee voice (Grant & Mayer, 2009). We tap into this issue and develop a contingency model to investigate the moderating role of perceived voice level in the work context. We further delineate its moderating influence on the effects of OC and IM motives on voice. Specifically, we found that employees driven by OC motives were more likely to speak up when they perceived little voice from others in the work context. This finding resonates with Hussain et al.’s (2019) work, which suggests that voice may be subject to a bystander effect. Only those who have strong courage or other-serving motives tend to emerge and voice. In contrast, employees with IM motives tend to conform to others and speak up when others in the work context are also speaking up. These findings add to the voice literature by further clarifying how different motives can lead to different levels of voice in different contexts.

Practical Implications

This study offers a few implications for practice. Managers who are often the recipients of voice should be aware of that employees may engage in voice due to both other-serving and self-serving concerns. Managers should tap into these different motives when cultivating more voice. For example, managers should emphasize that personal interests are most likely to be preserved when the organization is performing well. Such effort can engender OC motives in which employees are likely to prioritize organizational values and interests and show commitment to the organization (Rioux & Penner, 2001). Furthermore, managers should publicly recognize and applaud employee voice. As such, employees with IM motives are likely to be motivated to speak up.

Although voice should be welcomed regardless of employees’ underlying motives, we do encourage organizations to differentiate “good soldiers” from “good actors” (Bolino, 1999). Organizations in the long-run are almost likely going to benefit more from having genuine employees who have an other-serving mentality than self-serving individuals who tend to have a calculative mindset of maximizing their own interests. This is because “good soldiers” seek to help the organization improve in all aspects, whereas “good actors” are selective and only take actions when they believe they are likely to be rewarded.

Limitations and Directions for Future Research

This study has some limitations that should be addressed by future studies. First, due to the cross-sectional data, we were unable to draw definitive causal conclusions on the relationships between motives and voice. To alleviate this concern, we advise future studies to adopt a longitudinal study design or utilize experiments to further validate our findings. Another recommendation for future studies is to consider the potential influence of cultural background. According to Takeuchi et al. (2015), cultural background may play a significant role in determining employee voice. For instance, Chinese people tend to believe that “the lead bird is most likely to be shot." Therefore, they may naturally be more likely to “read the wind” and be less inclined to voice opinions first. As this study was conducted in a Chinese work setting, we suggest future studies to draw samples from different cultural settings to gain wider knowledge about the generalizability of our findings.

We acknowledge the fact that the snowball sampling method has some shortcomings. For example, it cannot assess the response rate and may generate a biased sample with self-selected participants. In addition, it largely prevented us from collecting and obtaining group samples. Hence, all of the main constructs were operationalized at the individual level. To eliminate the potential contaminations of our findings, we sought to enhance the data quality by emphasizing the data collection process (e.g., clearly stating the procedures at the beginning of the surveys and used unique identification codes to match the surveys). Despite these shortcomings, the snowball sampling technique is widely used in organizational research (e.g., Bayl-Smith & Griffin, 2015; Ezeofor & Lent, 2014; Haar et al., 2014; Zacher et al., 2015) due to its convenience and ability to obtain diverse samples, which can enhance the generalizability of study results (Kausel et al., 2016). Nevertheless, we encourage future studies to use probability sampling method to collect data and compare findings to those of ours.

Furthermore, since perceived voice level is a perceptual measure, it depicts individual, idiosyncratic beliefs and differs from group-level voice constructs, such as group voice climate. Group climates illustrate group beliefs, often justified by empirical evidence of sharedness (e.g., Kao et al., 2021; Morrison et al., 2011).Footnote 5 Indeed, group voice climate has been measured by aggregating individual responses of voice (e.g., Frazier & Bowler, 2015; Morrison et al., 2011). As such, group voice climate may offer a more objective assessment of the actual voice level in a group. In addition, future studies can investigate the roles of other relevant contextual factors, such as psychological safety. As psychological safety is derived from calculating the potential risks of showing one’s true self (Kahn, 1990), employees with OC motives are less likely to be affected by feelings of psychological safety when deciding to voice due to their greater desire to genuinely help the organization progress. In contrast, those with IM motives are more likely to be affected by feelings of psychological safety because they are less tentative to put themselves at risk.

Finally, both OC and IM motives were treated as distinct constructs,suggesting that individuals can inherently possess both motives. The extent to which these motives are activated depend on contextual influences. This notion is supported by Rioux and Penner’s (2001) work, which suggests that the same behavior can serve multiple motives. In addition, due to the nature of the voice measurement, we do not assume a difference in voice quality between employees of different motives. Rather, voice quality can be reflected in whether others accept the suggestions and these accepted suggestions do indeed bring about effective changes. Hence, future research is encouraged to examine voice endorsement (i.e., implementation of voice by others; King et al., 2019) and its subsequent influences on organizational functioning with respect to the citizenship motives framework.

Conclusion

Drawing upon the citizenship motives framework and voice literature, we examined the effects of OC and IM motives on promotive and prohibitive voice. Moreover, adopting a contingency perspective, we investigated the differential moderating influences of perceived voice level in the work context on the associations between motives and voice. This study offers insights and implications for both theory and practice. We hope that this study can inspire other scholars to make further contributions to the voice literature.