Abstract
This study advances voice research by offering a social-relational view of the drivers of voice, a theoretical approach to voice that is seldom considered within the current paradigm largely focused on personality traits, job conditions, and organizational characteristics. One overlooked yet important social-relational antecedent of voice is received respect. Our core premise is that when employees believe they are respected by coworkers, they experience psychological changes to their control beliefs (representing “can-do” proactive motivation) and positive mood (representing “energized-to” proactive motivation), which then motivate voice. We further consider another social-relational variable—perspective taking—as a predictor of received respect and therefore as an indirect predictor of voice. Through a multimethod, multisample research program comprising four studies (two experiments involving more than 400 subjects in total, a sample involving more than 700 matched employee-coworker and subordinate-supervisor dyads, and a 9-week within-person field investigation of more than 400 university alumni), we provide evidence to support the proposed model. That is, received respect was associated with employees’ voice through control beliefs and positive mood, and perspective taking was a prominent predictor of received respect.
Contemporary organizations rely on employee voice to improve, adapt, and innovate (Bashshur & Oc, 2015; Grant, 2013; Li, Liao, Tangirala, & Firth, 2017). Constructive voice (or simply “voice”) refers to change-oriented communication intended to advance an organization’s interests (LePine & Van Dyne, 1998). Much evidence has shown the value of voice for key organizational outcomes (Bashshur & Oc, 2015; Morrison, 2014), which in turn has led scholars to seek to understand what drives voice. Personality traits (e.g., LePine & Van Dyne, 2001; Tangirala, Kamdar, Venkataramani, & Parke, 2013), job conditions (e.g., Fuller, Marler, & Hester, 2006; Schmitt, Den Hartog, & Belschak, 2016), and organizational characteristics (e.g., Conway, Fu, Monks, Alfes, & Bailey, 2016; Knoll & Redman, 2016) have all been identified as critical antecedents of voice.
However, one important gap remains in the current voice research. Namely, the social-relational antecedents of voice have been underrepresented. Chamberlin, Newton, and LePine (2017) reviewed the voice literature and identified 32 main predictors; only 8 were social-relational in nature. Six of these focused on leadership (transformational leadership, leader-member exchange, ethical leadership, leader openness, trust in leader, low abusive supervision) and 2 concentrated on peers (social support, group identification), with corrected effect sizes ranging from .15 to .33. Thus, it is clear that only a narrow spectrum of the social-relational antecedents of voice have been considered, especially in terms of peer factors. A social-relational view is needed because voice is delivered through interpersonal communication where there is always a “receiver” to whom a “voicer” sends his or her suggestions (Bashshur & Oc, 2015).
Nevertheless, the ultimate decision to engage in voice is still an individual one, influenced by an individual’s motivation. Voice is self-initiated and change-oriented and, as such, is a form of proactivity in the workplace (Parker & Collins, 2010). However, like other forms of proactivity, voice involves psychological hurdles because others may not welcome it (Burris, 2012; Frese & Fay, 2001). Consequently, it is not always initiated readily or easily (Milliken, Morrison, & Hewlin, 2003), and individuals need active and positive motivation to speak out. As Parker, Bindl, and Strauss argued, because of the challenging nature of proactive behavior such as voice, individuals “need to feel confident they can both initiate proactive goals and deal with their consequences before they act” (2010: 834).
Crucially, to advance voice research, more work is needed to connect the social-relational context within which one operates to the personal decision to engage in voice. As Morrison elaborated, we need “to consider more fully the effects of one’s colleagues and relationships with one’s colleagues on the decision of whether to engage in voice or remain silent, because it is likely that voice is shaped by social and relational factors” (2014: 195). Our goal in this study is to advance voice research by linking one’s positive social standing at work (received respect) to one’s motivation to engage in voice. We argue that when employees have a high level of received respect in the workplace, they feel capable of making a difference by speaking out (what Parker et al., 2010, referred to as “can-do” proactive motivation) and they are in a more positive mood to do so (what Parker et al. referred to as “energized-to” proactive motivation). Next, we connect this respect-voice process to another important social-relational variable—perspective taking (PT)—by treating it as a predictor of received respect. Figure 1 (upper left panel) depicts this model. Below, we explain why we examine these components.
Theoretical Models
Received respect, which is one’s perception of one’s imputed worth (Rogers & Ashforth, 2017), is the starting point of our model. There are three reasons for this. First, employees are concerned with their imputed worth (Rogers, Corley, & Ashforth, 2017). As Vorauer noted, “People have an enduring, deeply entrenched need to know their standing with others that manifests itself in ongoing nonconscious monitoring of the social world for evaluative feedback” (2013: 72). Because employees are sensitive to status cues, received respect is an important variable in explaining many of the behavioral decisions in the workplace (Rogers & Ashforth, 2017; van Quaquebeke, Zenker, & Eckloff, 2009), including the decision to voice. Second, received respect captures whether one has a good reputation (Bartel, Wrzesniewski, & Wiesenfeld, 2012; Fuller, Hester, et al., 2006), and such social currency can strengthen one’s interest in voice, knowing that there are fewer objections or obstacles to face. This partly reflects the notion that received respect connotes an elite status advantage over others, making this variable unique compared with variables like coworker exchange quality (Sherony & Green, 2002) and team cohesion (Mathieu, Kukenberger, D’Innocenzo, & Riley, 2015). Third, received respect from coworkers operates rather independently from leadership, thereby extending the social-relational view of voice to include a prominent peer factor.
The next component of our model, proactive motivation, connects received respect to voice. When individuals feel respected by their peers, they are likely to experience positive changes in the motivational states that facilitate voice. First, received respect strengthens employees’ perceptions that their voice makes a difference. Consistent with this reasoning, in their model of proactive motivation, Parker et al. (2010) argued that one’s beliefs that one could bring about change and make a difference, or one’s control beliefs, were an important aspect of “can-do” motivation that stimulated proactivity. Second, because employees strongly value being respected (van Quaquebeke et al., 2009), we propose that received respect emotionally energizes them to become proactive rather than reactive. Such an energizing role of received respect maps onto the “energized-to” pathway theorized by Parker et al. in which activated positive affect fuels proactive behavior. We thus rely on both control beliefs and positive mood to explain why received respect is linked to voice. In so doing, we capture two of the three key motivational states that Parker et al. found proximally drive proactive behavior. 1 The importance of these motivational states has been largely neglected in other voice studies.
Finally, we examine PT (perspective taking) as an antecedent of received respect. PT has been defined as the cognitive process of imagining the world from another’s point of view (Parker & Axtell, 2001). It is a social-relational factor that matters to received respect because it captures a relational competence that helps individuals earn others’ appreciation and recognition. It is likely that perspective takers garner respect from others because they are viewed as more socially competent. For instance, PT fosters affiliative behavior such as being considerate and understanding in conversation (Goldstein, Vezich, & Shapiro, 2014). In turn, coworkers feel they are being listened to and appreciate the perspective takers’ thoughtful attempts. Thus, PT can enhance received respect and indirectly promote voice. However, despite its relevance, to date little has been known about the roles of PT in the nomological network of voice.
This study contributes to voice research in several ways. First, we test a social-relational view of voice that has been sidelined by the predominant focus on dispositional, job, and contextual factors as the key antecedents. Among those scant studies that have taken a social-relational perspective, the predominant focus has been on leadership factors (Chamberlin et al., 2017). Only social support and group identification have been examined as prominent peer factors that promote voice. Our study unveils the overlooked role of received respect in motivating voice. By doing so, it also extends the current focus of voice research on perceptions of social relationship quality (e.g., support, identification) to include individual perceptions of social standing as predictors of voice. In other words, it is not merely interpersonal warmth that motivates voice; one’s social standing, as reflected in received respect, also matters.
Second, we uncover two proactive motivations underlying employees’ decision to voice. Given that voice is challenging in nature (Burris, 2012; Lebel, 2016) and entails much work even after the suggestion has been made (Baer, 2012), many employees choose to remain silent (Brinsfield, 2013; Milliken et al., 2003). Thus, proactive motivation is arguably the most salient and proximal psychological determinant of voice. Yet voice researchers have seldom directly incorporated the specific types of proactive motivation into their theories.
Third, we identify PT as an antecedent of received respect. Addressing PT here helps direct researchers’ attention to the personal qualities that create better social standing, further enacting a social-relational view of voice that focuses on building one’s elite social status, not just amicable social relationships. Although there are other relational competence variables, such as emotional intelligence, empathy, political skills, and self-monitoring, PT is arguably the most important form of relational competence in the voice context because it effectively lowers interpersonal risk. Engaging in PT allows individuals to truly understand interpersonal risks from others’ perspectives, devise the best possible strategies to avoid them, and quickly adjust their behavior by putting themselves in the others’ shoes when they observe early signs of resistance. Other types of relational competence do not necessarily enhance this contextual knowledge that is so important to removing interpersonal risk (e.g., emotional intelligence or empathy focus more on empathizing with others’ emotions and feelings rather than understanding others’ perspectives, opinions, and arguments that lead to risks; political skills or self-monitoring entail adjusting behavior in different situations without necessarily engaging in high-level cognitive processing conducive to risk removals). Importantly, because PT is an interpersonal skill that can be acquired through training (Audia, 2012; Grizenko et al., 2000; C. J. Russell & Kuhnert, 1992), incorporating it into our model has provided a potential point of intervention for enhancing voice. That is, organizations can target and foster voice by shaping their employees’ PT tendencies.
Hypothesis Development
Received Respect, Control Beliefs, and Voice
Received respect is the degree to which an employee receives esteem from his or her peers (Smith & Tyler, 1997), often derived from the employee’s competence (Spears, Ellemers, & Doosje, 2005; Spears, Ellemers, Doosje, & Branscombe, 2006). It is the perceived worth accorded to a focal employee by others (Rogers & Ashforth, 2017; Sleebos, Ellemers, & De Gilder, 2007). A high level of received respect suggests that one is in good standing at work (Bartel et al., 2012) and affirmatively answers the question “Am I valued and held in high regard?” (Fuller, Hester, et al., 2006). Importantly, received respect is a form of particularized or appraisal-based respect (Rogers et al., 2017) that reflects the collective opinions others hold of the focal person’s worth (Bird, 2004; Huo, Binning, & Molina, 2010). Thus, gaining respect is distinct from developing amicable relationships. In essence, received respect must be “earned” by demonstrating competence.
Employees who receive more respect from others are more likely to engage in voice because received respect inspires greater “can-do” proactive motivation. According to the proactive motivation model (Parker et al., 2010), “can-do” motivation refers to one’s belief that one can successfully engage in proactive behavior. Control beliefs represent such motivation (Lebel, 2017). They refer to an employee’s belief that he or she can produce desired outcomes and prevent undesired outcomes (Skinner, 1996; Skinner, Chapman, & Baltes, 1988). As shown in Figure 1 (upper left panel), we propose a mediation process: employees’ received respect is likely to relate to their voice through their strengthened control beliefs at work.
Received respect → control beliefs
Employees’ received respect from coworkers should promote their control beliefs at work. The main reason is that received respect confirms the employee’s worth as a capable worker and signals to the employee that taking greater control over his or her work environment is possible given his or her high ability. Because respect must be earned (Spears et al., 2005; Spears et al., 2006), when an employee receives respect, it indicates that he or she is a competent and valued member of the organization (Bartel et al., 2012; Fuller, Hester, et al., 2006). This confirmation of worth is important because it cultivates the employee’s belief that he or she has what it takes to exert greater control in the workplace and obtain desired outcomes (Clarke, 2011; Eden, 1984; Greenberger & Strasser, 1986). This confirmation of worth also assures such employees that their colleagues are comfortable with their taking the lead to create outcomes. Received respect, in some sense, represents “the right to exert influence” (van Quaquebeke, Henrich, & Eckloff, 2007: 194). Empirical evidence has shown that those who earn greater respect as a result of their competence have stronger influence over their peers (Anderson & Kilduff, 2009; Cheng, Tracy, Foulsham, Kingstone, & Henrich, 2013) and are more likely to become leaders (Clarke, 2011).
The worth-confirmation mechanism leads to the expectation that received respect from coworkers is likely to relate to employees’ beliefs that they can control what happens in the workplace, that is, their control beliefs at work. This premise is consistent with the argument by Huo et al. that received respect can be viewed as “social currency” (2010: 202). Cheng et al. went further, claiming that “influence cannot be attained through coercive tactics such as bullying or intimidation, and instead is derived only from one’s apparent value to the group” (2013: 104-105). Sleebos et al. asserted that “individuals with respect-induced self-esteem are more confident that they can perform well, and in turn will address their efforts on behalf of their team” (2007: 331). These arguments suggest that the respect employees receive from colleagues becomes their currency or license to exert greater control at work.
Control beliefs → voice
On the basis of the premise that “can-do” motivation promotes more proactive behavior (Lebel, 2017; Parker et al., 2010), we propose that employees’ control beliefs relate to voice. Specifically, employees’ control beliefs increase the perceived probability that their voice attempts are likely to be effective. Control beliefs help employees see the work environment as structured, orderly, and surmountable (Bhanji, Kim, & Delgado, 2016; Duckworth, Peterson, Matthew, & Kelly, 2007; Friesen, Kay, Eibach, & Galinsky, 2014). The psychological benefits flowing from control beliefs motivate employees to more proactively engage in voice to initiate change, as they believe work outcomes are contingent on their behavior rather than on external, uncontrollable factors (Lefcourt, 1976; Phares, 1976).
This is consistent with expectancy theory (Vroom, 1964), in which employee behavior is portrayed as a function of whether their efforts are expected to effect change or achieve desired objectives. If employees believe they can control their work environment, they should be proactive in making changes to it (Kopelman & Thompson, 1976). Conversely, when they believe their efforts cannot effectively change the work environment in ways they want (i.e., having weak or no control beliefs), it creates passivity among them. Under such a scenario, employees have few reasons to speak up.
Hypothesis 1a (H1a): The stronger the employee’s received respect, the stronger the employee’s control beliefs.
Hypothesis 1b (H1b): The stronger the employee’s control beliefs, the stronger the employee’s voice.
Received Respect, Positive Mood, and Voice
Another reason why employees who receive respect are likely to engage in voice is that it stimulates their “energized-to” proactive motivation, referring to the affective rationale that prompts proactive goal pursuits (Parker et al., 2010). Positive mood precisely captures this type of motivation, as it represents a positive affective state that people experience (George, 1989). As shown in Figure 1 (upper left panel), we propose a second mediation process: employees’ received respect relates to their voice via their enhanced positive mood.
Received respect → positive mood
Employees’ received respect is likely to be associated with greater positive mood because received respect can fulfill many internal needs that lead to satisfaction. Specifically, receiving respect from coworkers sends important social cues to employees that their contributions are acknowledged and appreciated, that they are viewed as valued and praiseworthy members of the organization, that they are well integrated or occupy elite status positions in the social network at work, and that their future with the organization is likely to be promising and long lasting (Anderson & Kilduff, 2009; Cheng et al., 2013; Sleebos et al., 2007; van Quaquebeke et al., 2007). These positive social cues, all of which emanate from the respect received from colleagues, fulfill at least two prominent internal needs of employees that create positive mood.
First, all individuals have a need to belong, or the desire to establish meaningful and satisfying social relationships (Baumeister & Leary, 1995). Received respect signals to employees that their personal qualities are accepted and appreciated by other employees, fulfilling their need to belong (De Cremer & Tyler, 2005; Huo et al., 2010). As De Cremer and Mulder noted, “People value respect out of a concern to belong” (2007: 441-442). Second, all individuals have a need for competence, or the desire to effectively manage their external environment (Patrick, Knee, Canvello, & Lonsbary, 2007; Sheldon, Ryan, & Reis, 1996). Received respect can also fulfill this need because it signals to individuals that they are capable, influential, and reputable members of their organization (De Cremer & Tyler, 2005; Huo et al., 2010). De Cremer and Mulder emphasized that “people value respect because respect caters to reputational concerns” (442). These two needs are noteworthy because need satisfaction strongly promotes positive mood and more broadly, affective well-being (Deci et al., 2001; Ryan, 1982; Ryan & Frederick, 1997; Sheldon & Schuler, 2011). Consequently, employees’ received respect is likely to relate to their positive mood.
Positive mood → voice
We propose that employees’ positive mood is associated with voice because positive mood activates one’s behavioral approach system. The behavioral approach system is an intrapsychic mechanism that governs approach-oriented behavior; it directs people toward experiences and circumstances that bring rewards, improvement, and excitement (Watson, Wiese, Vaida, & Tellegen, 1999). When it is activated, individuals become approach oriented; for example, they actively seek out resources or the means to control the allocation of resources to attain meaningful goals. Put another way, the behavioral approach system governs one’s readiness for action (J. A. Russell, 2003).
Positive mood is one of the most proximal activators of the behavioral approach system (Erez & Isen, 2002; Ruud & Henk, 2005). That is, positive mood is likely to activate an approach-oriented mind set (Brockner & Higgins, 2001; Keltner, Gruenfeld, & Anderson, 2003). Watson, Clark, McIntyre, and Hamaker (1992) observed that people in a positive mood were more likely to actively seek out different social activities to participate in. Watson et al. (1999) further pinpointed the neurological pathway of this link by showing that individuals who were in a positive mood experienced greater activation in the part of the brain where approach-oriented behavior was initiated and monitored. Accordingly, we propose that employees’ positive mood relates to proactive behavior such as voice. Voice is considered an important approach-oriented behavior in the workplace, directing employees toward desirable goals (e.g., an improved workplace, strengthened team performance, a high-performing employer; Thomas, Whitman, & Viswesvaran, 2010; Tornau & Frese, 2013).
Hypothesis 2a (H2a): The stronger the employee’s received respect, the stronger the employee’s positive mood.
Hypothesis 2b (H2b): The stronger the employee’s positive mood, the stronger the employee’s voice.
PT as a Predictor of Received Respect
PT entails a deliberate cognitive process in which employees, in a nonjudgmental way, try to understand the thoughts and motives of others (Grant & Berry, 2011; Parker & Axtell, 2001). PT is widely regarded as “a valuable social skill” (Galinsky, Gilin, & Maddux, 2011: 105), suggesting that not everyone can effectively execute it. Perspective takers must adjust their anchored perspectives to accommodate the viewpoints of others (Epley, Keysar, Van Boven, & Gilovich, 2004). They must actively comprehend others’ thoughts, discern their targets’ points of view, alternate between mental frames, integrate the perspectives of themselves and others, and step outside their mental routines to engage in more demanding information processing (Janczyk, 2013; Ku, Wang, & Galinsky, 2015). They can then use the information obtained through their PT skills to anticipate the behavior and reactions of others (Davis, 1983) and devise a corresponding behavior (Fingerman, 1995; Galinsky, Wang, & Ku, 2008). Ryskin, Benjamin, Tullis, and Brown-Schmidt explained that PT “is not something that individuals do effortlessly” (2015: 899). According to Ryskin et al., it requires “a tremendous variety of higher level cognitive skills” (898). Thus, not all employees will engage in PT, and because PT is an appreciated social skill, those who engage in it more should receive greater respect from others.
How do targets know when a person has engaged in PT directed at them or others and that it in turn deserves their respect? Although PT is cognitive in nature, it is often targeted at others, leading to many observable cues during social interactions. Targets can discern a person’s PT attempts through such behavioral cues as whether the person patiently listens to their point of view, constantly refers back to their orientation, sincerely agrees with their opinions, analyzes the situation at hand or dissects a relevant issue in detail for them, and provides helpful advice from their perspective (Goldstein et al., 2014; Gutenbrunner & Wagner, 2016; Tuller, Bryan, Heyman, & Christenfeld, 2015). Targets can also find cues of another’s PT in group settings, such as when the focal person considers others’ input before making decisions (Galinsky & Moskowitz, 2000; Todd, Bodenhausten, Richeson, & Galinsky, 2011), whether the person remains open to diverse opinions throughout communications (Hoever, Van Knippenberg, Van Ginkel, & Barkema, 2012), and whether the person tries to reach satisfactory win-win agreements with others in negotiations (Galinsky, Maddux, Gilin, & White, 2008; Trotschel, Huffmeier, Loschelder, Schwartz, & Gollwitzer, 2011).
The relational competence of perspective takers observed by their targets forms the basis of the targets’ respect for them. Particularized respect emerges precisely because of competence (Spears et al., 2005). This competence can be based on either task or social success. Much of the research has argued that workplace respect is earned primarily through task competence (Grover, 2013). However, we argue that PT, a different type of competence, is also impactful for respect perceptions. Rogers and Ashforth (2017) note that particularized respect can be based on the target’s attributes and behaviors rather than merely achievements (i.e., task competence). In other words, meeting or exceeding prototypical character standards is important, in its own right, for particularized respect. Crucially, PT may signal that a person is meeting or exceeding good character standards (e.g., willing to serve the interest of others), with people’s views about character being aspects that are “deeply held and transcend the organizational context” (Rogers & Ashforth, 2017: 1591-1592) and hence important sources of particularized respect. van Quaquebeke and Eckloff (2010) similarly found that one’s positive leadership style (e.g., considering subordinates’ needs, being open to subordinates’ opinions) earned competence-based respect from subordinates. Thus, it appears that one’s relational competence, captured by PT, is likely to draw others’ respect. Others’ respect for perspective takers in turn becomes the received respect of the latter. Observable social cues of received respect include others’ deference to one’s decisions, hearing others’ expressions of confidence and trust in oneself, one’s opinions being sought, and being offered compliments on one’s achievements and milestones (Clarke, 2011; Dunning, Anderson, Schlosser, Ehlebracht, & Fetchnhauer, 2014; Rogers et al., 2017; van Quaquebeke & Eckloff, 2010).
Hypothesis 3 (H3): The stronger the employee’s PT, the stronger the employee’s received respect.
Research Program Overview
Following the “full-cycle organizational research” paradigm (Chatman & Flynn, 2005), we conducted experiments in a more controlled setting and subsequent field studies with a high level of generalizability. In Study 1 (N = 225), we examined the respect-voice link by manipulating received respect to see whether voice changed accordingly. In Study 2 (N = 248), we tested for the effects of PT on respect by manipulating PT to see whether PT subjects received greater respect than control subjects. Study 3 involved 735 dyads (either employee-coworker dyads or subordinate-supervisor dyads) and tested the entire proposed model.
Study 4 was a within-person design that revealed the degree to which domains covaried over time within a person (Hoffman, 2007). Between-person designs should be complemented by within-person designs (Molenaar, 2004; Ployhart & Vandenberg, 2010) because “many, perhaps even most research questions in psychology and micro-organizational behavior are in reality within-person questions” (Dalal, Bhave, & Fiset, 2014: 1399). A within-person design is applicable to this study because its variables have tended to change over time. Most importantly, received respect can fluctuate depending on what an employee accomplishes and whether the employee is able to live up to prescribed standards of behavior. For instance, an employee may garner respect for helping resolve a dispute, but that respect may be eroded when the employee displays personal bias in a subsequent team meeting. Rogers et al. (2017) observed that when strong respect cues were consistently given to new hires, the respect they felt was strong. This implies that respect cues can fluctuate and that it is not always possible to garner and maintain respect. Rogers and Ashforth (2017) also urged researchers to study how negative situations can undo positive situations that build respect and how low respect can motivate individuals to build up respect, hinting that received respect can change over time. Control beliefs, mood, voice, and PT have also been shown to fluctuate over time (Helgeson, Novak, Lepore, & Eton, 2004; Heller & Watson, 2005; Kilpatrick, Bissonnette, & Rusbult, 2002; Lin & Johnson, 2015; Liu, Song, Li, & Liao, 2017). Thus, in Study 4, we sampled 227 employees and conducted a 9-week within-person investigation.
Study 1
Participants and Research Design
Study 1 aimed to establish the causal effects of received respect on voice. We recruited 225 subjects from Amazon Mechanical Turk to take part in an experimental vignette (65% male, mean age = 37 years old, mean work experience = 15 years, 82% had college education).
The vignette began with the priming of received respect. In the treatment condition (n = 111), subjects were asked to imagine that they received strong respect from coworkers as a result of their ability and competence and that their coworkers always deferred to their decisions, always expressed confidence in them, always sought their opinions and views, and always complimented them on their achievements. In the control condition (n = 114), subjects were asked to imagine that they did not receive respect from their coworkers based on their ability and competence; coworkers seldom deferred to their decisions, seldom expressed confidence in them, seldom sought their opinions and views, and seldom complimented them on their achievements.
After being primed with one of these two conditions, all of the subjects were asked to report their control beliefs and positive mood if they worked within the described social environment. Items were adapted from the scales we used in our field studies (reported later) to measure control beliefs (α = .94; e.g., “Whether or not I am able to get what I want in this workplace is in my own hands”) and positive mood (α = .97; e.g., “Feeling determined at work”). We also measured manipulation checks using the received respect scale we adopted in our field studies (see the online supplemental material; α = .99; e.g., “My coworkers respect me”).
Subjects were then asked to continue to imagine themselves facing the following scenario. Their company had recently developed a profitable new product, although there had been a shortage of the materials needed to produce it. Thus, the company’s R&D department, along with the other departments, were meeting with the top managers in a product development meeting to discuss the production issues. Both the treatment and the control groups were then told that they represented the R&D department and were to attend the product development meeting. After reading this scenario, all of the subjects were asked to fill out a survey that measured their voice tendency in the product development meeting, using five items adapted from Maynes and Podsakoff (2014; α = .95). An example item is “In the current and upcoming product development meetings, I will suggest changes to work projects in order to make them better.”
Results
Our manipulation was successful. Subjects in the experimental condition reported a higher level of received respect (M = 4.39, SD = 0.61) than those in the control condition (M = 1.75, SD = 1.03; F(1, 223) = 544.80, p < .01). Moreover, a series of analyses of variance showed that subjects in the experimental condition reported a higher level of control beliefs at work (M = 4.12, SD = 0.68) than those in the control condition (M = 2.22, SD = 1.03), a higher level of positive mood at work (M = 4.04, SD = 0.71) than those in the control condition (M = 2.06, SD = 1.02), and a higher level of voice tendency (M = 4.16, SD = 0.77) than those in the control condition (M = 3.22, SD = 1.18). The between-group differences were significant in all three cases (p < .01).
Mediational analyses showed that control beliefs mediated the relationship between the manipulation and voice tendency. That is, manipulation (0 = low received respect, 1 = high received respect) positively predicted control beliefs (β = 0.95, SE = 0.06, p < .01), which in turn positively predicted voice tendency (β = 0.37, SE = 0.07, p < .01). The indirect effect was significant: effect = .35, SE = .07, 95% confidence interval (CI) = [.22, .48]. Thus, H1a and H1b were both supported.
Positive mood demonstrated similar mediation effects. The manipulation was positively related to positive mood at work (β = 0.99, SE = 0.06, p < .01), which in turn was positively related to voice tendency (β = 0.35, SE = 0.07, p < .01). The indirect effect was significant (effect = .35, SE = .08, 95% CI = [.19, .51]). H2a and H2b were thus supported. In sum, the above results provide causal support for our core premise that received respect positively predicts control beliefs, positive mood, and voice tendency.
Study 2
Participants and Research Design
Study 2 aimed to establish the causal effects of PT on received respect. A total of 248 undergraduate students from a local university participated in this lab experiment (mean age = 20 years old, 65% female). Upon arrival, the subjects were randomly assigned into Group A (n = 124) or Group B (n = 124). The two groups (now in different rooms) were then asked to study a work-related scenario. The scenario described a fictional company that had recently developed a profitable new product, although it could lead to potential long-term health concerns for the production workers. As a result, the company’s R&D and marketing departments decided to hold a meeting. From here, Group A subjects were instructed to take on the role of “Allen,” a representative from the R&D department, and Group B subjects were instructed to take on the role of “Bryan,” a representative from the marketing department. The scenario ended with Bryan offering several specific arguments to support the launch of the new product despite the health concerns (e.g., the production processes still met current legal regulations; the production processes could still be improved after the product was launched).
PT manipulation
After reviewing the scenario, Group A subjects (or Allen) were asked to respond to Bryan’s arguments. Following the PT manipulation developed by Galinsky, Maddux, et al. (2008) and Galinsky, Magee, Rus, Rothman, and Todd (2014), half of the Group A subjects (n = 62) were randomly assigned and instructed to take Bryan’s perspective when “interacting” with him, and the other half (n = 62) were instructed to remain objective and not adopt Bryan’s perspective.
Once the Group A subjects learned about the social approaches they needed to adopt, they completed a nine-item scale (α = .94) on how they intended to interact with Bryan during the meeting. The scale was adapted from past PT research (Ingoglia, Lo Coco, & Albiero, 2016; Peloquin & Lafontaine, 2010); sample items include “I will imagine myself to be Bryan and try to understand why he insists on launching the product,” and “Before criticizing Bryan, I will imagine how he would think and feel if I were in his position.” Next, the answers of the Group A subjects were distributed to their paired Group B subjects for review.
Group B subjects’ ascribed respect to Group A subjects
After reviewing the answers from their paired Group A subjects, Group B subjects (or Bryan) evaluated their Group A counterparts. We created a 15-item scale (α = .95; see the online supplemental material) based on previous studies of cues of particularized respect (e.g., De Cremer & Tyler, 2005; Rogers et al., 2017; Spears et al., 2005). These items captured competence-based respect, which has been the focus of our theory. Using competence-based items to measure respect also ensured that Group B subjects were not merely reciprocating with polite and civil responses to those who treated them politely and civilly (Meier & Semmer, 2013), such as perspective takers, and, thus, were evaluating Group A subjects positively. Instead, they evaluated whether Group A subjects (or Allen) were seen as socially competent workers who deserved their particularized respect. Sample items are “I think Allen is socially competent” and “I will give praise to Allen for the way he handles his job, if I were to work with him closely.” After Group B subjects completed this survey of ascribed respect, their answers were returned to their paired Group A subjects.
Group A subjects’ received respect from Group B subjects
Upon receiving the evaluations of the Group B subjects, the Group A subjects were instructed to complete a scale measuring their received respect from their paired Group B subjects. The nine-item scale (α = .95) was adapted from the received respect scales of Ng (2016) and Bartel et al. (2012). These two scales were also used in our field studies (see the online supplemental material). The sample items included “I think Bryan values my skills and abilities” and “I think Bryan respects me.” The lab experiment ended when both groups of participants finished their assigned tasks. They were paid, thanked, and dismissed.
Results
A manipulation check showed that our PT manipulation was successful; Group A subjects in the PT condition (M = 3.78, SD = 0.63) reported a significantly greater tendency to engage in PT in their interactions with Bryan compared to the Group A subjects in the control condition (M = 2.89, SD = 0.92). The difference was significant (F(1, 122) = 39.31, p < .01).
The PT manipulation among the Group A subjects resulted in Group B subjects’ according higher ascribed respect to the perspective takers (M = 3.80, SD = 0.60) than to the non–perspective takers (M = 3.13, SD = 0.71). The difference was significant (F(1, 122) = 31.68, p < .01). More importantly, the PT manipulation among the Group A subjects resulted in the perspective takers’ rating received respect (M = 3.75, SD = 0.87) higher than the non–perspective takers (M = 3.16, SD = 0.86). The difference was significant (F(1, 122) = 14.27, p < .01). A mediational analysis revealed that the ascribed respect Group B subjects gave to Group A subjects fully mediated the relationship between the PT manipulation (0 = no PT, 1 = PT) and Group A subjects’ received respect (indirect effect = .64, SE = .12, 95% CI = [.41, .87]). In summary, Study 2 demonstrated the causal effect of PT on received respect (with PT targets’ ascribed respect given to the perspective takers as the mediator), supporting H3.
Study 3
Participants and Research Design
Study 3 involved a convenience sample of 735 focal employees, along with their coworkers or supervisors. Two subsamples were recruited. In Sample A, there were 558 dyads, including 344 employee-coworker and 214 subordinate-supervisor dyads. In Sample B, an additional 177 employee-coworker dyads were recruited. The online supplemental material shows the details of our samples, including their demographics, job industries, and reported variables.
Phase 1
Using a convenience sampling strategy (Butts, Becker, & Boswell, 2015; Panaccio & Vandenberghe, 2012), we recruited participants from our own personal and professional networks. The study’s purpose, design, and incentives (US$10) were explained in invitation e-mails. Thereafter, interested subjects were asked to recommend a coworker they had regular interactions with to join the project and form a dyad. The coworkers were contacted by e-mail, invited to join, asked to confirm their coworker status through a declaration, assured of data confidentiality, and promised a small incentive (US$10) for their participation. We checked to ensure that no participants had more than one registration record. Participants were also encouraged to recommend other individuals or dyads to join the project. Ultimately, there were 344 independent employee-coworker dyads. Similarly, we recruited managers from our personal and professional networks. The managers were instructed to invite a direct subordinate to join the project. Ultimately, 214 independent subordinate-supervisor dyads were formed. Taken together, the above recruitment efforts resulted in 558 focal employee participants (Sample A). Their coworkers or supervisors provided voice ratings.
Phase 2
We then collected more employee-coworker dyads (Sample B) using the same recruitment strategy we followed for Sample A. We collected this additional sample to triangulate the findings, adopting different measures of our three main variables: received respect, voice, and PT. In addition, we included more correlates of voice as control variables in this sample. In the end, we gathered 177 independent employee-coworker dyads for our sample.
Measures
Received respect
In Sample A, received respect was measured by five items from Ng’s (2016) scale (α = .82; see the online supplemental material). An example item is “My coworkers respect me.” In Sample B, received respect was measured by four items from the Bartel et al. (2012) study (α = .86; see the online supplemental material). An example item is “My coworkers value my skills and abilities.”
Control beliefs
The respondents’ control beliefs were measured using a four-item scale from Mittal and Griskevicius (2014) in both Samples A and B (α = .80 and .75, respectively). An example item is “Whatever happens on my job mostly depends on me.”
Positive mood
Positive mood was measured using the 10-item Positive and Negative Affect Schedule (PANAS) scale of Watson, Clark, and Tellegen (1988) in both Samples A and B (α = .92 and .88, respectively). Examples of the emotions measured are enthusiasm, excitement, and interest.
Voice
In Sample A, employee voice was assessed by either their coworkers or supervisors, using the 5-item scale of Maynes and Podsakoff (2014; α = .93). An example item is “This employee makes suggestions about how to improve work methods.” In Sample B, coworkers provided voice ratings of the focal employees, using Liang, Farh, and Farh’s (2012) 10-item voice measure, which breaks down constructive voice into two components: promotive (α = .89) and prohibitive (α = .84) voice. Separating the two components allowed us to examine whether the proposed theoretical model applied to different subtypes of constructive voice.
Promotive voice consists of employee suggestions for improving the overall functioning of their organization (e.g., “This employee makes constructive suggestions to improve the organization’s operation”). Prohibitive voice comprises employees’ expressions of concern about work practices, incidents, or behavior that are harmful to their organization (e.g., “This employee proactively reports coordination problems in the workplace to the management”). Although both are well-intended changes, promotive voice is future oriented, focused on opportunities and improvement, whereas prohibitive voice is present and past oriented, focused on identifying and removing existing problems or constraints (Chamberlin et al., 2017). Our premise is that when employees feel respected, their proactive motivation is heightened. This manifests in a tendency to make different types of suggestions that help the organization improve, whether the suggestions are opportunity driven or based on problem removal. That is, respected employees are proactive both in looking for opportunities for growth (and hence promotive voice) and in identifying problems and constraints that hinder growth (and hence prohibitive voice).
PT
In Sample A, the respondents’ PT was measured by four items modified from Peloquin and Lafontaine’s (2010) scale (α = .82). An example item is “I try to put myself in coworkers’ shoes.” In Sample B, PT was measured by four items adapted from the scale of Ingoglia et al. (2016; α = .80). An example item is “Before criticizing my coworkers, I try to imagine how I would feel if I were in their place.”
Control variables
We controlled for several variables that were theoretically relevant to voice. First, better relationships with the rater could result in more favorable performance ratings that were independent of the ratees’ actual performance (Duarte, Goodson, & Klich, 1994). Thus, when the voice rating was provided by coworkers, we controlled for coworker exchange quality, which was assessed by Sherony and Green’s (2002) six-item scale. When the voice rating was provided by supervisors, we controlled for leader-member exchange, which was measured by Graen and Uhl-Bien’s (1995) six-item scale.
In Sample B, we collected extra measures of predictors of voice to more vigorously test our model. In a review of the voice literature by Chamberlin et al. (2017), the following variables were identified as being the most important in their respective categories: personal initiative and core self-evaluation in the personality category, felt responsibility and job autonomy in the job factor category, work engagement in the attitudinal predictor category, and transformational leadership in the leadership factor category. Personal initiative was measured by Frese, Fay, Hiburger, Leng, and Tag’s (1997) 7-item scale (α = .87); core self-evaluation was measured by Judge, Erez, Bono, and Thoresen’s (2003) 12-item scale (α = .85); felt responsibility was measured by the 5-item scale of Liang et al. (2012 ; α = .85); job autonomy was measured by Sadler-Smith, El-Kot, and Leat’s (2003) 3-item scale (α = .85); work engagement was measured by Schaufeli, Bakker, and Salanova’s (2006) 9-item scale (α = .89); and transformational leadership was measured by Tepper and Percy’s (1994) 12-item scale (α = .92).
Results
Table 1 provides the descriptive statistics. We first assessed measurement models through confirmatory factor analyses (CFAs), using the Tucker-Lewis index (TLI), comparative fit index (CFI), and root mean squared error of approximation (RMSEA). All of the fit indices are provided in Table 2. The CFA model for both samples had acceptable fit (TLI = .97/.93, CFI = .97/.94, RMSEA = .04/.06), suggesting that our scales had fairly good psychometric properties.
Means, Standard Deviations, and Correlations Among Study Variables in Studies 3 and 4
Note: Correlations for the two samples in Study 3 are between-person correlations; correlations for Study 4 are within-person correlations aggregated across Weeks 2 through 9 except for impression management, which was measured at Week 1. Coefficient alpha values are shown on the diagonal in parentheses.
p < .05.
p < .01.
Fit Indices
Note: In Study 3, we shortened three lengthy scales (positive mood, core self-evaluation, transformational leadership; all of them have 10 items or more) by either using parcel indicators or treating them as composites. Δχ2 = change in chi-squared value compared to the proposed model; Δdf = change in degree of freedom compared to the proposed model; TLI = Tucker-Lewis index; CFI = comparative fit index; RMSEA = root mean squared error of approximation; RR = received respect; VO = voice; CB = control beliefs; PM = positive mood.
Next, we observed that the proposed model had acceptable fit for Sample A (TLI = .96, CFI = .96, RMSEA = .05). The fit was marginally acceptable for Sample B (TLI = .91, CFI = .92, RMSEA = .06), perhaps due to its smaller sample size compared with Sample A (177 vs. 558). The RMSEA for Sample B, however, met the common cutoff (Hu & Bentler, 1999). The standardized path estimates are shown in Figure 1 (upper right and lower left panels).
In Sample A, H1a was supported. Received respect was related to control beliefs (β = 0.49, p < .01). However, control beliefs were not related to other-rated voice, despite a significant zero-order correlation (r = .13, p < .05). Thus, H1b was only marginally supported. Both H2a and H2b were supported; received respect was associated with positive mood (β = 0.42, p < .01), which in turn was related to other-rated voice (β = 0.17, p < .01). Finally, H3 was supported, as employee PT was related to received respect (β = 0.37, p < .01).
In Sample B, H1a and H1b were supported; received respect was related to control beliefs (β = 0.51, p < .01), which in turn was related to promotive (β = 0.32, p < .01) and prohibitive voice (β = 0.19, p < .05). H2a and H2b were partially supported. Received respect was related to positive mood (β = 0.26, p < .01), which in turn was associated with promotive voice (β = 0.21, p < .01) but not with prohibitive voice. Employee PT was related to received respect (β = 0.37, p < .01), supporting H3.
Table 3 illustrates the mediation effects embedded in the proposed model using Hayes’s (2013) bootstrapping approach. Especially noteworthy is the finding that (a) control beliefs marginally mediated the respect-voice link in Sample A (indirect effect = .05, 95% CI = [.00, .10]), (b) control beliefs marginally mediated the respect–promotive voice link in Sample B (indirect effect = .04, 95% CI = [.00, .08]), (c) positive mood mediated the respect-voice link in Sample A (indirect effect = .07, 95% CI = [.03, .11]), (d) PT was indirectly related to voice via received respect and positive mood in Sample A (indirect effect = .02, 95% CI = [.01, .03]), and (e) PT was indirectly and marginally related to promotive voice via received respect and control beliefs in Sample B (indirect effect = .01, 95% CI = [.00, .02]).
Mediation Relationships
Alternative models
In Alternative Model 1, received respect was specified to directly relate to voice to ascertain whether it affected voice through proactive motivation as argued. In both samples, the change in the chi-squared value was not significant, nor were the newly added paths, supporting our proactive motivation process. In Alternative Model 2, control beliefs were specified to relate to positive mood to determine whether the marginal effect of control beliefs on voice in Sample A could be explained by the effect being transferred via positive mood. We found in Sample A (not in Sample B) that the control beliefs–positive mood link was significant (β = 0.38, p < .01), corroborating our conjecture.
Study 4
Participants and Research Design
Study 4 adopted a 9-week within-person design wherein the variables were measured weekly starting from the 2nd week. With the help of our alumni office, we contacted alumni through e-mail. In the invitation e-mail, we explained the purpose and design of the study. The alumni were offered US$25 in return for complete participation in nine waves of weekly surveys. Close to 1,000 e-mail invitations were sent, and 261 alumni who were employed at the time were successfully recruited (response rate = 26%). During the first weekend, the respondents were invited to fill out a general survey. Over the next eight Friday evenings, they were sent identical surveys. The survey response rate was between 95% and 99% across the eight weekends (N = 227 employees; 2,043 week-level responses). We observed that the sociodemographic background of those who completed all of the weekly surveys was not significantly different from those who did not.
Measures
The measures used in Study 3 (Sample A) were also used in Study 4, with a slight adaptation to the weekly context. The average reliabilities across weeks were received respect (α = .86), control beliefs (α = .84), positive mood (α = .92), voice (α = .91), and PT (α = .87).
Using the respondents’ self-ratings of voice is defensible. First, employees are cognizant of the subtleties of their voice and can gauge whether their suggestions are fundamentally or marginally valuable because their mental representations of their own behavior incorporated the unique historical and contextual factors that affected that behavior (Janssen, 2001). Second, employees may have accurate knowledge of their voice intensity when they make suggestions to others that are unknown to their supervisors or coworkers. Third, prior meta-analyses showed that the nomological network of self-ratings and other ratings of job behavior were not that drastically different (Berry, Carpenter, & Barratt, 2012; Carpenter, Berry, & Houston, 2014).
Control variables
Impression management was controlled for, given that self-ratings of voice could partially reflect such a tendency (Fuller, Barnett, Hester, Relyea, & Frey, 2007). The 10-item scale of Blasberg, Rogers, and Paulhus (2013) was administered at Week 1 (α = .71).
Results
Our weekly variables (at Level 1) were obtained from multiple individuals (at Level 2); each respondent reported the study variables eight times over 8 weeks. Thus, we used multilevel techniques to analyze the data. First, weekly received respect was specified to affect weekly control beliefs and positive mood, which in turn were specified to affect weekly voice. Second, weekly PT was then specified to affect weekly received respect. Third, as these weekly relationships were nested within higher-order clusters (i.e., “persons”), we specified “persons” at Level 2 to control for broad between-person differences that might confound the Level 1 relationships. Fourth, we specified between-person differences in impression management at Level 2 to affect weekly voice. In so doing, the findings about weekly voice removed the bias induced by individuals’ impression management tendency. This is the only cross-level effect that we specified in the multilevel testing model.
As shown in Table 2, our multilevel CFA model had acceptable fit (TLI = .94, CFI = .95, RMSEA = .04). Measurement invariance tests also showed that none of the items used had significantly different factor loadings across time points. In addition, we observed that a large portion of the variance in each variable was attributable to within-person variation: received respect (44%), control beliefs (54%), positive mood (45%), voice (49%), and PT (50%).
Hypothesis testing
Our proposed model had a marginally acceptable fit, although the RMSEA indicated a good fit (TLI = .90, CFI = .91, RMSEA = .04). The standardized estimates are provided in Figure 1 (lower right panel). Weekly received respect was related to weekly control beliefs (β = 0.42, p < .01), which in turn was related to weekly voice (β = 0.19, p < .01). These results supported H1a and H1b. Weekly received respect was also related to weekly positive mood (β = 0.35, p < .01), which in turn was related to weekly voice (β = 0.16, p < .01), supporting H2a and H2b. Weekly PT was related to weekly received respect (β = 0.44, p < .01), supporting H3. We then examined the mediation effects embedded in the model using Preacher, Zyphur, and Zhang’s (2010) multilevel mediation tests. The results, shown in Table 3, suggest that (a) the mediating effect of control beliefs in the respect-voice link was significant (indirect effect = .10, 95% CI = [.05, .16]), (b) the mediating effect of positive mood in the respect-voice link was significant (indirect effect = .07, 95% CI = [.03, .11]), and (c) PT was indirectly related to voice via received respect and control beliefs (indirect effect = .05, 95% CI = [.02, .07]) or via received respect and positive mood (indirect effect = .03, 95% CI = [.01, .05]).
Alternative models
The same two alternative models were tested. In Alternative Model 1, weekly received respect was specified to relate to weekly voice. The nonsignificant change in fit suggests that as in Study 3, the relationship between received respect and voice was largely explained by proactive motivation, as expected. In Alternative Model 2, weekly control beliefs were specified to relate to weekly positive mood. The decrease in chi-square value and the newly added path were significant (β = 0.50, p < .01), suggesting that it was reasonable to specify this relationship. This result was also consistent with what was observed in Study 3.
Supplementary analyses
To make stronger causal inferences, we tested a model that controlled for each variable’s baseline measure taken at Week 1 so that the initial effect was removed. That is, the Week 1 baseline measures were specified to relate to their respective variables subsequently measured. Furthermore, we added 1 week of lag between each stage of effects to provide stronger causal evidence: received respect in Week 3 was specified to relate to control beliefs and positive mood in Week 4, which in turn were specified to predict voice in Week 5. PT in Week 2 was specified to relate to received respect in Week 3.
We observed that in such a cross-lagged model, received respect in Week 3 predicted both control beliefs (β = 0.27, p < .01) and positive mood (β = 0.25, p < .01) in Week 4, controlling for baseline control beliefs (which explained 16% of the variance) and positive mood (which explained 18% of the variance). However, although control beliefs in Week 4 did not predict voice in Week 5, positive mood in Week 4 did (β = 0.23, p < .01), while controlling for baseline voice (which explained 17% of the variance). Moreover, as expected, PT in Week 2 had a significant effect on received respect in Week 3 (β = 0.28, p < .01) while controlling for baseline received respect (which explained 26% of the variance). These significant findings suggest that the causal directions implied by our proposed model are largely reasonable.
Plots of effects of mediators on voice
One of the key proposed mediation effects was that of proactive motivation (control beliefs, positive mood) in the relationship between received respect and voice. In the online supplemental material, we plotted the effects of control beliefs and positive mood on voice, controlling for the effects of received respect. Researchers suggest that this plot is useful for illustrating the strength of mediation, especially the importance of the mediators to the outcome variables (MacKinnon, Fairchild, & Fritz, 2007).
General Discussion
Implications for Theory Development
We gathered evidence from four studies to show that voice could be explained by a social-relational mechanism revolving around received respect, proactive motivation, and PT. In Study 1, subjects who were primed to feel greater received respect reported greater voice (through the mediating effects of proactive motivation) than the control subjects. In Study 2, subjects who were in the PT condition were more likely to receive respect from their peers than those in the control condition. In Study 3 (Sample A), employees’ received respect predicted their other-rated voice through their positive mood, whereas the mediating effect of control beliefs were marginally significant. PT predicted received respect, as expected. In Study 3 (Sample B), control beliefs marginally mediated the effects of received respect on both promotive and prohibitive voice. Moreover, PT predicted received respect, as expected. In Study 4, weekly control beliefs and positive mood significantly mediated the effects of weekly received respect on weekly self-rated voice. Weekly PT predicted received respect, as expected.
The aforementioned findings contribute to voice research in various ways. First, and somewhat surprisingly, the social-relational antecedents of voice have been examined much less frequently than other factors, evidenced by the limited number of correlates observed in the Chamberlin et al. (2017) review. Moreover, beyond leadership, the effects of other relational realities on voice have largely remained undiscovered. For instance, in terms of peer factors, only social support and group identification have been consistently examined (Chamberlin et al., 2017). Recently, some researchers explored how one’s social network relates to voice (Venkataramani, Zhou, Wang, & Shi, 2016). All of these studies were interested in the effects of the quality or quantity of social relationships on voice. Our findings move the field of voice forward in an important way. Namely, this study shows that a person’s social standing (rather than relationship quality) affects voice. It is important to acquire some form of social reputation or elite status in executing voice, as when employees feel respected for their competence, they are more prone to engage in voice. Received respect acts as a confidence booster and an energizer, prompting employees to become proactive. In some sense, received respect gives employees the social currency or license to engage in proactive behavior that can disrupt the status quo. Our study demonstrates that such a social currency or licensing mechanism goes beyond the prevailing interest in the quality and quantity of coworker relationships.
The second contribution of this study is its revelation of the proximal psychological rationales preceding voice. Although researchers have acknowledged that proactive motivation underlies proactive behavior (Parker et al., 2010), it has been less common for researchers to incorporate these motivational mechanisms into their model testing. This also applies to voice research. We contribute to voice research by measuring and testing two main types of proactive motivation, including “can-do” motivation (captured by control beliefs) and “energized-to” motivation (captured by positive mood), as the precursors of voice. A lack of understanding of how respect psychologically activates employees may be the core reason why, despite its importance, respect has not received much attention from organizational researchers (Rogers & Ashforth, 2017). Our study resolves this issue and shows that received respect affects one’s voice because it activates psychological change to “can-do” and “energized-to” proactive motivation.
One finding on control beliefs is noteworthy. We observed that control beliefs did not have positive effects on voice in Study 3 (Sample A), although such a relationship emerged as expected in Sample B and in Study 4. Tangirala and Ramanujam (2008) reported that a low level of control motivated employees to speak up to gain more control, whereas a high level of control also empowered employees to actively engage in voice. As a result, the net effect of control beliefs on voice may sometimes be weak. Another explanation is that the effects of control beliefs on voice are partially transmitted through positive mood, which was what we found in our alternative model testing. Research has shown that cognitive beliefs about control are closely intertwined with the positive affective experience stemming from such beliefs (Breckler, 1984; Chaiken & Baldwin, 1981; Diener, Kanazawa, Suh, & Oishi, 2015).
This study also contributes to the enactment of a social-relational view of voice by considering PT as an indirect predictor of voice via received respect. PT not only determines peer relationship quality but also represents one’s relational competence that can indirectly affect voice by elevating one’s social status. One consistent finding that emerged in this study was that perspective takers were more likely to be respected by others. This was seen in Study 2, Study 3 (both Samples A and B), and Study 4. Thus, the predictor role of PT in the respect-voice relationship is quite robustly supported. Future theories can use this robust finding as the foundation for identifying other social-relational antecedents of received respect and voice.
Despite the supportive findings, there are two important ways to further strengthen the proposed social-relational model of voice that can be explored in future theory building. The first is to examine reverse causation. We presume that received respect relates to employees’ voice. Nevertheless, it remains possible that voice precedes respect such that individuals who speak out more may earn respect more easily from others. This is similar to findings that individuals who speak up more receive better performance ratings (Grant, 2013; Van Dyne & LePine, 1998; Whiting, Maynes, Podsakoff, & Podsakoff, 2012). In addition, we have assumed that PT precedes received respect. It is possible that one’s status affects the degree to which one engages in PT, with higher-status individuals being more likely to engage in it (Blader, Shirako, & Chen, 2016). It is also possible that individuals have more mental resources for PT only after they have become settled in their roles as well-respected employees; the pursuit of respect may absorb those mental resources. Under both of these scenarios, received respect should precede PT. Thus, future theories need to consider alternative ways in which the respect-voice link unfolds.
The second way is to consider how PT shapes the respect-voice process beyond its role as a predictor of respect. Its moderator role is especially intriguing because there are counteracting forces. On one hand, it can be argued that PT enhances the respect-voice process. Although proactive motivation (as a result of received respect) should relate to voice, interpersonal risks remain, such as the uncertainty surrounding a response or fear that others will not be receptive to one’s voice (Burris, 2012; Edmondson, 1999). PT reduces the uncertainty surrounding voice by helping the individual imagine how the other party will respond to his or her voice, turning “can-do” or “energized-to” motivation into “will do,” and removing barriers to action. In other words, PT enables employees to engage in voice in ways they imagine will be well received.
On the other hand, it can be argued that PT limits the respect-voice process. For instance, perspective takers may overthink others’ opinions of their voice and become reluctant to speak out as a result of such magnified concerns. Under that scenario, the respect-voice process may be hindered. Moreover, when a person adopts the perspective of a threatening target, it creates dissonance that may result in psychological distance from the target (Okimoto & Wenzel, 2011; Sassenrath, Hodges, & Pfattheicher, 2016). Depending on who the target is, PT can make perspective takers feel that their ego or self-concept is threatened, weakening the respect-voice relationship. Contextual conditions can also play a role. For instance, in a highly competitive environment, PT activates self-interested motives, generating self-centered rather than prosocial behavior (Epley, Caruso, & Bazerman, 2006). That is, understanding others’ motives in a competitive environment accentuates the need to protect or serve one’s own motives and therefore counteracts the potentially positive effects of respect on voice. The above discussion suggests that much more work is still needed on the moderating roles of PT.
Study Limitations
This research has some limitations. First, Studies 1 and 2 used hypothetical work-based scenarios, which limit its generalizability. However, the experimental design allowed for a stronger causal inference. Future research should consider manipulating respect and PT in a field setting. Second, in Study 3, the data were collected concurrently, albeit from different sources. Absolute causality could not be inferred. Future research in which the measurement of the respect-voice link is separated may better disentangle the causes and effects.
Third, in Study 4, all of the five core variables were self-reported and, thus, vulnerable to the influence of common method variance. However, as noted before, self-ratings of voice have some advantages (Janssen, 2001). Moreover, this study controlled for the impression management tendency and was longitudinal in design. Future research may consider more objective measures of voice (e.g., the number of suggestions, counting instances of speaking up).
Fourth, this study focused only on constructive voice (including promotive and prohibitive voice). Other types of voice, such as defensive voice, whistleblowing, complaining, and bad-mouthing (Maynes & Podsakoff, 2014), were not examined. Study 3 (Sample B) showed that the proposed respect-voice mechanism was stronger for promotive rather than prohibitive voice, perhaps because received respect and heightened proactive motivation directed employees’ attention to opportunities for improvement. Future research may clarify the effect of respect on various forms of voice and identify the respective mechanisms involved.
Fifth, this study sampled respondents from mixed job industries. Future research should examine whether the respect-voice link is stronger in certain types of work groups or contexts than others. For instance, this link may be stronger in jobs where cooperation is key (Van der Vegt, Emans, & Van de Vliert, 2001). Received respect may be important to successfully initiating change in such an environment.
Finally, our samples in Studies 3 and 4 were relatively young; more than 80% of participants were younger than 30 years old and a majority had tenure of less than 3 years. This might have affected our results. For instance, older and longer-tenured workers should have greater work knowledge (Ng & Feldman, 2010) and therefore should receive more particularized respect in the workplace. Having said that, older workers may also become stereotyped at times as being less competent and motivated in their jobs (Ng & Feldman, 2012). There is also evidence that older and longer-tenured workers are more likely to engage in citizenship behavior (Ng & Feldman, 2008, 2010), including voice, although the effect sizes are not strong. PT may also be affected by age and tenure. For instance, the ability to engage in PT may gradually decline in old age (Rakoczy, Wandt, Thomas, Nowak, & Jana, 2018), although there is also evidence that young adults do not frequently engage in PT (Flannery & Smith, 2017). In addition, gaining tenure allows employees to more easily put themselves in others’ shoes because they have experienced many different roles or tasks (Tesluk & Jacobs, 1998). However, it can also be argued that junior, short-tenured employees may be more fearful of taking social risks at work and therefore try to engage in PT more often to overcome such fears (Loschelder, Trotschel, Huffmeier, Schwartz, & Gollwitzer, 2011). Although we did not observe any systematic effects of age and tenure on respect, voice, and PT across our samples, we urge researchers to become aware of the potential influences of age and tenure on our proposed social-relational model of voice and to replicate our results in other samples.
Implications for Managerial Practice
Voice has been widely encouraged to promote organizational innovation and adaptation. However, practitioners still find it difficult to motivate employees to speak up (Matsunaga, 2015; Milliken et al., 2003). Thus, one implication for managerial practice is that instead of blindly focusing on encouraging employees to use their voice—which may adversely lead to employee stress and resistance—organizational leaders may want to focus on fostering a respectful culture where employees feel able and energized to speak up. As we have demonstrated, received respect is the core reason why employees are more likely to engage in voice, especially promotive voice that serves to improve organizations. Managers seldom prioritize the promotion of respect in the workplace. Although employees must be competent to garner greater respect from others, an organizational culture that publicly emphasizes respect for the technically and socially competent may be more likely to link received respect and voice. Encouraging this link, for example, by giving employees verbal praise and nonverbal approval, showing appreciation for their contributions, celebrating their achievements, and elevating their public image, is particularly important (Carmeli, Dutton, & Hardin, 2015).
This study has demonstrated one further and important way to promote employees’ feelings of received respect. We showed that PT was a consistent predictor of received respect. Thus, managers are strongly encouraged to promote PT among their employees (Audia, 2012). To that end, the selection of employees with high PT orientation, managers’ demonstrating strong PT behavior (thereby serving as role models), and PT training (Grizenko et al., 2000; C. J. Russell & Kuhnert, 1992) may be beneficial to organizations promoting change and continuous improvement. Employees who have doubts about engaging in voice can first consider engaging in PT. This promotes received respect that in turn activates their proactive motivation to voice.
Overall, the significant respect-voice relationship is good news for both employees and managers. Employees often need active and positive motivation to speak out in the workplace (Morrison, 2014; Parker et al., 2010), and such motivation is not necessarily easy to find (Brinsfield, 2013). This study reveals that garnering respect from others, such as engaging in more PT, is a good starting point. With heightened received respect, employees find the motivation to engage in different types of proactive behavior, including voice. Our results thus suggest to managers that received respect is an important social-relational antecedent of voice, that respected employees have strengthened proactive motivation to make stronger contributions, and that PT can significantly enhance received respect.
Supplemental Material
Online_Supplement_aggregate – Supplemental material for Received Respect and Constructive Voice: The Roles of Proactive Motivation and Perspective Taking
Supplemental material, Online_Supplement_aggregate for Received Respect and Constructive Voice: The Roles of Proactive Motivation and Perspective Taking by Thomas W. H. Ng, Dennis Y. Hsu and Sharon K. Parker in Journal of Management
Footnotes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
