The Effect of Transformational Leadership and Job Autonomy on Promotive and Prohibitive Voice: A Two-Wave Study

Abstract

Although there is a vast amount of research on leadership and improvement-oriented voice behavior, the amount of cross-lagged research on leadership that also incorporates more challenging forms of voice is sparse. This article reports on a two-wave study of white-collar workers in a Norwegian medical technology company, investigating the relationship among employees’ perceived transformational leadership behaviors, job autonomy, and promotive and prohibitive voice. Testing our results cross-lagged, we demonstrate that perceived transformational leadership is significantly related to prohibitive voice over time, whereas this effect worked in the opposite direction for promotive voice. We also explore the boundary conditions of transformational leadership, demonstrating that perceived job autonomy strengthens the effect of transformational leadership on prohibitive voice. Implications for theory and research are discussed.

Keywords

transformational leadership promotive voice prohibitive voice job autonomy

Introduction

In 2014, a scandal at General Motors resulted in the recall of nearly 30 million cars worldwide and the company paying compensation for 124 deaths. Later investigations revealed how employees in the company had been aware of the technical problems causing the scandal but had failed to communicate them upwards. The company’s CEO, Mary Barra, stated that main reasons for the scandal were

the failure of individual employees in several departments to address a safety problem . . . [and that] employees repeatedly failed to disclose critical pieces of information that could have fundamentally changed the lives of those impacted. . .

The GM scandal and the above quotations reflect the importance and practical necessity of what the literature refers to as voice behavior. Voice behavior is defined as employees’ discretionary communication of ideas, suggestions, and concerns at work with the purpose of improving organizational functioning (LePine & Van Dyne, 1998; Morrison, 2011) and is crucial for an organization’s self-improvement, renewal, and crisis prevention (Liang, Farh, & Farh, 2012). However, voice behavior can also be a risky endeavor for the employee as it challenges the status quo in the company and, thus, the supervisors responsible. Accordingly, as was the case in the GM scandal, many employees choose not to express their concerns or constructive ideas regarding the organization’s efficient functioning, because they feared negative response from their superiors (Milliken, Morrison, & Hewlin, 2003). Correspondingly, leadership is underlined as an important antecedent of voice, as employees will “read the wind” to establish whether it is safe and worthwhile to speak up (Detert & Burris, 2007).

The past two decades have seen a large number of studies regarding the relationship between leadership and voice behavior (cf. Detert & Burris, 2007; Duan, Li, Xu, & Wu, 2017; Dutton, Ashford, Lawrence, & Miner-Rubino, 2002; Liu, Zhu, & Yang, 2010; McClean, Burris, & Detert, 2013; Svendsen & Joensson, 2016; Svendsen, Jønsson, & Unterrainer, 2016; Tangirala & Ramanujam, 2012). Despite this progress, we argue that research on leadership and voice can be improved in three specific ways.

First, research has focused almost exclusively on promotive voice—that is, the employees’ expression of ideas for new products or projects (Liang et al., 2012)—rather than prohibitive voice—defined as the employees’ communication of concerns and factors that may harm the organization (Liang et al., 2012; Morrison, 2011; Van Dyne, Ang, & Botero, 2003). Keeping the GM scandal in mind, this lack of research is unfortunate, as prohibitive voice may be of even more relevance for practice than promotive voice. Theoretical work (Morrison, 2011) and empirical findings (Liang et al., 2012; Lin & Johnson, 2015; Svendsen et al., 2016) have suggested that the motivation of the two forms of voice may vary. Specifically, challenging the status quo by calling attention to factors that are harmful to the organization may implicate a direct failure of someone who is responsible (Burris, 2012; Liang et al., 2012), which increases the interpersonal risk taken by the reporting employee. Accordingly, self-protective motives are found to be more important for prohibitive than promotive voice (Liang et al., 2012). A leader is an important source for employees to establish the degree of interpersonal risk associated with speaking up (Detert & Burris, 2007). However, despite recent theory and empirical studies highlighting the importance of studying promotive and prohibitive voice separately, research on the relationship between leadership and prohibitive voice is sparse (Svendsen et al., 2016).

Second, although most research on leadership and voice proposes that leadership precedes voice, these studies have a cross-sectional design (see Detert & Burris, 2007, for a notable exception), which precludes the ability to explore the dynamics between leadership and voice over time or to investigate the reciprocal relationships between the two constructs. For example, the relationship between leadership and voice may work in the opposite direction, in that employees who express themselves may also come to see their leader as open to employee voice (Duan et al., 2017). Correspondingly, there is a lack of time-lagged, autoregressive research into the reciprocal relationship between leadership and voice.

Last, much of the research concerning voice and leadership has focused on leadership as an isolated predictor variable and has not considered how different proximal motivations may be present simultaneously when individuals decide whether to voice (Tangirala & Ramanujam, 2012). Therefore, researchers have called for more studies on how contextual variables interact with leadership to increase or decrease the effect of leadership behavior (Ashford, Sutcliffe, & Christianson, 2009). In light of these three points raised, the main goal of our study is to explore how leadership relates to promotive and prohibitive voice over time, and how contextual variables may affect this relationship.

In order to accomplish this goal, we have developed and tested a model in which we propose that the employees’ perceived transformational leadership is positively related to promotive and prohibitive voice. We also conducted post hoc analyses to explore the potential differential and reciprocal effect between the transformational leadership on promotive as well as prohibitive voice. The research on transformational leadership and promotive voice has been positive but inconclusive. For example, Detert and Burris (in the cross-sectional part of their 2007 study), Liu et al. (2010), and Duan et al. (2017) all found a positive, direct relationship between transformational leadership and improvement-oriented voice. On the other hand, Rank (2006), Svendsen and Jønsson (2016), and Detert and Burris (in the time-lagged part of their 2007 study) found no significant direct relationship between the two factors. These inconclusive results suggest a need for more cross-lagged studies on the topic in order to explore the relationship between the two phenomena. Furthermore, to the best of our knowledge, the relationship between transformational leadership and prohibitive voice remains unexplored. In light of the importance of prohibitive voice (Liang et al., 2012; Morrison, 2011), this knowledge is crucial in order to understand how leaders can stimulate this type of employee behavior.

There may be several contextual variables that affect the relationship between transformational leadership and promotive as well as prohibitive voice. Drawing on theory of planned behavior (Ajzen, 1985, 1991, 2011; Ajzen & Madden, 1986) and self-determination theory (Gagné & Deci, 2005; Ryan & Deci, 2000, 2006), we propose that the experience of job autonomy—that is, the degree or level of freedom and discretion an employee experiences on the job (Jønsson & Jeppesen, 2012)—interacts with transformational leadership to have a positive impact on the relationship with promotive and prohibitive voice. Theory of planned behavior states that attitudes, norms, and behavioral control play a role for intentions and overt behavior (Ajzen, 1985, 1991, 2011). Furthermore, self-determination theory (Gagné & Deci, 2005; Ryan & Deci, 2000, 2006) propose autonomy to be pivotal in creating own values and norms, choosing actions, and co-creating and internalizing norms. Chatzisarantis, Hagger, and Brickell (2008) found support for these notions in a prospective study. They discovered that autonomy support directly and indirectly affected behavior (mediated by intentions and attitudes). In line with this, Ajzen and Madden (1986) argue that both intention and control are necessary for performance of a behavior such that the effect of intention on behavior depends on perceived behavior control. Thus, we argue that transformational leadership, which stimulates and motivates the employee directly, will be positively related to voice behaviors, conditioned on the level of autonomy so that high autonomy enhances the effect of transformational leadership.

The present study makes several unique contributions to the research on both voice behavior and transformational leadership. First, we extend previous voice research by exploring how leadership is directly related to promotive as well as prohibitive voice. By incorporating prohibitive voice in our understanding of voice, and making post hoc analyses of the difference between them, we respond to the call for more research on how leadership may affect promotive and prohibitive voice differently (Chamberlin, Newton, & LePine, 2017; Liang et al., 2012; Morrison, 2011; Svendsen et al., 2016). Second, the present study contributes to the understanding of the causal dynamics between transformational and voice behavior over time. Our investigation suggests that transformational leadership can both shape the individuals’ propensity to voice and be shaped by the employees’ previous voice. However, this effect is dependent on the prohibitive or promotive content of the employee’s voice. Thus, we give a broader perspective on the reciprocal effect of transformational leadership on voice that also incorporates how followers shape the perception of their leader based on previous voicing. This exploration also provides a perspective on why previous time-lagged research has failed to find a direct effect of transformational leadership on promotive voice (Detert & Burris, 2007) and serves as a response to researchers who have highlighted the possibility for reverse or reciprocal causality between transformational leadership and voice (Duan et al., 2017). Last, our investigation of the moderating effect of job autonomy increases the understanding of how the effect of transformational leadership may be strengthened in certain environments. To date, transformational leadership has often been investigated as an isolated predictor variable; the present study renders important information about how transformational leaders may be more or less effective in eliciting voice from their employees, depending on the employees’ perception of job autonomy.

Theory and Hypotheses

Voice as Both Promotive and Prohibitive

Hirschman (1970) initially introduced the concept of voice as a way for an employee to express organizational dissatisfaction. However, the literature on voice has mainly focused on organizational improvements, such as ideas for new products or projects (Morrison, 2011). Correspondingly, recent empirical and conceptual developments have underlined the lack of focus on employees’ upward communication regarding factors that may harm the organization (Morrison, 2011; Van Dyne et al., 2003). Therefore, Liang et al. (2012) introduced the distinction between promotive and prohibitive voice. Although the concepts share important similarities—that is, they both aim to change the status quo to improve organizational functioning—there are also some important differences. Promotive voice is future-oriented and focuses on “realizing ideals and possibilities” (Liang et al., 2012, p. 75), whereas prohibitive voice is both past- and future-oriented and concerns “stopping and preventing harm” in the organization (Liang et al., 2012, p. 75). Importantly, prohibitive voice involves more personal risk for the employee because speaking up regarding possible harm to the organization may highlight the failure of the individual responsible for this dysfunction. In addition, the good intentions behind prohibitive voice may not be as easily recognized and the employee may risk being seen simply as a complainer (Liang et al., 2012). Chamberlin et al. (2017) found meta-analytical support for this notion and illustrated how leaders rate employees who voice prohibitive as lower performers than employees who voice their promotive ideas. Correspondingly, Morrison (2011) argued that the strength of the motivating forces predicting voice may differ depending on its promotive or prohibitive content. Specifically, Morrison (2011) suggested that self-protective motives may be an especially important motivating factor with respect to prohibitive voicing, due to the increased interpersonal risk associated with it. Liang et al. (2012) empirically supported the notion that self-protective motives are more important for prohibitive voice by demonstrating how psychological safety is significantly stronger related to prohibitive voice compared with promotive voice.

To our knowledge, the only study that has investigated the relationship between leadership and promotive as well as prohibitive voice is a cross-sectional study by Svendsen et al. (2016). They found that participatory leadership is related to prohibitive voice differently than to promotive voice. Specifically, they found a direct effect on promotive voice and an indirect effect, through psychological safety, on prohibitive voice, which aligns with the notion that self-protective motives are more important for prohibitive than for promotive voice (Liang et al., 2012; Morrison, 2011). Accordingly, we test the hypothesis regarding promotive and prohibitive voice separately in our study and conduct post hoc analysis of their potential differential effect to account for the recent conceptual and empirical developments (Liang et al., 2012; Morrison, 2011; Svendsen et al., 2016; Van Dyne et al., 2003).

Transformational Leadership Behaviors and Promotive and Prohibitive Voice

Transformational leadership is one of the most investigated and applied leadership constructs (Yukl, 2008). It consists of the four following dimensions of behaviors: idealized influence, inspirational motivation, intellectual stimulation, and individualized consideration (Liu et al., 2010). A key characteristic of transformational leaders are their positive orientation toward change and willingness to challenge status quo in organizations (Bass, 1985; Herold, Fedor, Caldwell & Liu, 2008). The specific transformational leadership behaviors should all be related to the individuals’ propensity to voice, both promotive and prohibitive. For example, a leader who shows individualized consideration by listening and personally interacting with employees may stimulate promotive and prohibitive voice by giving the employees room and access to express themselves. By experiencing consideration from the leader, an employee may reciprocate by displaying consideration for the good of the organization and voice concerns and ideas to safeguard the organization’s interests. Moreover, intellectually stimulating transformational leadership behaviors may also encourage promotive and prohibitive voice by motivating employees to look at things differently and critically, thereby increasing the number of ideas or concerns generated by the employees (Liu et al., 2010). Furthermore, a leader who shows intellectual stimulation aims to use the inputs from his or her employees (Bass & Riggio, 2006), which may motivate employees to express themselves both promotively and prohibitively. Last, the transformational leader also commits, inspires, and empowers employees in order to accomplish organizational goals (inspirational motivation and idealized influence) and may increase the employees’ motivation to go beyond formal job description and express themselves both promotively and prohibitively to reach these goals (Bass & Riggio, 2006; Detert & Burris, 2007; Liu et al., 2010). In sum, the transformational leader behaviors should inspire employees to believe that their supervisors are oriented toward the future rather than to preserve the status quo, and thus stimulate both promotive and prohibitive voice (Detert & Burris, 2007).

Overall, the results from studies on the relationship between transformational leadership and promotive voice are mixed, albeit generally positive. In a cross-sectional study, Liu et al. (2010) found that transformational leadership has a positive direct relationship with improvement-oriented voice in MBA students in China. Furthermore, in a study of U.S. restaurant employees, Detert and Burris (2007) found evidence that transformational leadership is cross-sectional but not longitudinally related to improvement-oriented voice. Additionally, a recent study by Duan et al. (2017) found that transformational leadership has a positive relationship to promotive voice.

In sum, we have argued that theoretically, transformational leadership behaviors, as perceived by the employees, should have a positive direct relationship to both promotive and prohibitive voice behavior. Based on these theoretical arguments and prior empirical studies, we propose the following hypotheses:

Hypothesis 1a: Transformational leadership, as perceived by employees, is positively related to promotive voice.

Hypothesis 1b: Transformational leadership, as perceived by employees, is positively related to prohibitive voice.

Previous research on promotive and prohibitive voice has often compared the effect of the predictive variables in order to explore the potential differential effect between them and to investigate whether the two interrelated forms are indeed distinguishable constructs in their own right. Liang et al. (2011) and Svendsen et al. (2016) found that self-protective motives such as psychological safety are more strongly related to prohibitive than promotive voice. Previous studies confirm that transformational leadership is significantly related to psychological safety (Detert & Burris, 2007), suggesting that transformational leadership may be especially well suited for eliciting prohibitive voice. In line with this research and to establish whether transformational leadership has a differential effect on promotive and prohibitive voice, we make a post hoc analysis comparing the relative effect of transformational leadership on promotive and prohibitive voice.

The Moderating Effect of Job Autonomy

The use of job autonomy is based on the theoretical framework of the theory of planned behavior (Ajzen, 1985, 1991, 2011; Ajzen & Madden, 1986) and self-determination theory (Gagné & Deci, 2005, Ryan & Deci, 2000, 2006). According to these theories, the perception of an autonomy supportive environment makes a unique contribution to an individual’s intentions and motivation to exert a behavior. Specifically, research on the theory of planned behavior has found that an autonomy-supportive context increases an individual’s experience of nonpressuring forms of social influence, which in turn increases the individual’s intention to carry out a social behavior such as voice (Ajzen, 1991, 2011). Moreover, research on self-determination theory has shown that individuals operating under autonomy supportive contexts are more intrinsically motivated, which increases the individuals’ propensity to act and make a continuous effort on a task or behavior. This form of motivation and continuous effort may positively affect an individual’s propensity to voice (Gagné & Deci, 2005; Ryan & Deci, 2006).

In line with this idea, Yukl (2008) argued that certain contexts may support and reinforce the effect of leadership behaviors. One such job characteristic may be perceived job autonomy, which is an important contextual factor for a number of positive outcomes, such as job satisfaction, turnover intentions, and organizational commitment (Humphrey, Nahrgang, & Morgeson, 2007). Job autonomy pertains to employee influence on how to decide on, perform, and schedule work tasks (Hackman & Oldham, 1980; Humphrey et al., 2007).

Dust, Resick, and Mawritz (2014) argued that a context of high job autonomy has fewer formal or explicit behavioral expectations, which enables employees to feel that they can accept the leader’s influence, thereby increasing the effect of transformational leadership on behavioral outcomes. We follow this line of argumentation and suggest that employees’ job autonomy may strengthen the relationship between transformational leadership and voice. For example, job autonomy may strengthen the effect of individualized consideration and intellectual stimulation by constituting a context that increases the employees’ perception of being listened to and valued for their input (Vroom, 1964). Along the same lines, a study by Choi (2007) found that employees who experience high job autonomy tend to define their work roles in a more broad and inclusive manner, and therefore feel that speaking up is one of their job responsibilities (Choi, 2007). Accordingly, the experience of job autonomy may also strengthen the effect of inspirational motivation and idealized influence by giving the employees the space to go beyond their formal job description in order to reach the organizational goals. In sum, job autonomy is therefore likely to strengthen the employee’s belief and motivation to support the transformational leaders’ challenge of the status quo, which is crucial for creating motivation to voice. However, to the best of our knowledge this relationship has never been investigated. Therefore, with regard to the present study, we argue that autonomy and transformational leadership have an interactional effect on both promotive and prohibitive voice. Accordingly, we posit the following hypotheses:

Hypothesis 2a: Job autonomy positively moderates the relationship between perceived transformational leadership and promotive voice, so that the effect of perceived transformational leadership will be stronger when job autonomy is also high.

Hypothesis 2b: Job autonomy positively moderates the relationship between perceived transformational leadership and prohibitive voice, so that the effect of perceived transformational leadership will be stronger when job autonomy is also high.

Method

Sample and Procedure

We invited all white-collar workers to participate in a survey at a global medical technology company based in Norway. The survey had a two-wave panel design, with 10 months between data collections. The invitation to participate was sent to employees via their professional email and emphasized that participation was voluntary and confidential, but that e-mail addresses would be used to match participants with their supervisors. Employees were not organized in teams, but one leader represented each unit.

At both time points, 261 employees were invited to participate. Response rates were 73.18% (n = 191) at Time 1 (T1) and 51.34% (n = 134) at Time 2 (T2). After removing all individuals who had missing values on entire scales, we reached a total sample of 173 employees at T1 and 124 at T2. Finally, we matched the data of the two time points, which resulted in a longitudinal sample of 120 employees. This longitudinal sample included fewer than 10% missing values per person (according to Kline, 1998), which were imputed using the expectation-maximization algorithm using the NORM software (Graham, Cumsille, & Elek-Fisk, 2003) to prevent biases caused by not having completely random missing data processes. On average, the employees in the final sample were 42 years old and had worked for this company for 11 years. Sixty-one percent were males, and 39% were females. Seventy-five percent had a bachelor degree or higher and 91% did not have leadership responsibilities.

We computed chi-square tests and independent t tests to check for a systematic dropout of respondents in the two waves on the cleaned sample. Respondents who answered at T1 but not at T2 (n = 53) differed significantly from employees who responded at both times (n = 120) in leadership responsibility only. The dropout rate for employees with leadership responsibility was larger than for those without leadership responsibility: χ²(1) = 5.79, p = .016. There was no gender difference between responders and nonresponders, χ²(1) = 0.42, p = .516; education, χ²(1) = 1.22, p = .269; age, t(170) = 0.32, p = .749; job satisfaction, t(170) = −0.49, p = .626; transformational leadership, t(171) = 0.18, p = .855; promotive voice, t(171) = 1.55, p = .124; prohibitive voice, t(171) = 0.32, p = .750; and job autonomy, t(171) = -0.73, p = .467.

Measures

All the applied scales in this study had been previously published and validated. Scales that were originally formulated in English were translated to Norwegian and then back to English (Brislin, 1986). All the scales showed satisfactory reliabilities at both measurement times. The means, standard deviations, zero-order correlations, and reliability coefficients are shown in Table 1.

Table 1.

Means, Standard Deviations, and Cross-Lagged Correlations for T1 and T2 (N =120).

	M	SD	1	2	3	4	5	6	7	8	9	10	11	12	13
T1
1. Gender	0.61	0.49	—
2. Age	42.22	9.56	−.03	—
3. Supervisor responsibility	0.09	0.29	.14	.17	—
4. Education	0.75	0.43	.29**	−.19*	.12	—
5. Job satisfaction	3.95	0.78	.06	.05	.07	−.16	—
6. TL	3.59	0.70	−.05	−.05	.21*	.02	.21*	(.96)
7. Job autonomy	3.84	0.80	.05	.11	.11	.04	.21**	.35**	(.75)
8. Promotive voice	3.86	0.55	.22*	.00	.22*	.29**	−.15	.16	.09	(.82)
9. Prohibitive voice	3.65	0.53	.03	.12	.19*	.02	.06	.25**	.12	.40**	(.73)
T2
10. TL	3.58	0.66	.09	−.04	.16	.19*	.11	.60**	.31**	.28**	.20*	(.95)
11. Job autonomy	3.90	0.73	.19*	.08	.19*	.10	.28**	.30**	.55**	.28**	.19*	.34**	(.78)
12. Promotive voice	3.88	0.55	.14	−.17	.13	.32**	.04	.17	.14	.54**	.34**	.24**	.25**	(.88)
13. Prohibitive voice	3.65	0.56	.07	.11	.14	.00	.12	.33**	.18	.41**	.59**	.35**	.22*	.35**	(.80)

Note. TL = transformational leadership.

p < .05. **p < .01.

As shown in Table 1, promotive voice was significantly correlated with education, transformational leadership (at time point two), and job autonomy. Prohibitive voice, on the other hand, was significantly correlated with transformational leadership at both time points and job autonomy.

Transformational Leadership

Transformational leadership behavior was measured using the Multifactor Leadership Questionnaire Form 5x (Bass & Avolio, 1995). Idealized influence (attributed and behavioral), inspirational motivation, intellectual stimulation, and individualized consideration were measured with four items per subdimension, resulting in 20 items. A sample item is “My leader provides a compelling vision for the future.” Responses were made using a 5-point Likert-type scale (from 1 = not at all to 5 = frequently, if not always). Because we had no hypotheses relating to the different subdimensions, we combined them into one overall composite.

Promotive and Prohibitive Voice Behavior

We measured promotive and prohibitive voice behavior using a 10-item scale developed by Liang et al. (2012), with five items to measure each construct. The questions were rated on a 5-point Likert-type scale (1 = strongly disagree to 5 = strongly agree). This scale captured the divergent aspects of promotive and prohibitive voice behavior. The wording was changed slightly to make it suitable for self-reporting. Sample items include “I suggest new projects which may be beneficial to the work unit” (promotive voice) and “I speak up honestly with problems that might cause serious loss to the work unit, even when/though dissenting opinions exist” (prohibitive voice).

Job Autonomy

Job autonomy was measured using a three-item scale developed by Jønsson and Jeppesen (2013). The three items concentrated on issues proximal to the employee and represented major issues pertaining to decisions about work, addressing both doing and organizing work. A sample item is “How much influence do you experience that you have on how the daily work tasks are organized?” Questions were rated on a 5-point Likert-type scale (1 = nothing to 5 = a lot).

Control Variables

Based on previous literature, we added age, gender, supervisory responsibility, job satisfaction, and education as control variables (Morrison, 2011). Age was measured by years and months, and gender was dummy-coded (0 = female, 1 = male). Supervisory responsibility was dummy-coded based on organizational records (0 = not supervisor and 1 = supervisor). Job satisfaction was measured by the item “All together, how satisfied are you with your job?” (Wanous & Reichers, 1997). Responses were made using a 5-point Likert-type scale (1 = very dissatisfied to 5 = very satisfied). Education was measured on five categories (elementary school, high school, undergraduate studies, bachelor degree, and master’s degree or postgraduate studies). Since education is not a continuous variable we dummy-coded elementary school, high school, and undergraduate studies into the nonacademics category (0—referring to not holding an academic degree). Bachelor degree and master’s degree or higher was coded into academics (1—referring to holding an academic degree).

Analytical Approach

According to previous research, such as Schriesheim, Castro, Zhou, and DeChurch (2006), a large effect of transformational leadership may exist at the individual level. Consequently, we were interested in the employees’ individual perceptions of their supervisors, not the supervisors’ general leadership styles. This approach also made sense contextually because the work units were not organized in teams, which meant that employees did not have high levels of social interaction or work interdependence and interrater agreement was likely to be low (Klein, Smith, & Sorra, 2001). Nevertheless, because our study consisted of employees drawn from separate units represented by one leader, our sample violated the independence assumption and may result in spuriousness due to data clustering (Raudenbush & Bryk, 2002). Therefore, we made a cluster robust, complex model in Mplus, version 7.11, using manifest variables to adjust for the clustered nature of our data. The calculated ICC1 for transformational leadership was small (T1: ICC1 = .09), which indicates controlling for the clusters, but computing the analyses on an individual level.

Prior to hypotheses testing, we conducted confirmatory factor analyses (CFA) to provide evidence for discriminant validity of the measured variables at both time points, using the statistical software, Mplus version 7.11, and applying Maximum Likelihood Robust estimation of model fit. First, we compared an overall measurement model that included four latent factors (transformational leadership, job autonomy, promotive voice, and prohibitive voice) with a single-factor model.

The proposed four-factor model demonstrated satisfactory model fit¹ at T1, χ²(475) = 694.81, p < .001, root mean square error of approximation (RMSEA) = .06, 90% confidence interval (CI) [0.05, 0.07], comparative fit index (CFI) = .89, Tucker–Lewis index (TLI) = .88, standardized root mean square residual (SRMR) = .07, as well as at T2, χ²(475) = 691.36, p < .001, RMSEA = .06, 90% CI [0.05, 0.07], CFI = .90, TLI = .89, SRMR = .07. The CFAs of the single-factor models revealed poor model fits at both time points²: T1: χ²(486) = 1110.76, p < .001, RMSEA = .10, 90% CI [0.10, 0.11], CFI = .70, TLI = .67, SRMR = .11; T2: χ²(486) = 1279.44, p < .001, RMSEA = .12, 90% CI [0.11, 0.12], CFI = .65, TLI = .62, SRMR = .12. A Satorra–Bentler scaling-corrected χ² difference test (Satorra & Bentler, 1999) showed that, at both times, the four-factor model was significantly better than the single-factor model (T1: Δχ²[11] = 229.24, p < .001; T2: Δχ²[11] = 415.92, p < .001). These results provide some evidence against bias from common method variance (Podsakoff, MacKenzie, Scott, Lee, & Podsakoff, 2003).

In addition, to assess whether prohibitive and promotive voice were separate constructs, we conducted CFAs with the two latent factors and tested them against a single-factor model. The two-dimensional model showed a satisfactory fit at both times: T1, χ²(34) = 38.46, p = .275, RMSEA = .03, 90% CI [0.00, 0.08], CFI = .98, TLI = .98, SRMR = .05, and T2, χ²(34) = 51.67, p = .027, RMSEA = .07, 90% CI [0.02, 0.10], CFI = .95, TLI = .94, SRMR = .06. In contrast, the model fits of the single-factor model were not acceptable: T1, χ²(35) = 110.01, p < .001, RMSEA = .13, 90% CI [0.11, 0.16], CFI = .72, TLI = .65, SRMR = .11, and T2, χ²(35) = 147.75, p < .001, RMSEA = .16, 90% CI [0.14, 0.19], CFI = .71, TLI = .62, SRMR = .12. The Satorra–Bentler scaling-corrected χ² difference test revealed significantly better model fits for the two-factor model than for the single-factor model at T1 (Δχ²[1] = 40.63, p < .001) and at T2 (Δχ²[1] = 368.83, p < .001).

We also tested whether the used measurements were invariant across both times. We conducted measurement invariance analyses for the four constructs: transformational leadership, job autonomy, promotive voice, and prohibitive voice. Following Vandenberg and Lance (2000) in the sequence of testing nested models, we first performed an unconstrained baseline model (configural invariance) showing that the constructs were conceptualized in the same way at both time points. Second, we constrained the factor loadings (metric invariance) to be equal at both times, and then tested whether the item intercepts were invariant (scalar invariance) across time. Based on Vandenberg and Lance’s (2000) method, we used the χ²-difference test and ΔCFI comparisons to evaluate the change in model fit within the nested models. Measurement invariance across time is given if Δχ² is not significant and if changes in CFI are −.01 or less (Vandenberg & Lance, 2000). Table 2 shows that the CFI comparisons were below the recommended cutoff points for all constructs.

Table 2.

Confirmatory Factor Analysis Models Testing Measurement Invariance Across Time.

Model	χ²	df	p	CFI	RMSEA	Model comparison	Δχ²	Δdf	p
Transformational leadership
1. Configural invariance	480.50	302	<.001	.938	.070	Model 2 ↔ 1	27.49	15	.025
2. Invariance of the factor loadings	507.88	317	<.001	.934	.071	Model 3 ↔ 1	45.89	30	.032
3. Invariance of the item intercepts	526.51	332	<.001	.933	.070	Model 3 ↔ 2	17.95	15	.265
Promotive voice
1. Configural invariance	17.45	10	.065	.974	.079	Model 2 ↔ 1	1.92	4	.751
2. Invariance of the factor loadings	20.10	14	.127	.978	.060	Model 3 ↔ 1	3.32	8	.913
3. Invariance of the item intercepts	22.38	18	.216	.985	.045	Model 3 ↔ 2	1.37	4	.850
Prohibitive voice
1. Configural invariance	16.55	10	.085	.975	.074	Model 2 ↔ 1	6.01	4	.198
2. Invariance of the factor loadings	22.60	14	.067	.967	.072	Model 3 ↔ 1	7.33	8	.502
3. Invariance of the item intercepts	24.34	18	.144	.976	.054	Model 3 ↔ 2	1.08	4	.898
Job autonomy
1. Configural invariance	0.00	0	<.001	1.000	.000	Model 2 ↔ 1	0.43	2	.806
2. Invariance of the factor loadings	0.43	2	.806	1.000	.000	Model 3 ↔ 1	2.64	4	.621
3. Invariance of the item intercepts	2.64	4	.621	1.000	.000	Model 3 ↔ 2	2.36	2	.308

Note. N = 240. CFI = comparative fit index; RMSEA = root mean square error of approximation.

The χ²-difference tests were significant for transformational leadership only when comparing the configural with the metric and scalar models. However, Cheung and Rensvold (2002) suggested relying on ΔCFI because χ²-difference tests are very sensitive to sample size and nonnormality. Hence, the CFI comparisons of all constructs provided evidence for psychometrically equivalence across both time points.

Results

The Direct Effect of Transformational Leadership on Promotive and Prohibitive Voice

To test the time-lagged results, we conducted a cross-lagged analysis that enabled us to explore the reciprocal dynamics between the constructs over time (Podsakoff et al., 2003). Table 3 illustrates the results of the conducted autoregressive model (χ²[12] = 23.50; p = .024; RMSEA = .09; 90% CI [0.32, 0.14]; CFI = .93; TLI = .77; SRMR = .04).

Table 3.

Regression Results of the Autoregressive Model (Dependent Variables: Promotive Voice T2; Prohibitive Voice T2; Transformational Leadership T2); N = 120.

Variable	Promotive voice T2	Prohibitive voice T2	Transformational leadership T2
Age	−.16* (.07)	.06 (.09)
Education	.18* (.08)	−.05 (.06)
Gender	−.02 (.10)	.07 (.08)
Leadership responsibility	.02 (.09)	−.02 (.08)
Job satisfaction	.16* (.07)	.07 (.07)
Transformational leadership (TL) T1	.04 (.10)	.20* (.10)	.53** (.12)
Promotive voice (PromV) T1	.43** (.08)		.18* (.08)
Prohibitive voice (ProhV) T1		.51** (.09)	−. 02 (.08)
Job autonomy (PPI) T1	.09 (.07)	.06 (.07)	.11 (.09)
TL × PPI	.13 (.09)	.18* (.09)
PromV × PPI			.03 (.07)
ProhV × PPI			−.06 (.07)
R ²	.35	.40	.40

Note. N = 120. Reported coefficients of the autoregressive model are standardized (standard errors in parentheses). TL = transformational leadership; PPI = job autonomy. PromV = promotive voice; ProhV = prohibitive voice.

p < .05. **p < .01.

As shown in Table 3, Hypothesis 1a was not supported because we found no significant effect of transformational leadership on promotive voice. Interestingly, we did find support for the opposite, as the self- perception of promotive voice (T1), significantly predicted the employees’ perceived transformational leadership (T2): β = .18, p = .02. However, Hypothesis 1b was supported as we found a significant effect of transformational leadership on the self-perception prohibitive voice (T2) when controlling for self-perception of prohibitive voice (T1): β = .20, p = .04. As a post hoc analysis, to test whether transformational leadership had a significantly different effect on promotive versus prohibitive voice, we conducted a likelihood ratio test in which we compared a model that constrained the two coefficients to be equal with one where they were allowed to vary freely. If the model where they vary has significantly better fit (by comparing the difference in chi-squared values with one degree of freedom), this is evidence that the coefficients are indeed different. In conducting this analysis, we found a marginally significant effect for transformational leadership, meaning that the effect was only significant at a p >.10 level (Satorra-Bentler scaled Δχ²[1] = 2.74, p = .10).

The Moderating Role of Job Autonomy

As shown in Table 3, we found no significant interactional effect of perceived transformational leadership and job autonomy on promotive voice; therefore, Hypothesis 2a was not supported. However, as illustrated in Figure 1, Hypothesis 2b was supported because we found a positive and significant moderating effect of job autonomy on prohibitive voice; β = .18, p = .05.

Figure 1.

Interaction plot.

Discussion

The goal of this study was to examine the time-lagged effect of transformational leadership and job autonomy on the employees’ self-perception of promotive and prohibitive voice. Our results provide support for the notion that transformational leadership affects prohibitive voice positively. For promotive voice, our cross-lagged analysis revealed that the dynamics work in the opposite direction, in that employees who saw themselves as voicing promotive in general at Time Point 1 came to rate their leader as more transformational at Time Point 2. Post hoc analysis further showed a marginally significantly stronger relationship to prohibitive than promotive voice, indicating that transformational leadership had a stronger effect on the self-perception of prohibitive voice than on promotive voice. Last, we found no support for the moderating role of job autonomy on promotive voice, although job autonomy significantly strengthened the effect of transformational leadership on prohibitive voice. Below, we elaborate on how our results give important contributions to literature on both voice and transformational leadership.

Theoretical Contributions

This study offers four key implications for theory and research on transformational leadership and voice. First, by being the first study to look at transformational leadership and voice by applying a time-lagged analysis, we provide a novel perspective on what kind of leadership behaviors may be most crucial for eliciting voice from employees. In accordance with Detert and Burris (2007), we did not find a significant relationship between transformational leadership and temporal increases in promotive voice. Therefore, our results and those of Detert and Burris (2007) may suggest that specific leadership behaviors such as leadership openness or participative supervisory behavior may be more important than transformational leadership for stimulating promotive voice (Detert & Burris, 2007; Svendsen et al., 2016). In their longitudinal study, Detert and Burris (2007) found that leadership openness was significantly related to improvement-oriented voice. Such participative supervisory behaviors are potentially more explicit signals that encourage employees to express themselves promotively than transformational leadership (Tangirala & Ramanujam, 2012).

Second, our study extends the theoretical understanding of the causal dynamics between transformational leadership and voice, and provides a new perspective on how this relationship may be depending on what the employee wish to express. Our results indicate that employees who perceived themselves as high in promotive voice at Time Point 1 perceived their leader as more transformational at Time Point 2. An explanation for this may be that the employees who perceived themselves as high in promotive voice also felt they were more likely to be listened to and acknowledged by their leader, leading them to perceive their leader as more transformational at Time Point 2. The dynamics for prohibitive voice over time may be different, in that the employee needs to gain an individual perception of their leader as someone who listens to them and cares for their individual needs before they feel safe to voice their concerns regarding the organizational functioning. Accordingly, our results align with and further develop the conceptual work of Morrison (2011) and others (Van Dyne, Ang & Botero, 2003; Liang et al., 2012; Svendsen et al., 2016) by illustrating the importance of investigating the antecedents of promotive and prohibitive voice separately.

Third, our study also provides a novel perspective on earlier research on promotive voice. Importantly, our results shed light on the failure to find a direct significant time-lagged relationship between transformational leadership and promotive voice (Detert & Burris, 2007, in the time-lagged part of their study). Although other studies have found a significant direct cross-sectional relationship between transformational leadership and voice, these cross-sectional results may still appear, even though the direction of the results may be opposite of what was expected (Detert & Burris, 2007; Liu et al., 2010). Critics of transformational leadership theory argue that a problematic aspect of transformational leadership is how the theory confounds the relationship between cause and effect (Van Knippenberg & Sitkin, 2013). Our results validate this point further by suggesting how the relationship between transformational leadership and self-perceptions of promotive voice may indeed be vice versa, and they also stress the need for more research that allows for cross-lagged analyses in order to explore this finding further.

Fourth, our results extend the theoretical knowledge on how contextual factors may affect the relationship between leadership and voice. According to our results, the perception of job autonomy significantly strengthened the effect of transformational leadership on prohibitive voice over time, but not on promotive voice. This finding suggests that the moderating effect of job autonomy may be more important for prohibitive voice than promotive voice in terms of increasing the influence of transformational leadership. An interpretation of this is that the perception of an autonomy supportive context strengthens the transformational leaders’ effort to convince the employees into believing that expressing oneself is indeed welcomed and will not lead to negative consequences (Ajzen, 1985, 1991). This stimulation of self-protective motives, resulting from a combination of transformational leadership and job autonomy, may be more important for prohibitive voice than for promotive voice (Liang et al., 2012). Furthermore, in line with self-determination theory, job autonomy may have contributed to more intrinsically motivated employees, which in turn may have resulted in a greater effort and persistence from the employees (Gagné & Deci, 2005). This persistence and effort may be especially crucial when voicing prohibitive concerns, as the probability for negative reactions and resistance to change is more likely to occur (Liang et al., 2012). Thus, our results indicate that job autonomy may strengthen the transformational leaders’ efforts to create intrinsically motivated employees who contribute and make sacrifices for the benefit of the group. Accordingly, employees are more motivated to adjust their working roles to be broader and more inclusive, which would include speaking up about factors that may harm the organization (Chao, 2012; Choi, 2007; Major, Kozlowski, Chao, & Gardner, 1995). This intrinsic motivation to move beyond the formal job description, initiated by the combination of transformational leadership and job autonomy, may not be as important for promotive voice because presenting ideas and suggestions may be considered part of the job description (such as suggestion boxes or workshops to generate ideas).

Practical Implications

This study also has some important findings for practitioners looking to increase their employees’ voice. First, we found that transformational leadership, which is one of the most applied leadership development tools (Yukl, 2012), may be especially important for increasing employees’ prohibitive voice. However, more specific leadership behaviors, such as leadership openness or leadership consulting behavior, could be more effective at increasing promotive voice (Detert & Burris, 2007; Svendsen et al., 2016; Tangirala & Ramanujam, 2012). In addition, the perception of job autonomy may be essential for increasing the effect of transformational leadership on prohibitive voice in the long run. Accordingly, organizational initiatives that give employees autonomy on proximal issues may be an important tool for increasing the effect of transformational leadership on prohibitive voicing.

However, our results suggest that transformational leadership may not be a panacea for increasing promotive voice. Accordingly, managers should also consider other contextual factors when trying to increase promotive voice. For example, leaders could strive to make a coherent organizational culture where voice is rewarded both emotionally (e.g., praise, public recognition, increased performance evaluations) and instrumentally (pay, bonus, promotion etc.).

Limitations and Further Research

This study has several strengths, such as a time-lagged design and a statistical methodology, which enable us to take the clustered design of our data into account. However, it also has some limitations that are worth noting. First, all data came from a single source (the employees), which may introduce common method bias into the results. However, controlling for the initial value should dampen concerns for elevated relationships owing to common method bias (Podsakoff et al., 2003). Furthermore, recent simulation studies have confirmed that common method bias only deflates interaction effects, which suggests that our interaction effect should be robust (Podsakoff, MacKenzie, & Podsakoff, 2012; Siemensen, Roth, & Oliveira, 2010). In addition, the divergent results between promotive and prohibitive voice should not be attributed to common method bias (Podsakoff et al., 2003). Accordingly, in order to dampen the concern for common method bias and to more accurately capture the causality between the constructs, future studies could apply alternative methods such as diary studies to record actual occurrences of voice in the organization.

Second, our results were obtained from a sample of white-collar workers in Norway, a group that has a culture characterized by low power distance and a long tradition of employee participation. Hence, the Norwegian culture itself may convey the appropriateness of expressing oneself, which may have lowered the effect of both transformational leadership and job autonomy. Accordingly, these results may not be generalizable to cultures with high power distance or to blue-collar workers. Therefore, future studies could examine cultural influences on the effect of transformational leadership on promotive and prohibitive voice.

A third limitation is the way we measured and analyzed our hypotheses at an individual level, although transformational leadership may exist as a group-level construct (Bass & Riggio, 2006) and have different effects at the group versus individual levels (Avolio & Bass, 1995). Future research could examine transformational leadership at these two levels and sort out potential differences.

Fourth, in accordance with previous theory and research, we combined the behavioral factors in transformational leadership into one construct in our study. However, Van Knippenberg and Sitkin (2013) highlighted how transformational leadership theory and research do not specify the unique contributions of each behavioral dimension and how moderators may affect these dimensions uniquely and differently. Hence, it may be that some dimensions of transformational leadership are more important for promotive voice (i.e., intellectual stimulation), and that the moderating effect of job autonomy would have been relevant to a specific dimension but not for others. Because we combined these behavioral dimensions into one construct, we have potentially precluded the effects of some of the transformational leadership behaviors and their contingency on job autonomy as a potential moderator.

Fifth, although our study shows how leadership and job autonomy affect promotive and prohibitive voice differently, it does not provide insights into the mediational mechanisms that may create these effects. Therefore, future studies should explore whether the stronger effect of transformational leadership on prohibitive voice may be due to the stimulation of self-protective motives such as trust, psychological safety, and psychological contract. Moreover, in our study we focused on job autonomy as an important boundary condition for transformational leadership, but there may be other exciting factors that can strengthen or hamper the relationship between transformational leadership and voice. For example, if developing the ideas or methods proposed by the employee leads to an increased workload, this could hamper the relationship between transformational leadership and promotive voice. Accordingly, if the leader or context do not reward the employee (i.e., give the employee a share of the increased profit or at least publicly acknowledge him or her for the idea), this may affect whether the employee believes it is worthwhile speaking up. For prohibitive voice, however, there may be exciting avenues for research in exploring how contextual factors evoking personal morals may affect the relationship with transformational leadership. For example, it may be that contexts conveying moral symbols that evoke the employee’s moral identity (i.e., words, images or mundane objects) will strengthen the employee’s perception of the leader as someone who values openness and integrity and thereby stimulate prohibitive voice (Desai & Kouchaki, 2017).

Last, considering the study’s results, it would also be interesting to explore the individual differences that may affect the relationship between transformational leadership and promotive and prohibitive voice. It could be that employees who are high in self-efficacy or have a proactive personality will benefit less from transformational leadership behaviors because they would voice their ideas or concern regardless of the leaders’ behavior, whereas the opposite may be the case for individuals who are low in self-efficacy or do not have a proactive personality. Furthermore, it could be interesting to explore the effect of the perceived power distance between the transformational leader and the employee. Some authors have argued that a highly charismatic transformational leader may lead to a high level of personal identification with the leader, which may create a dependency on the leader (Kark, Shamir, & Chen, 2003). Accordingly, a leader who scores high on these characteristics may be considered intimidating for the employee, which could decrease the employee’s propensity to voice opinions. Correspondingly, further research could explore the potential nonlinear or bell-shaped curve of transformational leadership effectivity, in order to establish whether a dependency or intimidating effect may occur in employees who have a highly charismatic, transformational leader.

Conclusion

In light of the GM scandal mentioned at the start of this article, the main goal of our study was to explore how leadership relates to promotive and prohibitive voice over time, and how contextual variables may affect this relationship. The results of our study add important knowledge to the leadership–voice literature and practice by being the first study to show how transformational leadership and job autonomy relate differently to promotive and prohibitive voice over time. We encourage other researchers to continue exploring the various motivational backgrounds for promotive and prohibitive voice and to conduct additional time-lagged research in order to get a better grasp on the dynamics between the constructs.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Mari Svendsen is an associate professor at Kristiania University College in Norway. Her research interests include leadership, participation and employee proactive behavior.

Christine Unterrainer is a postdoc researcher in Work and Organizational Psychology at the Institute of Psychology, University of Innsbruck, Austria. Her research includes participation, distributed leadership, democratic organizations, socio-moral climate, and employee behavior.

Thomas Faurholt Jønsson is an associate professor in Work and Organizational Psychology at Department of Psychology, Aarhus University, Denmark. He has done research in Distributed Leadership, Employee Involvement, motivation, and social identity.

References

Ajzen

(1985). From intentions to actions: A theory of planned behavior (pp. 11-39). In Kuhl

Beckmann

(Eds.), Action control: From cognition to behavior. Berlin, Germany: Springer-Verlag.

Ajzen

(1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179-211.

Ajzen

(2011). The theory of planned behaviour: Reactions and reflections. Psychology and Health, 26, 1113-1127.

Ajzen

Madden

T. J.

(1986). Prediction of goal directed behavior: Attitudes, intentions, and perceived behavioral control. Journal of Experimental Social Psychology, 22, 453-474.

Ashford

Sutcliffe

K. E.

Christianson

M. K.

(2009). Speaking up and speaking out: The leadership dynamics of voice in organizations. In Greenberg

Edwards

M. S.

(Eds.), Voice and silence in organizations (pp. 175-202). Bingley, England: Emerald.

Avolio

B. J.

Bass

B. M.

(1995). Individual consideration viewed at multiple levels of analysis: A multi-level framework for examining the diffusion of transformational leadership. The Leadership Quarterly, 6(2), 199-218. doi:10.1016/1048-9843(95)90035-7

Bass

B. M.

(1985). Leadership: Good, better, best. Organizational Dynamics, 13(3), 26-40. doi:10.1016/0090-2616(85)90028-2

Bass

Avolio

(1995). MLQ Multifactor Leadership Questionnaire. Redwood City, CA: Mind Garden.

Bass

B. M.

Riggio

R. E.

(2006). Transformational leadership. Mahwah, NJ: Erlbaum.

10.

Brislin

R. W.

(1986). The wording and translation of research instruments. In Lonner

W. J.

Berry

J. W.

(Eds.), Field methods in cross-cultural research (Vol. 8, pp. 137-164). Beverly Hills, CA: Sage.

11.

Burris

E. R.

(2012). The risks and rewards of speaking up: Managerial responses to employee voice. Academy of Management Journal, 55(4), 851-875. doi:10.5465/amj.2010.0562

12.

Chamberlin

Newton

D. W.

LePine

J. A.

(2017). A meta-analysis of voice and its promotive and prohibitive forms: Identification of key associations, distinctions, and future research directions. Personnel Psychology, 70, 11-71.

13.

Chao

G. T.

(2012). Organizational socialization: Background, basics, and a blueprint for adjustment to work. In Kozlowski

S. W. J.

(Ed.), The Oxford handbook of organizational psychology (pp. 579-614). New York, NY: Oxford University Press.

14.

Chatzisarantis

N. L. D.

Hagger

M. S.

Brickell

(2008). Using the construct of perceived autonomy support to understand social influence within the theory of planned behavior. Psychology of Sport and Exercise, 9, 27-44. doi:10.1016/j.psychsport.2006.12.003

15.

Cheung

G. W.

Rensvold

R. B.

(2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233-255. doi:10.1207/S15328007SEM0902_5

16.

Choi

J. N.

(2007). Change-oriented organizational citizenship behavior: Effects of work environment characteristics and intervening psychological processes. Journal of Organizational Behavior, 28, 467-484. doi:10.1002/job.433

17.

Desai

Kouchaki

(2017). Moral symbols: A necklace of garlic against unethical requests. Academy of Management Journal, 60, 17-28. doi:10.5465/amj.2015.0008

18.

Detert

Burris

(2007). Leadership behavior and employee voice: Is the door really open? Academy of Management Journal, 40, 869-884. doi:10.5465/AMJ.2007.26279183

19.

Duan

C. H.

(2017). Transformational leadership and employee voice behavior: A Pygmalion mechanism. Journal of Organizational Behavior, 38, 650-670. doi:10.1002/job.2157

20.

Dust

S. B.

Resick

R. J.

Mawritz

B. M.

(2014). Transformational leadership, psychological empowerment, and the moderating role of mechanistic–organic contexts. Journal of Organizational Behavior, 35, 413-433.

21.

Dutton

Ashford

Lawrence

Miner-Rubino

(2002). Red light, green light: making sense of the organizational context for issue selling. Organization Science, 13, 355-369.

22.

Gagné

Deci

E. L.

(2005). Self-determination theory and work motivation. Journal of Organizational Behavior, 26, 331-362.

23.

Graham

J. W.

Cumsille

P. E.

Elek-Fisk

(2003). Methods for handling missing data. Handbook of Psychology, 4, 87-114.

24.

Hackman

J. R.

Oldham

G. R.

(1980). Work redesign. Reading, MA: Addison-Wesley.

25.

Herold

D. M.

Fedor

D. B.

Caldwell

Liu

(2008). The effects of transformational and change leadership on employees’ commitment to a change: A multilevel study. Journal of Applied Psychology, 93(2), 346-357. doi:10.1037/0021-9010.93.2.346

26.

Hirschman

A. O.

(1970). Exit, voice, and loyalty: Responses to decline in firms, organizations, and states. Cambridge, MA: Harvard University Press.

27.

Humphrey

S. E.

Nahrgang

J. D.

Morgeson

F. P.

(2007). Integrating motivational, social, and contextual work design features: A meta-analytic summary and theoretical extension of the work design literature. Journal of Applied Psychology, 92, 1332-1356.

28.

Jønsson

Jeppesen

H. J.

(2012). Under the influence of the team? An investigation of the relationships between team autonomy, individual autonomy, and social influence within teams. International Journal of Human Resource Management, 24(1), 78-93. doi:10.1080/09585192.2012.672448

29.

Jønsson

Jeppesen

H. J.

(2013). A closer look into the employee influence: Organizational commitment relationship by distinguishing between commitment forms and influence sources. Employee Relations, 35, 4-19.

30.

Kark

Shamir

Chen

(2003). The two faces of transformational leadership: Empowerment and dependency. Journal of Applied Psychology, 88, 246-255.

31.

Klein

K. A.

Smith

D. B.

Sorra

J. S.

(2001). Is everyone in agreement? An exploration of within-group agreement in employee perceptions of the work environment. Journal of Applied Psychology, 86, 3-16. doi:10.1037/0021-9010.86.1.3

32.

Kline

R. B.

(1998). Principles and practices of structural equation modeling. New York, NY: Guilford.

33.

LePine

J. A.

Van Dyne

(1998). Predicting voice behavior in work groups. Journal of Applied Psychology, 83, 853-868. doi:10.1037/0021-9010.83.6.853

34.

Liang

Farh

I. C.

Farh

J. L.

(2012). Psychological antecedents of promotive and prohibitive voice: A two-wave examination. Academy of Management Journal, 55, 71-92. doi:10.5465/amj.2010.0176

35.

Lin

S. H.

Johnson

R. E.

(2015). A suggestion to improve a day keeps your depletion away: Examining promotive and prohibitive voice behaviors within a regulatory focus and ego depletion framework. Journal of Applied Psychology, 100, 1381-1397.

36.

Liu

Zhu

Yang

(2010). I warn you because I like you: Voice behavior, employee identifications, and transformational leadership. Leadership Quarterly, 21, 189-202. doi:10.1177/1742715012444055

37.

Major

D. A.

Kozlowski

S. W. J.

Chao

F. T.

Gardner

P. D.

(1995). A longitudinal investigation of newcomer expectations, early socialization outcomes, and the moderating effect of role development factors. Journal of Applied Psychology, 80, 418-431. doi:10.1037/0021-9010

38.

McClean

Burris

E. R.

Detert

J. R.

(2013). When does voice lead to exit? It depends on leadership. Academy of Management Journal, 56, 525-548. doi:10.5465/amj.2011.0041

39.

Milliken

F. J.

Morrison

E. W.

Hewlin

P. F.

(2003). An exploratory study of employee silence: Issues that employees don’t communicate upward and why. Journal of Management Studies, 40, 1453-1476. doi:10.1111/1467-6486.00387

40.

Morrison

E. W.

(2011). Employee voice behavior: Integration and directions for future research. Academy of Management Annals, 5, 373-412. doi:10.1080/19416520.2011.574506

41.

Podsakoff

P. M.

MacKenzie

S. B.

Podsakoff

N. P.

(2012). Sources of method bias in social science research and recommendations on how to control it. Annual Review of Psychology, 65, 539-569. doi:10.1037/0021-9010.88.5.879

42.

Podsakoff

P. M.

MacKenzie

S. B.

Scott

Lee

J. Y.

Podsakoff

(2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879-903. doi:10.1146/annurev-psych-120710-100452

43.

Rank

(2006). Leadership predictors of proactive organizational behavior: Facilitating personal initiative, voice behavior, and exceptional service performance (Unpublished doctoral dissertation). University of South Florida, Tampa, FL. Retrieved from http://scholarcommons.usf.edu/etd/2669

44.

Raudenbush

S. W.

Bryk

A. S.

(2002). Hierarchical linear models (2nd ed.). Thousand Oaks: Sage.

45.

Ryan

R. M.

Deci

E. L.

(2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68-78.

46.

Ryan

R. M.

Deci

E. L.

(2006). Self-regulation and the problem of human autonomy: Does Psychology need choice, self-determination, and will? Journal of Personality, 74, 1557-1586.

47.

Satorra

Bentler

P. M.

(1999). A scaled difference chi-square test statistic for moment structure analysis (Unpublished technical report). University Pompeu Fabra, Barcelona, Spain.

48.

Schriesheim

C. A.

Castro

S. L.

Zhou

DeChurch

L. A.

(2006). An investigation of path-goal and transformational leadership theory predictions at the individual level of analysis, Leadership Quarterly, 7, 323-352. doi:10.1016/j.leaqua.2005.10.008

49.

Siemensen

Roth

Oliveira

(2010). Common method bias in regression models with linear, quadratic, and interaction effects. Organizational Research Methods, 3, 456-476.

50.

Svendsen

Joensson

T. S.

(2016). Transformational leadership and change related voice behavior. Leadership & Organization Development Journal, 37, 357-368.

51.

Svendsen

Jønsson

Unterrainer

(2016). Participative supervisory behavior and the importance of feeling safe and competent to voice. Journal of Personnel Psychology, 15, 25-35.

52.

Tangirala

Ramanujam

(2012). Ask and you shall hear (but not always): Examining the relationship between manager consultation and employee voice. Personnel Psychology, 65, 251-282. doi:10.1111/j.1744-6570.2012.01248.x

53.

Vandenberg

R. J.

Lance

C. E.

(2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3, 4-69. doi:10.1177/109442810031002

54.

Van Dyne

Ang

Botero

I. C

. (2003). Conceptualizing employee silence and employee voice as multidimensional constructs. Journal of Management Studies, 40, 1359-1392. doi:10.1111/1467-6486.00384

55.

Van Knippenberg

Sitkin

S. B

. (2013). A critical assessment of charismatic-transformational leadership research: Back to the drawing board? Academy of Management Annals, 7(1), 1-60. doi:10.1080/19416520.2013.759433

56.

Vroom

V. H.

(1964). Work and motivation. New York, NY: Wiley.

57.

Wanous

J. P.

Reichers

A. E.

(1997). Overall job satisfaction: How good are single-item measures? Journal of Applied Psychology, 82, 247-252.

58.

Yukl

G. A.

(2008). How leaders influence organizational effectiveness. The Leadership Quarterly, 19(6), 708-722. doi:10.1016/j.leaqua.2008.09.008

59.

Yukl

G. A.

(2012). Leadership in organizations (8th ed.). Upper Saddle River, NJ: Prentice Hall.